7fdf3ac0-5b5c-4f0c-9959-d9a65d4de092

LIST OF CONTRIBUTORS
Joseph R. Blasi School of Management and Labor Relations,

Rutgers University, USA
Mollie Bowers Merrick School of Business,

University of Baltimore, USA
Robert Bruno Institute of Labor and Industrial Relations,

University of Illinois, USA
Victor G. Devinatz Department Management and Quantitative

Methods, Illinois State University, USA
Anita Jose Department of Economics and Management,

Hood College and also at
Centre for Conflict Resolution,
Salisbury University, USA
Bruce E. Kaufman Department of Economics,

Georgia State University, USA
Maya K. Kroumova School of Management,

New York Institute of Technology, USA
Douglas L. Kruse School of Management and Labor Relations,

David Lewin Anderson Graduate School of Man'agement,

University of California at Los Angeles (UCLA),
USA
Patrice M. Mareschal Graduate Department of Public Policy and

Administration, Rutgers University at Camden,
USA
vii
viii
E. Patrick McDermott Center for Conflict Resolution,

Kenneth McLennan Manufacturers Alliance/MAPI, USA
Daniel J. B. Mitchell Anderson Graduate School of Management &

School of Public Policy and Social Research,
University of California at Los Angeles (UCLA),
USA
Ruth Obar Department of Economics and Management,

Hood College and also at
Centre for Conflict Resolution,
Brian Polkinghorn Center for Conflict Resolution,
Stephen R. Sleigh International Association of Machinists and

Aerospace Workers, USA
James C. Sesil School of Management and Labor Relations,

Roland Zullo Institute of Labor and Industrial Relations,

University of Michigan, USA
INTRODUCTION
Volume 11 of Advances in Industrial and Labor Relations (AILR) contains

nine papers that deal with a variety of industrial relations topics. The first
four papers in this volume were winners of the 2001 AILR/Industrial Relations
Research Association (IRRA) "best papers" competition. ~ In the first of
these papers, Patrick McDermott, Anita Jose, Ruth Obar, Mollie Bowers and
Brian Polkinghorn provide the first evaluation of the U.S. Equal Employment
Opportunity Commission's mediation program, basing their analysis and
conclusions on field data obtained directly from program participants. In a
related paper that relies on participant observation and survey data, Patrice
Mareschal distinguishes and analyzes three approaches to mediation - broad,
narrow and bifocal - used by U.S. Federal Mediation and Conciliation Service
(FMCS) mediators. Then, Maya Kroumova, James Sesil, Douglas Kruse and
Joseph Blasi combine several secondary data sets to provide cross-sectional
and longitudinal analyses of the effects of employee stock option plans on
unionized and non-union companies' financial performance. In the fourth of
these best papers, Robert Bruno develops a conceptual model of union gover-
nance and draws on detailed survey data to analyze how members of a
Chicago-based Teamsters' local union assess their union's internal governance
during a period of an imposed trusteeship.
The next two papers in this volume also deal with selected aspects of union
behavior. Roland Zullo uses a pretest-posttest control group design to assess
effects of the Wisconsin AFL-CIO's political education outreach on union
members' perceptions of and preferences for political candidates in 1996
congressional races, finding support for a psychological voter model but not for
a rationale voter model. Then, using field research as well as archival and event
analysis to analyze the unsuccessful attempt of Illinois State University faculty
to create a faculty union, Victor Devinatz concludes that the university
administration effectively combated the union organizing drive by adopting
an adversarial approach that strongly emphasized the potential costs of
unionization to several key parties.
The final three papers in this volume provide retrospectives on selected
dimensions of 20th century industrial relations to forecast 21st century
developments. Scholar Daniel J. B. Mitchell analyzes the aging work force to
ix
x INTRODUCTION
determine if there will be a 21st century equivalent to the 20th century baby
boomer generation. Union staff officer Stephen Sleigh analyzes how worker
assets can be leveraged to develop a 21st century focus on corporate
governance, in contrast to the 20th century focus on workplace governance.
Management representative Kenneth McLennan draws out certain key lessons
for 21 st century labor markets based on still further development of the late
20th century trend toward globalization. 2
As a set, these nine papers provide an interesting mix of topics, invoke a
variety of theoretical perspectives and research designs, employ a wide range
of research methods, produce important findings, and yield well-reasoned and
sometimes provocative conclusions. In these respects as well as others, these
papers testify to the vibrancy and diversity of contemporary industrial relations
research.
NOTES
1. Shorter versions of these papers were presented at the national IRRA meeting, New
Orleans, LA, January 5, 2001.
2. Shorter versions of these papers will appear in Perspectives on Work, Vol. 5 (Fall)
2001. Champaign, IL: Industrial Relations Research Association and University of
Illinois Press.
David Lewin
Bruce E. Kaufman
Series Editors
HAS THE EEOC HIT A HOME RUN?
AN EVALUATION OF THE EQUAL
EMPLOYMENT OPPORTUNITY
COMMISSION MEDIATION
PROGRAM FROM THE
PARTICIPANTS' PERSPECTIVE
E. Patrick McDermott, Anita Jose, Ruth Obar,

Mollie Bowers and Brian Polkinghorn
I. INTRODUCTION
Alternative dispute resolution (ADR) programs are gaining increasing
acceptance in the workplace as more employers and employees become
concerned that settling disputes in a courtroom is a lengthy, time-consuming,
costly, and often-frustrating process that results in a win-lose outcome. The
report of the Commission on the Future of Worker Management Relations (The
Dunlop Commission), which included an investigation of the usefulness of A D R
methods in resolving workplace disputes, provided an additional impetus for
the examination of the merits of ADR. The unionized sector of the economy
has for decades used such techniques to resolve labor-management disputes.
Now, diverse organizations across the United States of America (USA) have
become receptive to ADR, and to mediation in particular.
Advances in Industrial and Labor Relations, Volume 11, pages 1--40.

© 2002 Published by Elsevier Science Ltd.
ISBN: 0-7623-0853-2
2 E. PATRICKMcDERMO'Iq"ET AL.
Mediation is a dispute resolution process in which a neutral third party,

who has no final decision-making authority, assists disputing parties in devel-
oping options for an acceptable resolution, t A central feature of a successful
mediation is the interaction among the concerned parties and the mediator to
arrive at a voluntary, mutually acceptable resolution. This can mean that the
parties, with the help of the mediator, engage in "give and take" resulting in
a compromise or, with the mediator's assistance, expand the possibilities
for resolution. In other words, mediation is a facilitated negotiation. Thus,
mediation attempts to focus on the relationship of the parties, rather than on
legalism. 2 This becomes important if parties want to preserve their relation-
ship, which may be destroyed or damaged by the adversarial nature of dispute
resolution settlement mechanisms, such as litigation. Experts opine that ADR
methods, such as mediation, become very important in non-discharge
employment cases, where the parties want to preserve their relationship with
their employer. 3
Mediation differs from other ADR processes and from litigation in significant
ways, including, but not limited to: (1) the parties maintain control over the
outcome if a settlement is reached; (2) in compromise situations, there is a
diminished sense of win-lose. 4 Furthermore, if the parties do not agree in medi-
ation, then they have the option of submitting their case for administrative
and/or legal consideration. 5
Mediation has been found effective in dealing with a number of different
types of disputes. It is effective not only in settling small claims, 6 but also in
resolving a variety of different disputes including labor grievances, 7 contractual
terms, construction damages, personal injuries, damages to property and the
environment, 8 and employment discrimination under the Title VII of the Civil
Rights Act. 9 There are a number of commonly held beliefs regarding the
perceived advantages to using mediation rather than the courts. Many
advocates of mediation state that it is often less costly, less time consuming,
can be less damaging to the emotional state of disputants, less damaging to the
relationship of the concerned parties or can even serve to improve them, and
by its very nature promotes a win-win solution.
The largest workplace mediation program is the Equal Employment
Opportunity Commission's (EEOC) voluntary mediation program, a detailed
description of which is presented in the background section of the paper. During
the first twelve months of the expanded mediation program (4/1/99 through
3/31/00), the EEOC conducted over 11,700 mediations; more than 7,500 charges
were resolved through the program. This early success leads into two next-step
questions. The first is does the mediation program help the EEOC system
address charges more effectively and secondly, if it does, will it become an
Has the E E O C Hit a Home Run ? 3
accepted institutionalized process? The ultimate acceptability of a dispute

resolution system has been described as one of the "ultimate tests" of the
system's viability. 1° The present research goes toward answering these next-
step questions.
The purpose of this research paper is to present the results of a comprehen-
sive survey of the participants (charging parties and respondents) in the EEOC
mediation program regarding their experience with and opinions of the program.
More specifically, this paper presents the analysis of the participants' opinions
regarding the procedural elements of the mediation, the performance of the
mediator, and participant satisfaction with the results of the mediation.
This paper is organized into several sections. After the introductory section,
we present a brief background of the EEOC, its strategies under different
administrations, and its mediation program. The third section of the paper
introduces our research and its premises. Next, we present a brief review of
the literature in the field of ADR program evaluation. The fifth section presents
the research methodology of our study. The sixth section details the results
of the study. The paper concludes with the discussion section, which focuses
on the implications of the results.
II. B A C K G R O U N D
The Equal Employment Opportunity Commission (EEOC) was established as

the federal agency responsible for the administration and enforcement of Title
VII of the Civil Rights Act of 1964 ("Title VII" or "the Act"), which prohibits
workplace discrimination based on race, sex, national origin, color, or religion.I l
It investigates, conciliates, and litigates complaints of discrimination arising
under the laws within its jurisdiction through its 50 field offices located
throughout the United States. Central to the agency's mission is the develop-
ment and implementation of charge resolution programs and processes.
Originally, the EEOC had no power to enforce its findings. Title VII
indicates that Congress intended the EEOC to quickly investigate allega-
tions of discrimination and to then act as a conciliator between the parties
to bring about a voluntary resolution of a dispute. 12 In 1972, the Equal
Employment Opportunity Act amended Title VII to allow the EEOC to file
suit in federal court to enforce Title VII. This amendment also expanded the
EEOC jurisdiction to include public and private educational institutions
and state and local government. This law continued to rely primarily on the
conciliation process to resolve charges that the agency investigated and found
to be meritorious. ~3
4 E. PATRICKMcDERMOTTET AL.
When Title VII was passed in 1964, Congress expected that the EEOC could
investigate and then immediately conciliate meritorious cases to a voluntary
resolution; this proved to be an unrealistic expectation. From its inception, the
EEOC had an inventory of roughly 3,000 cases. By the 1970s, it was clear that
the EEOC was overwhelmed with charges. In June 1972, the backlog rose to
53,000 charges. By April 1977, that backlog had risen to 130,000.14 This caused
the EEOC to lose credibility with the parties who appeared before the agency.
It was clear that something had to be done to more effectively and
efficiently process this growing backlog of cases.
A. The EEOC Strategies To Timely Process Charges
This section presents the various strategies that the EEOC has taken over the
last two decades to process the charges in a timely fashion. From Eleanor
Holmes Norton to Ida Castro, the different EEOC chairmen and women have
instituted different policies to ensure that the EEOC address the issue of the
timely processing of cases.
In May 1977, EEOC Chairwoman Eleanor Holmes Norton introduced a "rapid
charge processing" program. The program was designed to expedite the process-
ing of charges by seeking quick settlement prior to full investigation. This pro-
gram relied on no-fault settlement agreements with the agency in return for
payment of some money or other benefit to the charging party. This strategy
resulted in the expeditious processing of many cases. Also, this strategy enabled
the EEOC to expend its limited resources on "pattern and practice" and systemic
litigation. These cases are high profile and often result in the recovery of
large monetary settlements. From a cost - benefit perspective, such litigation
probably results in a larger monetary yield. ~5 To the critics of the program, any
charge, other than pattern and practice claims, was seen as being "for sale," i.e.
subject to quick resolution, by respondent payment of what the critics considered
to be a token sum of money. 16 It also appeared that individual charges were not
given the same attention or importance as the pattern and practice lawsuits.
In 1979, the EEOC's responsibilities were expanded to include enforce-
ment of discrimination charges made pursuant to the Age Discrimination in
Employment Act of 1967,17 the Equal Pay Act of 1963,18 and Section 501 of
the Rehabilitation Act of 1973.19
Even with the significant expansion of the EEOC jurisdiction during her
tenure, Chairwoman Norton's program reduced the case backlog. In FY 1981,
the inventory was 49,500; this was a noteworthy accomplishment given the
EEOC's increasing responsibilities, z° She left the agency in February 1981, and
was later succeeded by Clarence Thomas. 21
Has the EEOC Hit a Home Run? 5
After Clarence Thomas was appointed as the EEOC Chair in May 1982, he
was immediately met with a Government Accounting Office (GAO) report
critical of the EEOC rapid charge-processing program. The major criticism was
that meritorious cases were being settled too quickly and too cheaply while
non-meritorious cases were being settled for too much money. There seemed
to be little relationship between the underlying culpability of the respondent
and the amount of the settlement. 22
Thomas decided it was inappropriate for the EEOC to do little or no
investigation of the "garden variety" discrimination charge, while spending
substantial resources on the headline-grabbing, and more financially lucrative,
pattern and practice lawsuits. He believed that every person who filed a charge
with the EEOC was entitled to a full investigation. Accordingly, Thomas thought
the agency should move away from the claims adjuster image it had under
Norton and toward that of a law enforcement agency. 23
In December 1983, Thomas implemented a full investigation policy. All
charges were to be fully investigated to a determination of "reasonable cause"
to believe that there had been a violation of the law or a "no reasonable cause"
determination. A full investigatory report was written at the end of each
investigation. 24 The use of personnel to fully investigate every charge resulted
in litigation of fewer pattern and practice cases. When Thomas left the EEOC
in 1990, the case inventory was at 41,987. 25
Commissioner Evan J. Kemp, Jr., replaced Thomas as Chairman in March
1990. Also in 1990, President George Bush signed the Americans with
Disabilities Act (ADA) into law. 26 The EEOC responsibilities were further
expanded to include the investigation of claims made under Title I of the
newly enacted ADA. The Civil Rights Act of 1991 (CRA) further expanded
the 1964 Act. 27 The CRA made it more attractive to sue an employer for a
violation of Title VII. Under this new law, a victim of employment discrim-
ination was entitled to recover compensatory and punitive damages. 28
Previously these monetary damages were not available and a successful
litigant could obtain only injunctive relief backpay (less interim earnings),
reinstatement or front-pay in lieu of reinstatement, and respondent payment
of the employee's attorney fees and costs. Now, in addition to these remedies,
a successful litigant could obtain compensatory and punitive damages. Filing
of a charge with the EEOC continued to be a prerequisite to initiating a
lawsuit in federal court. Thus, the EEOC again experienced a significant
increase in the number of charge filings.
In addition to the rise in charges related to the ADA and the CRA of 1991,
national attention to sexual harassment, focused by the Clarence Thomas -
Anita Hill matter, resulted in a rise in charges of workplace sexual
harassment. In 1992, 72,302 cases were filed, the largest case intake since
1985. 29 The EEOC continued to adhere to the full case investigation approach
developed under Chairman Thomas. As a result, the volume of unresolved
charges as well as the time that it took to resolve a case continued to increase.
In 1994, the time from filing of a charge to resolution averaged 328 days. 3°
By October 1994, the EEOC had a new chair, Gilbert Casellas, who joined
the EEOC from the Air Force, where he had served as general counsel. Casellas
inherited an agency that could not move its cases in a timely fashion. By the
end of FY 1994, there was a backlog of 96,945 unresolved cases. 31
Chairman Casellas was faced with the challenge of reducing the backlog. He
established internal task forces to address the case-handling dilemma. The
resulting taskforce reports provided the information necessary to create a new
process for handling incoming charges. They were classified according to
"Priority Charge Handling Procedures" (PCHP), which divided charges into
three categories: (1) "A" charges are those that, based on the opinion of the
EEOC charge intake personnel, 32 were highly likely to result in an EEOC finding
of "reasonable cause" to believe that the law had been violated, and involved
pattern and practice or systemic issues, or were Equal Pay Act charges; (2) "B"
charges were those that have possible merit with the final finding contingent
on the results of the investigation; and (3) "C" charges were those without merit
on their face that should be dismissed outright. 33
Concurrent with this new classification procedure, in April 1995, the EEOC
voted to commit the agency to mediation as a voluntary alternative to normal
charge processing. Thus, mediation became the forum for alternative dispute
resolution (ADR) that was integrated into the EEOC charge processing
procedure. 34 Commissioners Paul Miller and R. Gaull Silberman headed the
ADR taskforce that made the recommendations. Mediation was selected largely
because of the success of a pilot program that was conducted in four field
offices in 1991-1992. 35
In addition to the PCHP, the Casellas administration introduced a National
Enforcement Plan (NEP), which was fully implemented by Fiscal Year (FY)
1997. The NEP employed a three-prong approach to address the agency's
mission. The prongs were: (1) the prevention of discrimination through enhanced
education, technical assistance, and outreach to the employer community,
charging party advocacy groups, and other stakeholders; (2) the eradication of
discrimination through investigation, conciliation, and litigation of charges with
significant impact; and (3) effective caseload and inventory management,
including effective use of alternative dispute resolution methods, to allow
the Commission to focus substantial resources on those matters having the
greatest impact. 36
Has the EEOC Hit a Home Run ? 7
The Priority Charge Handling Procedures allowed the EEOC to take great
strides in reducing its charge backlog. In FY 1995, immediately prior to
introduction of the PCHP, the agency had a backlog of 111,345 charges. At the
end of FY 1997, the inventory was reduced to 64,576 charges - a reduction of
o v e r 4 0 % . 37 By FY 1998, this backlog had been reduced to 52,011 - a 53%
reduction from the pre-PCHP backlog. The PCHP and the NEP of the Casellas
administration resulted in the reduction of average charge resolution time, a
reduction in the average age of the pending charge backlog, a reduction in the
average caseload of the EEOC investigator, and in other qualitative improve-
ments. 38
On October 23, 1998, Ida L. Castro was sworn in as the Chairwoman of
the EEOC. During her first year of tenure, the EEOC achieved many accom-
plishments, including cutting the backlog of private sector charges by 23% to
a 15-year low, reducing the average charge processing time, and reforming the
federal sector EEO complaint process. ''39
The agency nearly tripled the number of successful charge resolutions handled
through voluntary mediations to 4,833 at year-end FY 1999, up from 1,631 in
FY 1998. 40 Similarly, the program gained greater acceptance among its target
audience; FY 1999 saw an increase in the number of participants in the program.
Eighty-one percent of charging parties (up from 68% the year before) and 36%
of employers (up from 28%) accepted offers to mediate. The settlement success
rate for mediated cases was 65%, significantly higher than the 50% success rate
anticipated by the agency. The EEOC also met a key mediation goal by
mediating cases in an average of 87 days. a]
B. The E E O C Mediation Program
Beginning February 1991, the EEOC introduced an ADR pilot mediation

program as a strategy to manage its burgeoning caseload. The one-year pilot
tested the use of mediation in the Philadelphia, New Orleans, Houston, and
Washington field offices of the EEOC. 42 Statistical information was gathered
regarding the successes and deficiencies in the program. Professor Craig A.
McEwen of Bowdoin College evaluated the pilot program. He found a high
level of participant satisfaction with the pilot program. He concluded that
mediation resolved a significant number of charges in a fair manner. 43
The success of the pilot program resulted in the appointment of a mediation
task force to explore the possibility of expanding the use of mediation. In 1995,
the task force, under the direction of Commissioners Miller and Silberman,
concluded that mediation was a viable process. The EEOC adopted a policy
statement concerning ADR that included support for mediation. 44 Included in
the policy statement were certain core principles that governed the ADR
program administered by the EEOC. These principles are that the program
"must further the agency's mission to eliminate and remedy employment
discrimination" and it must be fair to the participants. Voluntariness, neutrality,
confidentiality, and enforceability were identified as the requirements to achieve
this fairness. 45
The EEOC uses the A-B-C classification system as the basis for selection of
cases that were deemed appropriate for mediation. Charges that are identified
as "A" charges usually are not selected for mediation. These charges involve
cases where a reasonable case finding is highly likely or where important pattern
or practice/systemic issues or other public policy concerns militate against the
use of pre-investigation mediation. Where a party requests that an "A" charge
be mediated, the District Director and Regional Director have the discretion to
allow such mediation. However, this is the exception and not the rule for such
cases.
Charges classified as "B" charges are charges where further investigation is
required to make a determination concerning their merit. In general "B" cases
are eligible for pre-investigation mediation. However, "B" cases that involve
the Equal Pay Act or pattern or practice/systemic allegations are not eligible
for pre-investigation mediation. Charges classified as "C" are considered
inappropriate for mediation since they are deemed unlikely to result in a
probable cause determination. 46
Intake personnel usually make the initial classification decision. Each office
has a process that reviews and ensures the accuracy of its classification
decisions. Equal Pay Act cases are not eligible for pre-investigation mediation.
In general, at the time a charge is filed, the charging party is advised that
voluntary mediation is available and is asked whether or not he/she is willing
to participate. Once the intake officer classifies the charge as a "B," and it is
confirmed that the charging party has agreed that he/she is willing to partici-
pate in mediation, the respondent is sent the charge along with a letter that
offers the respondent the opportunity to participate in mediation.
Since mediation is a pre-investigation dispute resolution procedure, the incen-
tive for the respondent to participate is that, in addition to all of the inherent
advantages of mediation, the respondent can postpone preparing a position
statement and/or responding to an EEOC information request. 47 Where media-
tion is not successful, the investigation process is reinstated and the respondent
is asked to submit a position statement and otherwise cooperate with the
investigation. If the respondent elects to mediate, this process is kept separate
from the pending, and now suspended, investigation. Thus, a "firewall" exists
that ensures that any confidential information obtained in the mediation process
is not disclosed to EEOC investigators. Investigators are instructed to cease any

line of inquiry with a charging party, if the party begins to discuss what occurred
in mediation.
The EEOC requires that the mediation process be completed within 100 days
from the time a charge is received in the office. The EEOC mediation program
uses EEOC staff employees as mediators as well as external mediators who are
either paid pursuant to an EEOC contract or serve on a p r o bono basis. The
EEOC provides extensive training to all mediators.
III. THE RESEARCH CATALYST
The EEOC decided in 1999 that it needed a comprehensive evaluation of its

mediation program based on the end users' perceptions of the process, the
mediator and the outcome. On August 18, 1999, Drs. McDermott, Jose, Obar,
and Bowers were retained by the EEOC to design, administer and analyze a
"customer satisfaction survey" of the EEOC mediation participants (EEOC
Order No. 9/0900/7632/2).
The purpose of the study was to examine how the participants of the program
- charging parties and respondents - evaluated the EEOC mediation program.
The Participant Satisfaction Survey, the instrument that was used to collect the
data, incorporated several research questions based on a comprehensive review
of the literature in the field. The main research questions are the following:
(1) What was the participant feedback regarding the mediation? More specifi-
cally, how did the participants view the procedural and distributive elements of
the mediation program? (2) Does participant feedback vary depending upon
variables such as mediator type, representation, mediation status, and media-
tion result? (3) Would the parties use the program again?
IV. LITERATURE REVIEW
A review of the various studies that evaluated the effectiveness of ADR

programs indicates that any evaluation of a dispute resolution method should
include participant feedback regarding their perceptions of procedural justice
(justice with regard to the process) and distributive justice (justice with regard
to the outcome).
Procedural justice can be defined as "the perceived fairness of the process
through which decisions are made. ''4~ The theory of procedural justice posits
"disputants prefer procedures that provide them with voice, control over the
outcome, and fair treatment by the third party. ''49 Thus, parties to a dispute
must first be given a fair chance to voice their concerns. Second, parties must
have control over the outcome of mediation since mediation is about
self-determination. Third, the mediator must be perceived as, and be, fair and
neutral. Measurement of participant perception of procedural justice regarding
mediation must consider whether the participants felt that they were treated
with fairness by the mediator, their voice was heard, and they had control over
the outcome.
Fairness is an essential element of procedural justice. It is especially crucial
to any mediation program because it is a necessary ingredient to induce
participation and settlement. 5° Hodges notes that the EEOC has recognized not
only the importance of actual and perceived fairness in its mediation program,
but also that "fairness requires adequate information, the opportunity for
assistance, knowing and voluntary participation, neutrality, confidentiality, and
enforceability." The EEOC ADR programs were developed according to these
principles of fairness. 51
As indicated above, an element of procedural justice is "knowing partici-
pation," i.e. participant understanding of the process. Research in organizational
theory has shown that understanding is an important factor in employee
attitudes towards organizational activities. 52 Hence, understanding of the
process is essential for participant satisfaction with the process and for their
perception that the process was fair. Hodges states "in order to make an
informed choice about whether to participate in mediation, the potential parties
need information about the mediation process, statutory rights and remedies,
and the advantages and disadvantages of both mediation and litigation." She
also points out that both the EEOC and Administrative Conference of the
United States (ACUS) recognize that ADR programs must be fair to the
participants, and fairness requires that extensive information be provided to
the participants about the process. It should be noted that one of the
recommendations of the Task Force on Alternative Dispute Resolution in
Employment, which established the "Due Process Protocol for Mediation and
Arbitration of Statutory Disputes Arising Out of the Employment
Relationship," is that participants be "given access to information about both
the process and the substance of the dispute. ''53
The "timing of mediation" and "representation" are two other important
variables that affect procedural justice. Prompt scheduling of a program is
considered to be an indication of effective program management. The timing
of mediation is important since one of the touted advantages of mediation is
that it is less time consuming than other methods of dispute resolution. Once
the case is referred to mediation and if it takes place promptly (i.e. before
positions harden), a settlement may be more likely. ''54
Fairness/procedural justice requires that there is opportunity for assistance

to the participants. Representation in the form of either attorneys or other
knowledgeable persons, for one or both parties, may serve to balance powerY
It is important not only to let the participants know that they can bring in
representatives, but also to notify each party as to who will be representing
the other party. 56 There is empirical evidence to support the notion that
when one party to a dispute appeared with an unanticipated representative,
the other party was concerned about the fairness of the process. 57
The mediator is a crucial assistant to the process of problem solving based
decision making. As the third party who assists the parties to a dispute, a mediator
facilitates communication, helps them to focus on the real issues of the dispute,
and assists them in generating options for settlement. 58 Mediators, unlike
arbitrators, do not mandate solutions to a dispute. Instead, they are facilitators
who enable parties to reach their own agreement. 59 As facilitators, it is essential
that the mediators understand the nature of the dispute in question and the
viewpoints of the parties to the dispute. This understanding becomes critical in
clarifying the issues to the parties, which is important to promote cooperative
problem solving. Mediators also serve as "a reality check, alerting parties to
unrealistic expectations. ''6°
One of the commonly included criteria in the definitions of mediation is
that the mediator be neutral or impartial. 61 Some researchers point out that
"neutrality, as traditionally practiced, actually includes two potentially conflict-
ing qualities: (1) impartiality, which refers to the mediator's ability to maintain
an unbiased relationship with the disputants; and (2) equidistance, which involves
the mediator temporarily becoming aligned with each party to encourage
disclosure and assist the party in expressing the case. ''62 Mediators must be
impartial, fair, and diligent in order to foster trust between the parties and
between the mediator and the parties. They must also maintain the confidentiality
of the parties. 63
Participant perception of distributive justice or justice with regard to the
outcome is another important element of participant satisfaction with the
program. It is not surprising that participants would perceive the outcome to be
more just when they received what they wanted than when they did not.
However, even when the mediation does not end in a settlement, it has the chance
to further the communication between parties since it brings the parties together.
Distributive justice measures include participant satisfaction with the outcome
of mediation and participant perceptions regarding the fairness of the outcome. 64
Research has shown that participants not only differentiate between procedural
justice and distributive justice, but also consider procedural justice to be more
important in certain situations. 65
Gordon, in his study of the issues of justice in a union-management grievance

procedure, emphasized the importance of procedural justice in workplace dispute
resolution. 66 In his survey of employees' overall satisfaction with the grievance
system, he found that the overall evaluation of the grievance system was
significantly correlated with the measures of perceived procedural and
distributive justice. He also found that the correlation was significantly higher
for the procedural justice factors as opposed to the distributive justice factors.
Thus, his research seemed to support the findings of Folger and Greenberg 67
that the "procedures followed, rather than the outcomes obtained, have the
greater influence on the overall evaluation of dispute-resolution systems. ''68
In conclusion, the literature review shows that the measurement of
procedural and distributive justice is essential to the evaluation of any
mediation program. Elements of procedural justice include variables such as
"knowing participation," representation, performance of the mediator, fairness
of the process, and opportunity to present participant views. Elements of
distributive justice include participant satisfaction with the outcomes and the
fairness of the outcomes.
V. RESEARCH M E T H O D O L O G Y
The primary method of data collection is a survey of the participants, who

were the charging parties (employees or ex-employees) and the respondents
(employers and ex-employers). We also use the internal documents from the
EEOC to collect information regarding the characteristics of the charges.
The participant feedback survey instrument is composed of a 22-item inven-
tory that includes 14 questions that employ a five-point Likert-type response
scale ranging from strongly disagree (1) to strongly agree (5). A Likert scale
is used because there is evidence that it is superior to other measurement formats
in measuring attitudes, and its underlying factorial structure is more stable across
situations and cultures. 69 Of the remaining eight questions, five require a yes
or no response, two allow for open-ended responses, and one incorporates
a multiple-choice set of responses. Care is taken to ensure that we measured
the EEOC's goals (where applicable) with regard to the fairness of its media-
tion program, i.e. to provide adequate information about the process, the
opportunity for assistance, knowing and voluntary participation, neutrality,
confidentiality, and enforceability. The survey was pilot-tested and modifica-
tions were made in order to improve its construct validity and when compared
to these results we are able to ascertain the survey's external validity or
reliability. 7°
The target population includes all the participants in the EEOC mediation
process conducted under the supervision of the 50 EEOC field offices from
approximately March 1 to July 3 l, 2000. Participants received the surveys at
the end o f their mediation sessions and were asked to complete them, place
them in an envelope, and seal them. Mediators then forwarded these surveys
to the local A D R coordinator along with the other required documentation
about the mediation. The local A D R coordinators were asked to mail the
surveys once a week directly to the research team. The mediators or field
office A D R coordinators were also asked to indicate when participants did
not fill out the survey. This was done to allow the researchers to measure the
non-response bias. A comprehensive protocol was developed to ensure that
data collection was done not only efficiently, but also in a manner to assure
client confidentiality.
The response rate for these field offices for charging parties was at least
46.25%. The authors received a total of 2,209 surveys completed by the charging
parties. Of these, 526 responses were not used in this study (while some were
not filled out or only partially completed, most rejected surveys either did not
indicate the case charge number or did not have a matching case number in
the EEOC database) resulting in a total of 1,683 usable surveys. Thus, the
effective participation rate was at least 35%.
The response rate for respondents was at least 50%. Of the 2,402 surveys
received from the respondents, only 1,572 were in a usable state resulting in
an effective participation rate of 33% by the respondents. The combined
effective participation rate of the total sample was approximately 34%.
The reason that we state our effective participation and response rate as "at
least" figures is that the EEOC district offices did not distribute the survey to
participants exactly on March 1. While we derived our figures by looking at
all the EEOC mediations held in all 50 field offices between March 1 and July
31, it is clear that not all mediations were surveyed due to this staggered rollout.
Also, we have not included those surveys (for mediations held by July 31)
which we received after our cutoff date. A review of these responses that were
not used indicates that they are, in all ways, similar to the results that we report
below.
Another factor that reduced our response rate was the need for cross-
tabulations. A number of surveys without charge numbers were excluded from
the sample since this preyented us from obtaining the information needed to
perform cross-tabulations. Finally, where parties at the site of the mediation
declined to fill out surveys, a "non-response log" was kept. Reviews of these
non-response logs indicate that there is no probative evidence that the reasons
for non-response would affect our reported results.
The data analyses used here are primarily summary statistics, such as frequen-
cies, percentages, and averages. The primary focus is to set forth the large scale
categories of participant responses. The appropriateness of this approach comes
from the nature of the study, which measured participant satisfaction with the
mediation program.
VI. RESULTS
There were 1,683 completed surveys from the charging parties and 1,572
completed surveys from the respondents. These numbers include only properly
completed questionnaires. As discussed earlier, in cases where the protocol and
the instructions were not strictly followed, the questionnaires were excluded
from the final sample.
A. Profile of the Sample
What are the characteristics of the sample? How representative is this sample
when compared to the cases mediated by the EEOC? To answer these
questions, we present a profile of our sample and a comparison of the profile
of cases mediated by the EEOC during the March 1-July 31 period. Information
was gathered about the company size of the respondents, the applicable statutes
(Title VII, ADA, ADEA), bases of the charge (religion, gender, national origin,
race, disability, age), issues (discharge, terms/conditions of employment, harass-
ment, promotion, sexual harassment, wages, discipline, and reasonable
accommodation), the mediator type (external and internal), the presence or use
of a legal or non-legal representative, and the status of the mediation at the
time the questionnaire was completed.
The majority of the participants were employed in companies with 15 to 100
employees (40.7% of charging parties, 41.7% of respondents) and companies
with more than 500 employees (30.2% and 30.3%, respectively). Title VII is
the statute that is applicable to the majority of charges (71.2% of charging
parties and 71.7% of respondents). The bases of the majority of charges were
race and gender, followed by disability, age, national origin, and religion. As
indicated in Table 1, the majority of the charges were filed under the issues of
discharge and terms and conditions of employment. It should be noted that
some parties filed a charge under more than one statute, basis, or issue.
For both groups in our sample, the vast majority (78%) of cases were
mediated by internal mediators. There is a marked difference in the use of
representation by the two parties. Respondents were more likely to have a
representative going into the mediation. Only 41% of the charging parties were
Table 1. Comparison of the Characteristics Between Our Sample and the

EEOC Database.
EEOC Mediation Cases Research Sample

(March 1-July 31) Charging Parties Respondents
Company Size
15 to 100 employees 40.2% 40.7% 41.7%
101 to 200 employees 13.7% 14.9% 14.1%
201 to 500 employees 11.3% 12% 11.3%
More than 500 employees 32% 30.2% 30.3%
Statute
Title VII 73.8% 71.2% 71.7%
Age 19.2% 21.4% 21.1%
Disability 22.3 % 23.5 % 24.1%
Basis
Religion 2.4% 2.5% 2.5%
Gender 33.3% 30.4% 30.7%
National Origin 8.5% 8.9% 9.6%
Race 35.2% 37.9% 37.7%
Disability 23.1% 23.3% 23.8%
Age 19.3% 21.2% 20.8%
Issue
Discharge 46.0% 48.6% 49.6%
Terms and Conditions 20.9% 19.0% 18.4%
Harassment 19.1% 17.1% 16.8 %
Sexual Harassment 12.1% 10.4% 10.6%
Promotion 10.4% 10.1% 10.0%
Wages 9.1% 9.2% 8.7%
Discipline 8.9% 9.2% 9.0%
Reasonable Accommodation 8.8% 8.6% 8.7%
Mediator Type
External 74% 78.4% 78.4%
Internal 25% 20.5% 20.6%
Representation
Yes n/a 41.3% 58.5%
No n/a 57.7% 39.8%
Mediation Status
Mediation is completed n/a 73.8% 81.0%
Mediation is on-going n/a 13.8% 13.3%
The percentages do not add up to 100 because some participants did not furnish the required
information.
represented, whereas 58% of the respondents were represented. For the majority
of participants (74% of charging parties and 81% of the respondents), their
mediation was completed at the time of the survey.
In summary, the discussion above demonstrates the very strong similarity
between the sample and population group with regard to the type of mediator
used, the company size, the statute the cases were filed under, the basis of the
charges, and the specific issue under consideration. These similarities demon-
strate that the results of this study are generalizable to the population.
B. Procedural Elements and EEOC Mediation
Four statements were used to measure the participants' satisfaction with the
mediation process. Of these, the first two were "pre-mediation session"
or "mediation preparation" statements regarding whether the participants
received an adequate explanation from an EEOC representative and whether
the session was scheduled promptly. The next two statements asked whether
the participants understood the process and had an opportunity to present
their views.
a. Explanation, Scheduling, and Voice

An analysis of the participant responses regarding their "pre-mediation" or
mediation preparation experiences shows that the participants had a positive
experience. Eighty-seven percent of the participants agreed with the state-
ment "prior to my attendance at this mediation session today, I received an
adequate explanation about mediation from an EEOC representative." The
mean score for the statement was 4.23. As indicated in Table 2, data also
show that both charging parties and respondents were in agreement regarding
the statement. Eighty-eight percent of the charging parties agreed with the
statement and 85% of the respondents agreed with the statement. The mean
score for the charging parties was 4.24 and the respondents was 4.21,
indicating that there is very little difference between the two groups regarding
their attitudes on this issue. Eighty-nine percent of the participants agreed
with the statement "the mediation was scheduled promptly." The mean score
was 4.34, indicating that there was strong agreement about the prompt sched-
uling of mediation. Ninety-one percent of the respondents and 88% of the
charging parties agreed with the statement. The mean score was 4.39 for the
respondent group and 4.28 for the charging party group. Analysis reveals
that the difference in the mean scores is significant, indicating that the respon-
dents were more satisfied than the charging parties with the prompt
scheduling of mediation.
Has the E E O C Hit a Home Run? 17
The authors believe that the reason for the difference in the mean scores is
significant is that for the charging parties, many of whom have had an adverse
personnel action taken against them, one simply can never get a hearing,
mediation, or other administrative action quickly enough. In general, a respondent
does not have the same pressing need for immediate action as the s t a t u s q u o is
acceptable to the respondent because the respondent determined that s t a t u s q u o
and not the charging party.
Participants also agreed with the statements regarding the mediation session.
Ninety-four percent of the participants agreed with the statement "after the
mediator's introduction, I felt that I understood the mediation process." The
mean score was 4.44, indicating that participants strongly agreed that they
understood the process. A comparison of the two groups shows that 96% of
the respondents and 92% of the charging parties agreed with the statement.
Analysis reveals that the responses of the participant groups varied significantly
and that the respondents (mean score of 4.53) agreed more strongly than the
charging parties (mean score of 4.35) that they understood the process.
While the overall results indicate that both parties understood the process,
the authors posit that the reason for the significant mean scores difference is
that in general the respondent may enter the mediation process with an
advantage in understanding the mediation process. This may be because the
respondent has usually spent some time considering whether, as the charged
party, mediation is appropriate. Also the respondent is often, but not always,
more experienced in dispute resolution and litigation.
Participants also felt that they had a voice during the process. The vast
majority of the participants agreed with the statement "I (or my representative)
had a full opportunity to present my views during the mediation process."
Ninety-five percent of the respondents and 90% of the charging parties agreed
with this statement, indicating that the opportunity to present one's views - one
of the essential elements of procedural justice - was present in the EEOC
mediation process. The average mean scores reveal that the respondents (4.57)
felt more strongly than the charging parties (4.39) that they had an opportunity
to present their views.
The authors are not sure why the difference in mean scores is significant.
Since the charging party is the "moving party" in such proceedings it is
interesting that significantly more did not feel as though they had a full
opportunity to present their views. A review of responses to an open-ended
survey question allowing for broad comments on the process did not provide
any insight into this result. One possible explanation is that the views that some
charging parties sought to present were not conducive to resolution of the dispute
and may have been deflected by or otherwise de-emphasized by the mediator.
18 E. PATRICK McDERMOTT ET AL.
Table 2. Participant Satisfaction with the Explanation, Scheduling, and

Voice.
Charging Parties Respondents

Mean Mean
Statements (n,%) (n,%)
Prior to my attendance at this mediation session today,

I received an adequate explanation from the EEOC 4.24 4.21
representative. (1637, 88.3%) (1516, 85.4%)
The mediation was scheduled promptly 4.28 4.39
(1673, 87.7%) (1559, 90.5%)
After the mediator's introduction, I felt that I 4.35 4.53
understood the mediation process (1676, 91.6%) (1561, 95.8%)
I (or my representative) had a full opportunity 4.39 4.57
to present my views during the mediation process. (1677, 89.8%) (1563, 95.0%)
Notes: Satisfaction is measured by the mean responses of the participants on a Likert scale (scale
of 1 [strongly disagree] to 5 [strongly agree]) and by the percentage of participants who agreed or
strongly agreed with the statements. The sample size (n) is also given for evaluation purposes.
Figures in bold refer to statements where a statistically significant difference (evaluated at 95%
confidence level) exists between the mean responses of the charging parties and respondents.
For example, if the view that the charging party sought to present was that
his boss was a "lying s.o.b." it makes sense that the mediation process may
have discouraged such venting or other adversarial conduct. Another possible
explanation is that, as one will see at page 23 infra, 26% fewer charging parties
stated that they obtained what they wanted at mediation. Finally, more
respondents were represented than charging parties. It may be that represented
parties are more likely to report that they had a full opportunity to present
their views.
b. The Mediator
The second set of questions focused on statements regarding the mediator's
performance. More specifically, participants were asked whether the mediator
understood their needs, helped to clarify their needs, remained neutral in the
beginning as well as throughout the process, helped to develop options for the
resolution o f their claim, and used procedures that were fair to them.
The majority of the participants felt that the mediator understood their needs
(87%) and helped to clarify their needs (82%). Eighty-six percent of the 1,669
charging parties and 87% of the 1,552 respondents agreed that the mediator
understood their needs. The mean scores of the charging parties (4.30) and the
respondents (4.31) indicate that the parties' responses were identical with regard
to the statement "the mediator understood my needs" the participants also felt
that the mediator helped clarify their needs. Eighty-four percent of the charging
parties and 79% of the respondents expressed either strong or very strong agree-
ment with the above statement. Although the agreement was strong among
participants in general, it was stronger among charging parties (mean of 4.25)
than among respondents (mean of 4.17).
An overwhelming majority of the participants agreed with the statements
regarding the neutrality of the mediator. Of the 1,674 charging parties who
responded to the statement "at the beginning of the mediation, I considered the
mediator to be neutral," 1,542 (92%) agreed with the statement. Similarly, 1,441
of the 1,566 respondents (92%) who answered indicated strong agreement with
the statement. The mean scores of 4.44 for the charging parties and 4.49 for
the respondents indicate the similar strength of the attitudes of participants.
Ninety-one percent of the charging parties and 89% of the respondents agreed
"the mediator remained neutral during the session." An analysis of the
mean scores of the participants (4.42 for the charging parties and 4.43 for the
respondents) shows that the participant groups felt almost exactly as strongly
that the mediator remained neutral during the session.
Participants also felt that the mediator played a very useful role in the
development of options for the resolution of the charge. 7~ Eighty-five percent
of the charging parties and 84% of the respondents agreed with the statement
"the mediator helped the parties develop options for resolving the charge." The
nearly identical mean scores of the participants (4.27 for the charging parties
and 4.23 for the respondents) attest to the fact that the attitudes of both partic-
ipant groups were almost exactly the same.
Participants agreed that the procedures used by the mediator were fair to
them. Of the 1,668 charging parties who responded to the statement "the
procedures used by the mediator in the mediation were fair to me," 1,476 (89%)
expressed agreement with the fairness of the mediation procedures. Similarly,
of the 1,564 respondents who answered, 1,436 (92%) expressed agreement.
Analysis of the mean scores reveals that the respondents (4.44) agreed more
strongly than the charging parties (4.33) regarding the fairness of the
procedures used by the mediator.
Overall, as shown in Table 3, the participants were very satisfied with the
role and conduct of the mediator. There were some differences between the
two participant groups regarding their perceptions of mediators. A higher
percentage of charging parties (84%) than respondents (79%) agreed that the
mediator helped to clarify their needs. This result suggests that the charging
parties may have entered the mediation with a greater need for assistance
from a third party to help them understand what they wanted. This is also
20 E. PATRICKMcDERMOTT ET AL.
Table 3. Participants' Satisfaction with the Mediator.

Mean Mean
The mediator understood my needs. 4.30 4.31

(1669, 86.4%) (1552, 86.9%)
The mediator helped clarify my needs. 4.25 4.17
(1665, 84.4%) (1504, 79.3%)
At the beginning of the mediation, 4.44 4.49
I considered the mediator to be neutral. (1674, 92.1%) (1566, 89.1%)
The mediator remained neutral throughout 4.42 4.43
the session. (1664, 90.7%) (1564, 89.1%)
The mediator helped the parties develop 4.27 4.23
options for resolving the charge. (1661, 85.1%) (1545, 83.9%)
The procedures used by the mediator 4.33 4.44
were fair to me. (1668, 88.5%) (1564, 91.8%)
Notes: Satisfactionis measured by the mean responses of the participants on a Likert scale (scale
of 1 [stronglydisagree] to 5 [stronglyagree]) and by the percentage of participants who agreed or
Figures in bold refer to statements where a statistically significantdifference (evaluated at 95%
consistent with the theory that, overall, respondents have a better idea of what
they want to get from mediation at the time that they agree to the mediation.
Since respondents can decline to go to mediation, and since for all practical
purposes the charging parties are strongly encouraged by the EEOC to give
mediation a chance, there is a greater self-selection process among respondents
who go to mediation.
Regarding the participant attitudes concerning fairness of the procedures used
by the mediator, 92% of the respondents agreed the procedures were fair,
whereas 89% of the charging parties agreed. This is an interesting result that
begs for further research. As posited earlier in this paper, it is possible that
some charging parties, in pressing their discrimination claims, may have
preferred procedures that allowed them to "argue" their claim but were not
necessarily conducive to resolution by mediation. For example, the use of an
immediate caucus that separates the parties for the majority of the mediation
may be an effective dispute resolution tool but may leave parties, particularly
charging parties, feeling as though they did not get their full say and therefore
the procedure was not perceived as fair.
As illustrated in this section on procedural elements, both the charging parties

and respondents were very satisfied with the fairness of the process. The mean
scores of both participant groups were over 4.00 on a 5-point scale, indicating
their strong agreement with the process statements. Thus, participants indicated
a high level of satisfaction with the EEOC mediation process.
C. D i s t r i b u t i v e E l e m e n t s a n d M e d i a t i o n
Participant satisfaction with the distributive elements of mediation was measured

using four questions concerning the results. More specifically, three questions
were asked about participant attitudes regarding the realistic nature of the options
developed during the mediation, their satisfaction with the fairness of mediation,
and their satisfaction with the results. The fourth distributive question was a
"yes/no" question concerning whether the participants obtained what they wanted
from the mediation. With the exception of this question, all other distributive and
procedural questions discussed above were measured using the above-referenced
Likert scale.
The first statement concerning the mediation outcome was "most of the options
developed during the mediation session were realistic solutions to resolving
the charge." The participants responded very similarly to this statement: 76%
of the respondents and 75% of the charging parties expressed agreement with
the statement. A larger percentage of participants (16% of respondents and 13%
of charging parties) remained neutral regarding this statement as opposed to
the other mediation process statements. The mean scores of 4.00 for the
respondents and 3.95 for the charging parties indicate that they agreed upon
the realistic nature of the options developed during the session and that
there was very little difference in the views of the participant groups regarding
this issue.
Eighty-seven percent of the respondents and 79% of the charging parties
agreed with the statement "I was satisfied with the fairness of the mediation
session. ''72 An analysis of the mean scores of the two groups indicates that
respondents agreed more strongly (4.31) with the statement than the charging
parties (4.07).
While the vast majority of the participants (83%) were satisfied with
the fairness of the mediation session, their satisfaction with the results (59%) of
the mediation was more tempered. Of the 1,547 charging parties who responded
to the statement "I was satisfied with the results of mediation," 55% agreed with
the statement, whereas of the 1,477 respondents, 63% agreed. Thus, the respon-
dents were more satisfied with the results of the mediation than the charging
parties. The mean scores reveal that respondents agreed more strongly (mean of
3.67) than the charging parties (3.38) with the statement. Table 4 presents these
findings. This result should be interpreted with some caution since our sample
includes participants from on-going mediations.
One possible explanation for the significant difference in the parties' response
regarding satisfaction with the results o f the mediation and the fairness of the
mediation m a y be found in the next set of results. More charging parties than
respondents reported that they did not get what they wanted from the mediation.
The literature to date has supported the conclusion that where a party does not
get what he or she wants the party is more likely to state that the process was
not fair or that he or she was not satisfied with the results.
The survey also sought to measure whether the participants obtained what
they wanted from mediation. This is a strong distributive measure. The
participants were asked whether they knew, before going into mediation, what
they wanted from mediation. If they stated that they did, then they were also
asked whether they obtained what they wanted. 73 O f the 79% o f the charging
parties who indicated that they knew what they wanted, 74 only 41% stated that
they obtained what they wanted. O f the 83% of the respondents who knew what
they wanted going into the mediation, 57% stated that they obtained what they
wanted.
Table 4. Participant Satisfaction with the Distributive Elements o f

Mediation.

Mean Mean
Most of the options developed during the

mediation session were realistic solutions to 3.92 4.00
resolving the charge. (1648, 75.2%) (1519, 75.6%)
I was satisfied with the fairness of the 4.07 4.31
mediation session. (1648, 78.9%) (1559, 86.9%)
I was satisfied with the results of the mediation. 3.38 3.67
(1547, 54.8%) (1477, 62.6%)
Mini - Win
Another distributive justice question targeted those participants whose disputes did
not get resolved during mediation. As indicated earlier, 26% of the charging
parties and 19% of the respondents did not resolve their claims. They were
asked whether progress was made in mediation toward the resolution of their claim.
Progress, if made, is a mini-win for the parties since it brings them closer together
toward the resolution of the charge. As shown in Table 5, of the 488 charging
parties who responded, 29% agreed that progress was made. Similarly, 28% of the
486 respondents agreed with the statement. The mean scores of 2.60 for the
charging parties and 2.72 for the respondents indicate that the respondents agreed
more strongly than the charging parties that progress was made in mediation.
In summary, the analysis of participant responses shows that the participants
were satisfied with both procedural and distributive elements of mediation. The
participants were more satisfied with the procedural elements than with the
distributive elements. Regarding the procedural elements, participants expressed
the highest degree of satisfaction with the neutrality of the mediator, the oppor-
tunity to present their views, and their understanding of the process. Conceming
the distributive elements, participants were more satisfied with the fairness of
mediation and with the realistic nature of the options developed during the
mediation than with the results of mediation.
D. The Influence o f Other Variables on Participant Satisfaction
We analyzed participant responses to see whether they varied based on medi-

ator type, representation, mediation status, and mediation result. The analysis
is based on within-group comparisons.
a. Type o f Mediator
The EEOC mediation used both internal mediators (EEOC staff) and external
mediators for mediation. We found only three significant differences (see
the bolded numbers in Table 6) in the responses of the charging parties and
Table 5. Responses of Participants Whose Claims Were Not Resolved in

Mediation Regarding Whether Progress was Made Toward Resolution.
Participant Total Mean Strongly Neither Agree/Strongly

Group Responses Rating Disagree/ Agree nor Agree
Disagree
Charging Parties 488 2.60 48.6% 22.1% 29.3%

Respondents 486 2.72 41.6% 30.9% 27.6%
Table 6. Participant Satisfaction Based on Mediator Type.
Statements Charging Parties Respondents

Mean (n,%) Mean (n,%)
Internal External Internal External
Procedural Elements
Explanation, Scheduling and Voice
Adequate Explanation 4.26 4.20 4.23 4.14

(1277, 88.4%) (343, 88.6%) (1189,86.7%) (311, 81.4%)
Prompt Scheduling 4.32 4.16 4.43 4.27
(1311, 88.6%) (346, 85.0%) (1226, 91.8%) (317, 85.5%)
Understood the process 4.37 4.31 4.54 4.50
(1314, 89.6%) (345, 91.3%) (1228,96.0%) (317, 95.0%)
Opportunity to present 4.39 4.37 4.57 4.57
views (1314, 89.6%) (346, 90.8%) (1227,94.9%) (320, 95.9%)
Mediator
Mediator understood 4.32 4.23 4.31 4.30

needs (1309, 86.6%) (344, 85.5%) (1221,87.1%) (315, 86.3%)
Mediator helped 4.28 4.18 4.17 4.16
clarify needs (1309, 85.1%) (339, 82.0%) (1182,79.8%) (307, 77.5%)
Mediator neutral in 4.45 4.41 4.47 4.56
the beginning (1313, 92.2%) (344, 92.2%) (1229,91.5%) (321, 94.4%)
Mediator remained 4.44 4.36 4.43 4.45
neutral (1305, 90.9%) (342, 90.6%) (1227,89.0%) (321, 90.3%)
develop options (1303, 85.0%) (341, 86.5%) (1215,83.9%) (314, 83.8%)
Mediator used 4.34 4.33 4.44 4.45
fair procedures (1309, 88.5%) (342, 88.9%) (1229,91.9%) (319, 92.2%)
Distributive Elements
Development of 3.99 3.83 3.99 4.02

realistic options (1295, 76.5%) (336, 71.1%) (1194,75.5%) (309, 75.5%)
Satisfaction with the 4.10 3.99 4.30 4.34
fairness of the session (1294,79.9%) (337, 75.4%) (1227,87.1%) (316, 85.4%)
Satisfaction with 3.39 3.37 3.68 3.62
the results (1213, 55.2%) (318, 53.8%) (1156,63.6%) (307, 59.0%)
confidence level) exists between the mean responses of the participants in the cases mediated by
external mediators and those mediated by internal mediators.
respondents based on mediator type. Both charging parties and respondents who
had an internal mediator expressed greater agreement about the prompt
scheduling of the mediation than those who had an external mediator. This data
suggests that there may be something in the mediation scheduling process that
favors the internal as opposed to the external mediators. According to Steve
Ichniowski, EEOC National ADR Coordinator, this is probably due to the fact
that the scheduling of external mediations can take more time because
documents have to be mailed, forms filled out and mediation space obtained
by the mediator. For internal mediations, Ichniowski indicated that the case file
can often be walked over to the mediator who can immediately contact the
parties and schedule them to meet at an available EEOC office space dedicated
to mediation.
Charging parties also rated intemal mediators higher than external mediators
regarding the realistic development of options. This data suggests that the exper-
tise of internal EEOC mediators, with their specialized dispute resolution
knowledge under the law of Title VII, results in the development of more
realistic options.
It should be noted that the comparison of mean scores from the six
statements regarding the performance of the mediators reveals that charging
parties were slightly more satisfied with the performance of the internal
mediator. Respondents on the other hand, though satisfied with both types of
mediators, gave external mediators a slightly higher score on neutrality. This
is especially true regarding their attitudes concerning the neutrality of the
mediators in the beginning (mean of 4.56 for external mediators and 4.47 for
internal mediators). However, as the mediation progressed, their attitudes
changed. This can be seen from the fact that on the question concerning whether
the mediator remained neutral during the session, the mean score was 4.45 for
external mediators and 4.43 for intemal mediators. It appears that respondents
were suspicious of the EEOC until exposed to the mediation program, at which
point trust appears to have been earned.
b. R e p r e s e n t a t i o n
During the mediation some participants were represented, whereas others were
not. Representation did not mean just legal counsel; representatives included
any individual who was present as an advocate. Our analysis of the mean scores
of the participant groups revealed several significant differences based on
representation. Among charging parties, those without representation were more
in agreement than those with representation regarding the mediator's role in the
clarification of needs and in the development of options. Similarly, charging
parties who were without representation agreed more strongly than those with
26 E. PATRICKMcDERMOTt ET AL.
representation that most of the options developed during the session were
realistic solutions to the resolution of charges. They were also more satisfied
with the results of the mediation.
Respondents differed significantly on eight statements concerning mediation
based on representation. Those who were without representation agreed more
strongly than those who were represented that they received an adequate
explanation of the process and that their sessions were scheduled promptly.
They also rated the mediator higher with regard to the mediator's understanding
of their needs, clarification of issues, assistance in the development of options,
and neutrality in the beginning. They were also more satisfied with the devel-
opment of realistic options and with the fairness of the process.
In general, it appears that participants who were without representation found
mediators to be more helpful than those who were represented. They were also
more satisfied with the outcomes. Overall, the results indicate that where
respondents are represented, the respondents are more likely to report that the
mediator did not perform as well in a broad range of areas. These results allow
for some speculation - Do representatives in general, and respondents' counsel
in particular, tend to criticize the mediator or the mediation process to their
clients, thus causing these results? Do representatives educate clients on process
or mediator flaws, thus causing a legitimate decline in clients' overall positive
opinions regarding the mediation process? This is an interesting area as some
mediation programs are designed with the view that attorney or other
representative participation is a negative factor and thus limit or exclude
such participation in the mediation. Table 7 shows the results based on
representation.
c. Charge Status
The attitudes of the participants towards mediation differed, significantly
at times, based on their mediation status. As explained before, the participants
in the study belong to one of three different groups based on the status of
their mediation session: (1) mediation is completed, and the charge has been
resolved in mediation; (2) mediation is completed, but the charge has not been
resolved and the parties will not continue the mediation; and (3) mediation is
ongoing.
As one would expect, the perceptions of the participants were affected by
the resolution status of their mediation. Table 8 indicates that satisfaction of
the charging parties with both the procedural and distributive elements varied
significantly based on mediation status. Charging parties who belong to the
second group ("mediation finished, charges not resolved") consistently rated all
the statements regarding mediation lower than the other two groups. Members
Table 7. Participant Satisfaction Based on Representation.

With Without With Without
Representation Representation Representation Representation
Procedural Elements
Explanation, Scheduling, and Voice
Adequate explanation 4.20 4.28 4.12 4.33

(669, 85.7%) (952, 90.1%) (877, 82.8%) (612, 88.7%)
Prompt scheduling 4.26 4.30 4.34 4.46
(664, 87.3%) (963, 88.1%) (909, 89.2%) (623, 92.3%)
(692, 91.6%) (968, 91.5%) (913, 95.0%) (621, 96.9%)
Opportunity to present 4.39 4.38 4.55 4.60
views (693, 88.9%) (968, 90.5%) (914, 94.3%) (623, 96.0%)
Mediator
Mediator understood 4.26 4.33 4.25 4.39

(690, 83.9%) (963, 88.2%) (906, 85.2%) (620, 89.4%)
Mediator helped clarify 4.18 4.30 4.11 4.25
needs (687, 80.6%) (962, 87.1%) (873, 77.8%) (606, 81.5%)
Mediator neutral in 4.44 4.43 4.44 4.55
the beginning (693, 91.8%) (965, 92.3%) (918, 91%) (621, 93.6%)
Mediator remained 4.41 4.42 4.40 4.48
neutral (690, 90.3%) (958, 91.0%) (916, 88.1%) (621, 90.3%)
develop options (687, 81.4%) (958, 87.8%) (900, 81.2%) (619, 87.9%)
Mediator used fair 4.34 4.32 4.41 4.48
procedures (691, 88.4%) (961, 88.6%) (916, 91.3%) (621, 92.6%)
Development of 3.86 4.01 3.95 4.07

realistic options (674, 70.9%) (958, 78.2%) (882, 74.8%) (612, 76.8%)
fairness of the session (685, 79.6%) (951, 78.2%) (917, 85.7%) (619, 88.5%)
results (630, 49.0%) (903, 58.5%) (858, 60.4%) (593, 65.9%)
Notes: Satisfaction is measured by the "mean responses" of the participants on a Likert scale (scale
confidence level) exists between the mean responses of the participants with representation and
without representation.
Table 8. C h a r g i n g P a r t i e s ' S a t i s f a c t i o n B a s e d o n M e d i a t i o n Status.
Statements Completed and Completed Ongoing

resolved not resolved
(Group 1) (Group II) (Group III)
Mean (n,%) Mean (n,%) Mean (n,%)
Procedural Elements
Adequate explanation 4.31 4.12 4.22

(915, 90.3%) (301, 85.0%) (227, 85.5%)
Prompt scheduling 4.36 4.16 4.31
(931, 89.4%) (304, 85.5%) (231, 88.3%)
Understood the process 4.43 4.22 4.35
(935, 93.5%) (303, 88.4%) (231, 90.9%)
Opportunity to present views 4.49 4.21 4.38
(932, 92.6%) (304, 84.5%) (232, 88.4%)
Mediator
Mediator understood needs 4.42 4.07 4.27

(931, 90.8%) (301, 78.1%) (232, 84.5%)
Mediator helped clarify 4.41 4.00 4.14
needs (927, 89.8%) (302, 75.2%) (228, 78.1%)
Mediator neutral in the 4.50 4.32 4.50
beginning (932, 93.7%) (304, 89.8%) (232, 93.1%)
Mediator remained neutral 4.50 4.25 4.45
(927, 92.7%) (303, 85.8%) (231, 91.3%)
Mediator helped develop 4.45 3.94 4.14
options (928, 91.9%) (299, 70.2%) (229, 81.2%)
Mediator used fair 4.41 4.20 4.37
procedures (927, 90.5%) (304, 85.9%) (232, 87.9%)
Development of realistic 4.23 3.47 3.68

options (930, 85.9%) (291, 55.3%) (226, 61.5%)
Satisfaction with the fairness 4.20 3.88 4.03
of the session (925, 83.9%) (300, 69.3%) (231, 76.2%)
Satisfaction with the results 3.87 2.24 2.99
(879, 70.5%) (287, 18.1%) (203, 37.9%)
confidence level) exists among the mean responses of the different groups.
Table 9. R e s p o n d e n t s ' S a t i s f a c t i o n B a s e d o n M e d i a t i o n Status.
Statements Completed and Completed Ongoing

resolved not resolved
(Group I) (Group II) (Group III)
Mean (n,%) Mean (n,%) Mean (n,%)
Procedural Elements
Adequate explanation 4.20 4.28 4.21

(922, 84.3%) (308, 87.7%) (203, 87.7%)
Prompt scheduling 4.41 4.41 4.33
(952, 90.8%) (314, 91.4%) (206, 89.3%)
Understood the process 4.54 4.54 4.48
(952, 95.9%) (314, 96.2%) (207, 94.2%)
Opportunity to present views 4.59 4.55 4.53
(957, 95.3%) (312, 95.5%) (207, 94.7%)
Mediator
Mediator understood needs 4.35 4.25 4.25

(950, 88.6%) (311, 83.3%) (204, 86.3%)
Mediator helped clarify needs 4.24 4.06 4.05
(920, 81.7%) (300,73.7%) (198, 76.3%)
Mediator neutral in the beginning 4.47 4.55 4.44
(956, 91.1%) (314, 94.6%) (208, 91.8%)
Mediator remained neutral 4.45 4.42 4.41
(956, 89.1%) (13, 89.1%) (207, 89.9%)
Mediator helped develop options 4.37 3.94 4.01
(947, 89.3%) (304, 71.4%) (206, 77.7%)
Mediator used fair procedures 4.46 4.44 4.38
(957, 92.2%) (314, 91.1%) (206, 92.7%)
Development of realistic options 4.26 3.43 3.73

(947, 87.1%) (292, 50.7%) (199, 63.3%)
Satisfaction with the fairness 4.37 4.21 4.24
of the session (956, 88.4%) (312, 82.4%) (208, 87.0%)
Satisfaction with the results 4.15 2.55 3.20
(923, 81.0%) (292, 19.9%) (187, 44.4%)
Figures in bold refer to statements where a statistically significant difference evaluated at 95%
confidence level) exists among the mean responses of the different groups.
of the first group had the most positive feedback. For these questions, a clear
continuum was shown for charging parties with the unresolved cases having
the lowest positive results, the ongoing cases landing in a middle range, and
the resolved cases registering the most positive results. As shown in Table 9,
the pattern becomes clear for the respondents only when it comes to the
distributive elements of mediation.
The data also show that the charging parties, regardless of their mediation
status, were able to differentiate between the procedural and distributive
elements of mediation. For example, analysis of the mean scores of the second
group reveals that the scores for all the procedural statements were either above
4 or close to 4 indicating participant satisfaction with the process. However,
their scores for the distributive elements were lower.
To summarize this section, participants' satisfaction with the distributive
elements varied with their mediation status; so did their perception of the role
of mediator in the clarification of needs and development of options. Among
the participants, the responses of charging parties varied more dramatically
based on the status of mediation. The bottom line is that where the dispute was
resolved, the ratings were higher.
d. Mediation Result
Participant responses were analyzed based on their satisfaction with mediation
results. For charging parties and respondents, the results on every question varied
significantly based on their satisfaction with the mediation results. As shown in
Table 10, participants who were satisfied with the results of mediation agreed
more strongly on the different procedural and distributive questions that were
asked than participants who were not satisfied with the results.
To summarize this section, participants rated the various elements of the
EEOC mediation program highly. In general, they gave higher marks to the
procedural aspects of the mediation program than to the distributive aspects.
Among the distributive aspects, participant satisfaction was high regarding the
fairness of the mediation. While participant responses sometimes varied
significantly based on the different variables discussed above, it should be
noted that their mean scores were almost always above four points (on a
five-point scale), indicating their satisfaction with the various elements of the
mediation.
E. Would the Parties Use the Program Again ?
As the literature review section indicated, a test of acceptability of a program

is the willingness of its participants to use the program again. The parties were
Has the E E O C Hit a Home Run ? 31
Table 10. Participant Satisfaction Based on Their Satisfaction with the

Mediation Result.

Satisfied Not Satisfied Satisfied Not Satisfied
Procedural Elements
Adequate explanation 4.43 3.94 4.31 3.94

(826, 93.5%) (395, 79.2%) (886, 88.6%) (253, 77.9%)
Prompt scheduling 4.47 4.01 4.49 4.08
(843, 92.2%) (398, 80.7%) (917, 92.6%) (259, 81.9%)
(848, 96.5%) (397, 83.1%) (918, 98.1%) (260, 89.2%)
Opportunity to present views 4.60 4.05 4.70 4.20
(845, 95.7%) (398, 80.7%) (923, 98.2%) (258, 84.9%)
Mediator
Mediator understood needs 4.58 3.89 4.50 3.84

(840, 95.4%) (397, 73.0%) (915, 94.3%) (257, 67.7%)
Mediator helped clarify needs 4.57 3.79 4.38 3.71
(838, 95.0%) (396, 69.2%) (881, 87.2%) (252, 59.9%)
Mediator neutral in the 4.61 4.16 4.62 4.15
beginning (845, 96.2%) (398, 84.7%) (922, 95.3%) (259, 82.2%)
Mediator remained neutral 4.64 4.09 4.61 4.01
(839, 97.1%) (396, 81.6%) (922, 94.6%) (258, 76.7%)
Mediator helped develop 4.58 3.75 4.48 3.59
options (841, 95.6%) (394, 65.2%) (911, 92.2%) (254, 60.6%)
Mediator used fair procedures 4.57 3.99 4.62 4.02
(839, 95.8%) (398, 77.9%) (922, 97.4%) (260, 78.8%)
Development of realistic 4.44 3.13 4.38 3.07

options (842, 93.5%) (389, 45.5%) (909, 90.8%) (246, 39.0%)
Satisfaction with the fairness 4.48 3.45 4.59 3.65
of the session (838, 95.1%) (390, 55.6%) (920, 97.2%) (258, 62%)
of 1 [strongly disagree] to 5 ]strongly agree]) and by the percentage of participants who agreed or
confidence level) exists between the mean responses of the participants who were satisfied with
the results and those who were not satisfied with the results of the mediation.
32 E. PATRICKMcDERMOTT ET AL.
asked whether, if they were party to a charge before the EEOC in the
future, they would be willing to participate again in the mediation program
(i.e. "willingness to return"). Ninety-one percent of the charging parties and
96% of the respondents indicated that they would be willing to use the program
again. This is a highly significant finding that demonstrates from the
participants' perspective the utility of mediation to effectively resolve EEOC
cases. This finding also lends support to the policy recommendation that the
program be supported and perhaps expanded.
One could argue that the ultimate test of a system is the willingness of the
parties who did not obtain what they wanted to use the system again. As
indicated in Table 11, regardless of whether the participants obtained what they
wanted from the mediation or not, they overwhelmingly indicated that they
were willing to participate in the program again (if the need arises). This can
be viewed as a very strong indication of their positive experiences with the
EEOC mediation program.
Table 11. Participants' Willingness to Participate in the EEOC Mediation

Program in the Future.
Total Total
Charging Respondents
Parties
For those who knew what they wanted

going into the mediation: Did you obtain
what you wanted going into the mediation? 100.0% 100.0%
Yes 40.8% 56.5%
No 55.5% 40.9%
No response given 3.7% 2.5%
For those who responded YES to the question
above: If you were a party to a charge before
the EEOC in the future, would you be willing
to participate in the EEOC's mediation program? 100.0% 100.0%
Yes 95.2% 97.5%
No 2.2% 0.7%
For those who responded NO to the question
above: If you were a party to a charge before
the EEOC in the future, would you be willing
to participate in the EEOC's mediation program? 100.0% 100.0%
Yes 88.9% 93.2%
No 9.8% 4.0%
Has the E E O C Hit a Home Run? 33
VII. CONCLUSIONS AND IMPLICATIONS
The participant evaluation of the EEOC mediation program shows a high degree
of participant satisfaction with the EEOC mediation program and its elements.
Both the participant groups - charging parties and respondents - gave high
marks to the various elements of the EEOC mediation program. The conclu-
sions of this research and their implications are the following:
The participants expressed strong satisfaction with the information they

received about mediation from the EEOC prior to their attendance at the
mediation session. Pre-mediation preparation and process education help
parties gain a sense of what is to come which may reduce the anxiety
associated with uncertainty, unpredictability, and sense of control.
Participants also feel very strongly that they understood the process after
the mediator's introduction. This guidance and education helps parties as
they move through the process. One of the EEOC goals of mediation is to
provide adequate information about mediation to the parties. The results
show that the EEOC was very successful in fulfilling this goal. As the
results indicate, "knowing participation," an essential element of procedural
justice, is present in the EEOC mediation process.
The vast majority of the participants agreed that their mediation was
scheduled promptly. Prompt scheduling is important for four reasons: (1)
it is an indicator of effective program management; (2) the likelihood of
a settlement increases if mediation takes place promptly before the parties
hardened their positions; (3) the timely attendance to these matters attests
to EEOC's determination to take the resolution of these charges seriously;
and (4) the charging parties' have evidenced a strong desire to have an
immediate hearing of their claim. The EEOC's prompt scheduling of
mediation sessions is indicative of effective program management. It also
increases the chances of dispute resolution in reaching an agreement.
Finally, it sets the EEOC mediation process apart from the formal
litigation process which takes much longer.
An overwhelming majority of the participants felt that they had a full

opportunity to present their views and be heard during mediation. This
"voice factor" is important for five reasons: (1) it provides the parties with
the satisfaction that their story is heard. Research on grievance systems
indicates that one of the main opportunities that participants seek is a forum
to present their views and tell their story; (2) it is crucial to the fairness
of the process, as indicated by the theory of procedural justice; (3) it

increases people's acceptance of the outcome, thereby contributing to their
overall satisfaction with the outcome. By taking part in the decision making,
parties take ownership rather than "buy-in" to the outcome; (4) there is a
psychological benefit for the parties to have the opportunity to have a fair
chance to express their issues, concerns, and needs in their own language
without having to resort to translating them into some legal argument or
other procedural language; (5) in some instances the mere process of venting
one's emotions can have a cathartic effect on the parties.
The participants were very satisfied with the role and conduct of the
mediators. They felt strongly that the mediators understood their needs,
helped to clarify their needs, and assisted them to develop options for
resolving the charge. They felt even more strongly that the procedures used
by the mediators were fair. The questions regarding the neutrality of the
mediators elicited some of the strongest responses from the participants,
who felt that the mediators were neutral not only in the beginning of the
process, but also remained neutral throughout the process. This finding is
significant for two reasons: (1) the perceived neutrality of the mediator is
important to foster trust between the parties and the mediator, which is
essential for the resolution of the charge and for the participant satisfac-
tion with the process; (2) one of the EEOC goals of mediation is neutrality.
As the participant responses indicate, the EEOC was successful in achieving
this goal.
Participant satisfaction with the distributive elements of mediation was more

tempered than their satisfaction with the procedural elements. This is indica-
tive of the fact that mediation is a facilitated negotiation process, that is
couched in a problem solving format, where parties do not usually obtain
what they wanted going into the negotiations. This result is also consistent
with the dispute resolution literature on distributive justice. Among those
elements classified as distributive, the participants were most satisfied with
the fairness of the mediation session. They also agreed that most of the
options developed during mediation were realistic solutions to resolving
the charge. The majority of the participants were also satisfied with the
results of mediation.
Participant satisfaction with the EEOC mediation program remained high

even when the participant responses differed, at times, based on whether
a party to mediation was represented, the type of mediator, the status of
the mediation, and satisfaction with the result. Our overall results indicated
that the EEOC participant feedback reported here was fairly consistent,
regardless of the influence of the various factors discussed above. The
only obvious exception was based on "participant satisfaction with the
mediation result." As one would expect, participants who were satisfied
with the result of their mediation gave higher ratings to their mediation
experiences than those who were not satisfied. However, it should be noted
that even in this case, the results indicate that the participants who were
not satisfied with the result of mediation expressed positive opinions
regarding the various procedural elements of mediation.
An overwhelming majority of the participants indicated that they would be

willing to participate in the mediation program again if they were a party
to an EEOC charge. Participants, regardless of their satisfaction with the
outcome of mediation, overwhelmingly indicated their willingness to return
to mediation. This is a strong indication of their satisfaction with the EEOC
mediation program. This finding also indicates that participants were able
to differentiate between the procedural and distributive elements of medi-
ation. The fact that willingness to return was high, even among participants
who did not receive what they wanted, indicates that a fair and neutral
process that provides participants with an opportunity to present their views
may be even more important than the obtained outcome. This finding is
consistent with the results of other empirical studies of procedural justice.
The satisfaction of the parties with the process, regardless of the outcome,
indicates that mediation is attractive for a number of reasons. Some are
presented here but it is certain that there are other reasons not identified
in this study; this is an area for further research. Another area of research
is based on the notion that "process is related to outcome" and the more
opportunities that exist for face-to-face constructive problem solving
dialogue the more likely, we theorize, the rate of agreement may rise. This
theory needs to be tested by examining the case stream analysis experi-
ences of EEOC personnel.
The results of this study so far indicate that mediation is indeed a process
that is well fit to manage the issues that typically arise in many EEOC
cases. One of the common ADR intake formulas is to determine case "fit"
by "matching the issue with the right process" and it appears that EEOC
is doing a good job. Seeking ways to improve the case stream analysis and
channeling may improve the rate of agreement.
W e have a single policy recommendation. The results of this study, which

is the most comprehensive of the EEOC mediation process to date, strongly
suggests that the EEOC, based on participant feedback, continue to closely
examine ways to make the program more attractive to parties and media-
tors. W e also suggest that EEOC explore ways to make the program
permanent given the strong results provided here that lend credible evidence
that the process is well received. Mediation should be made an institutional
alternative and not a supplement to the normal processes. This suggestion
is not just for cases deemed amenable to mediation but to other more
"challenging" cases. If facilitated negotiation is a key to the mediation
process acceptance and success then the EEOC should also ascertain parties'
disposition to negotiate as one means of channeling cases to mediation.
Expanding the program to include more complex cases, and types o f cases,
will allow A D R researchers and practitioners the opportunity to determine
just how successful mediation can be at the EEOC. The EEOC may discover
that the impressive results o f the present program can be achieved for a
broader class o f charges filed with the EEOC.
NOTES
1. Center for Dispute Resolution, Mediation (1989); Administrative Conf. U.S.,
"Implementing the ADR Act: Guidance for Agency Dispute Resolution Specialists" 5
(February 1992).
2. Dannin, E. J. "Contracting Mediation: The Impact of Different Statutory Regimes,"
Hofstra Labor and Employment Law Journal, (Fall 1999).
3. Hodges, A. C. "Mediation and the Americans With Disabilities Act," Georgia Law
Review, (Winter 1996).
4. Ettingoff, C. C. and Powell, G. "Use of Altemative Dispute Resolution in
Employment-Related Disputes," University of Memphis Law Review, (Spring 1996).
5. Id.
6. McEwen, C., "Note on Mediation Research." In: S. B. Goldberg, F. E. A. Sander
& N. H. Rogers (Eds), Dispute Resolution: Negotiation, Mediation, & Other Processes,
(2nd ed.). (1992), Boston: Little Brown. Moore, Christopher W. (1986). The Mediation
Process: Practical Strategies for Resolving Conflict. San Francisco: Jossey Bass.
7. Brett, J. M., & Goldberg, S. B. "Grievance Mediation in the Coal Industry: A Field
Experiment," Industrial and Labor Relations Review, 37, (1983), 4 9 ~ 9 .
8. Brett, J. M. Barsness, Z. I., & Goldberg, S. B. "The Effectiveness of Mediation:
An Independent Analysis of Cases Handled by Four Major Service Providers,"
Negotiation Journal, (July 1996), 259-269; Thibaut, J., & Walker, L. Procedural Justice:
A Psychological Analysis, (1975), Hillsdale, N. J.: Lawrence Earlbaum Associates.
Williams, Bruce A., & Albert R. Matheny. (1995). Democracy, Dialogue, and
Environmental Disputes: The Contested Language of Social Regulation. New Haven: Yale
University Press. Crowfoot, James E. and Julia M. Wondolleck. (1990) Environmental
Disputes: Community Involvement in Conflict Resolution. Washington, D.C.: Island Press.

Maser, Chris. (1996). Resolving Environmental Conflict: Towards Sustainable Community
Development. Del Ray Beach FL: St. Lucie Press
9. Dunlop, J. T., & Zack, A. M. Protocols for Employment Dispute Resolution:
Mediation and Arbitration of Employment Disputes, (1997), San Francisco: Jossey-Bass.
10. Kochan, T., Lautsch, B., & Bendersky, C. "An Evaluation of the Massachusetts
Commission Against Discrimination Alternative Dispute Resolution Program," Harvard
Negotiation Law Review, 233 (Spring 2000).
11. Pub. L. 88-352, July 2, 1964; 42 U.S.C. Section 2000(e)-2(a).
12. Pub. L. 88-352, July 2, 1964; 42 U.S.C. Section 2000(e)-4(g) and 5.
13. Where an agency investigation finds a charge to have merit, the agency issues a
Notice of Reasonable Cause that Title VII has been violated. Conciliation is offered after
this determination has been made.
14. Lundberg, K. "Reducing the Complaints Backlog at the Equal Employment
Opportunity Commission", Case C16-00-1562.0, Kennedy School of Government Case
Program, Harvard College, (Spring 2000), p. 1, citing Priority Charge Handling Task
Force/Litigation Task Force Report, EEOC, March 1998, p. 3.
15. Recent examples of EEOC pattern and practice cases include the Texaco and
Mitsubishi cases that were settled for large sums of money.
16. Most respondents are employers but this group also includes trade unions, govern-
ment entities, and employment agencies.
17. Pub. L. 90-202, Dec. 15 1967; 29 U.S.C. Sections 621-624.
18. Pub. L. 88-38, June 10, 1963; 29 U.S.C. Section 206.
19. Pub. L. 93-112, Sept. 26, 1973; 29 U.S.C. Section 705, 791 et. seq; this law is the
precursor to the Americans With Disabilities Act.
Opportunity Commission", Case C16-00-1562.0, Kennedy School of Government Case
Program, Harvard College, (Spring 2000), p. 3.
21. Due to statutory and political reasons beyond the scope of this paper, when one
Chairperson leaves, although an "acting" chair is named, there is usually a significant
period of time before a permanent successor assumes office.
22. Priority Charge Handling Task Force/Litigation Task Force Report, EEOC,
March 1998, p. 5.
Opportunity Commission," Case C16-00-1562.0, Kennedy School of Government Case
Program, Harvard College, (Spring 2000), p. 4.
This contemplated model was similar to the way that the National Labor Relations Board
operated at that time. Some of Thomas' top management had prior NLRB experience.
24. Id. Again, this final investigative report was used by the NLRB for all cases.
25. Id., p. 5.
26. Pub. L. 101-336, July 26, 1990; 42 U.S.C. 12101-12117, 12201-12213.
27. Pub. L. 102-166, Nov. 21, 1991, this law added amendments to various existing
civil rights statutes.
28. Compensatory damages are for pain and suffering. Punitive damages are awarded
to punish egregious defendant conduct.
Opportunity Commission," Case C16~00-1562.0, Kennedy School of Government Case
Program, Harvard College, (Spring 2000), 9.
38 E. PATRICKMcDERMOTF ET AL.
30. Id., p. 10
31. Id.
32. Every office has some system in place to review charge prioritization systems.
33. Many of these C cases involve situations where the EEOC does not
have jurisdiction. The classification of these A-B-C cases was for internal purposes
only; the specific designation assigned to a charge was kept confidential by the
EEOC. It should be noted that this is a general description of the charge classification
process.
34. The EEOC defines mediation as "a fair and efficient process to help you to resolve
your employment disputes and reach an agreement. A neutral mediator assists you in
reaching a voluntary, negotiated agreement." EEOC Internet Homepage, "Mediation,"
http://www.eeoc.gov/mediate/index.html (June 18, 2000).
35. McEwen, C. An Evaluation of the Equal Employment Opportunity Commission's
Pilot Mediation Program. Bowdoin College (1994).
36. EEOC Office of Field Programs, FY 1997, Year-End Report (Washington, D.C.).
37. EEOC Office of Field Programs, FY 1998, Year-End Report (Washington, D.C.).
38. Id.
39. EEOC Internet Homepage, "Ida Castro: Chairwoman," http://www.eeoc.
gov/castro.html
40. EEOC, Comprehensive Enforcement Program Highlights.
41. Id.
42. EEOC Internet Homepage, "History Of EEOC Mediation Program,"
http://www.eeoc.gov/mediate/history.html (June 18, 2000).
43. EEOC Purchase Order No. 910900176321G, Statement of Work, p. 4
(8/18/99).
44. EEOC Internet Homepage, "History Of EEOC Mediation Program."
45. EEOC. Chapter One: Mediation Policies and Procedures in the Mediation
Deskbook.
46. Id.
47. EEOC Internet Homepage, "History Of EEOC Mediation Program,"
http://www.eeoc.gov/mediate/history.html (June 18, 2000). "Facts About Mediation."
According to the EEOC homepage these advantages include saving time and money in
an efficient process, the benefit of a neutral third party, confidentiality, settlement agree-
ments obtained in mediation are not an admission of guilt, and mediation avoids lengthy
and unnecessary litigation.
48. M. Gordon, "Grievance Systems and Workplace Justice: Tests of Behavioral
Propositions About Procedural and Distributive Justice." In: B. D. Dennis (Ed.),
Proceedings of the Fortieth Annual Meeting of the Industrial Relations Research
Association held in Chicago 28-30, December 1987 (pp. 390-397). Madison, Wisconsin:
Industrial Relations Research Association, 1988.
49. Brett, J. M., Barsness, Z. I., & Goldberg, S. B. "The Effectiveness of Mediation:
An Independent Analysis of Cases Handled by Four Major Service Providers,"
Negotiation Journal, (July 1996), 259-269; Thibaut, J., & Walker, L. Procedural Justice:
A Psychological Analysis, (1975), Hillsdale, N. J.: Lawrence Earlbaum Associates.
51. Id.
52. Dulebohn, J., & Martocchio, J. J. "Employee Perceptions of the Fairness of Work
Group Incentive Plans," Journal of Management, 24(4), (1998), 469-489.
54. Id.
55. Id.
56. Id.
57. Id.
58. Kovach, K. K. Mediation: Principles and Practice, (1994).
59. Weckstein, D. T. "In Praise of Party Empowerment - And Of Mediator Activism,"
Willamette Law Review, (Summer 1997).
61. Id.
62. Weckstein (1997).
63. Ettingoff, C. C. and Powell, G. "Use of Alternative Dispute Resolution in
Employment-Related Disputes," University of Memphis Law Review, (Spring 1996).
64. Please refer to the next section of this literature review for a comprehensive
discussion.
65. Folger, R., & Greenberg, J. "Procedural Justice: An Interpretive Analysis of
Personnel Systems." In: K. M. Rowland & G. R. Ferris (Eds), Research in Personnel and
Human Resources Management (Vol. 3, pp. 141-183). Greenwich, Connecticut and
London, England: JAI Press, Inc., 1985.
66. Gordon, M. "Grievance Systems and Workplace Justice: Tests of Behavioral
Propositions About Procedural and Distributive Justice," In: B. D. Dennis (Eds),
Proceedings of the Fortieth Annual Meeting of the Industrial Relations Research
Association held in Chicago 28-30, December 1987 (pp. 390-397). Madison, Wisconsin:
Industrial Relations Research Association, 1988.
67. Folger, R., & Greenberg, J. "Procedural Justice: An Interpretive Analysis of
Personnel Systems." In: K. M. Rowland & G. R. Ferris (Eds), Research in Personnel and
Human Resources Management (Vol. 3, pp. 141-183). Greenwich, Connecticut and
London, England: JAI Press Inc., 1985.
68. Id. at p. 394.
69. Ashkanasy, N. M. "Rotter's Internal-External Scale: Confirmatory Factor Analysis
and Correlation with Social Desirability for Altemative Scale Formats," Journal of
Psychology and Social Psychology, 48, (1988), 1328-1341.
70. A copy of the survey is available from the authors upon request.
71. We have classified this as a distributive measure but are mindful that one could
view this question as a hybrid procedural/distributive measure.
72. This concept of getting what one wanted, did not attempt to ascertain whether the
party did not obtain a "target" or "resistance" point and did not attempt to determine
whether or not the settlement was within the range between these two points.
73. The results of the first part of this survey response may give solace to the dispute
resolution professional who has often wondered, in the heat of mediation, whether the
parties even know what they want.
74. Many of the implications are based on the literature review, which was presented
in Section II.
ACKNOWLEDGMENTS
The authors would like to thank Steven P. Ichniowski, National ADR

Coordinator, Equal Employment Opportunity Commission, for his assistance in
explaining the intricacies of the EEOC mediation program and for his feedback
on earlier drafts of this paper.
RESOLVING CONFLICT: TACTICS OF
FEDERAL MEDIATORS
Patrice M. Mareschal
ABSTRACT
This research examines the mediation process in the labor relations context
to identify the determinants of mediators' tactics. First, data collected
from secondary sources, informal networking with dispute resolution
professionals, participant observation of new mediator training sessions,
and qualitative interviews with Federal Mediation and Conciliation Service
(FMCS) mediators were used to develop a written survey. Next, the survey
was distributed to practicing mediators with the FMCS. In brief this
research compares the assumptions underlying the FMCS' training
curriculum with practitioner sentiments concerning good mediator
practice.
Six hypotheses were developed from the preliminary analysis. These
hypotheses were tested using linear regression. Four statistically
significant relationships were found. Three hypotheses were confirmed.
Survey results indicated that mediator tactics tend to cluster into two
groups, "the broad approach" and "the narrow approach." The survey
data suggest that the broad and narrow approaches to mediation are
complements to each other rather than substitutes for one another.
Therefore, I summed the "broad approach" and "narrow approach" scales
to create one measure of mediator tactics, which I named "the bifocal
approach." The following predictors of the bifocal approach were found
Advances in Industrial and Labor Relations, Volume 11, pages 41-68.

Copyright © 2002 by Elsevier Science Ltd.
All rights of reproduction in any form reserved,
ISBN: 0-7623-0853-2
41
42 PATRICE M. MARESCHAL
to be statistically significant: bargaining context, mediator acceptability,

bargaining chips, and unionization rate.
INTRODUCTION
Mediators often claim that like snowflakes, no two mediation situations are
exactly alike. Moreover, no two mediators would deal with the same dispute
in the same way. Practitioners describe mediation as an art with numerous
philosophies and approaches (Kolb, 1983; Kochan & Katz, 1988). According
to this line of argument, mediation is difficult to learn and not well-suited to
scientific study. Mediation is indeed a complex process. Nevertheless, social
scientists from a variety of disciplines have studied mediation. Through both
theoretical and empirical research social scientists have identified some system-
atic pattems in the mediation process (Kochan & Jick, 1978; Camevale & Pruitt,
1992; Wall & Lynn, 1993).
In recent years, mediation has become increasingly popular as a means to
resolve conflict. This phenomenal rise in the popularity of mediation as a form
of alternative dispute resolution (ADR) has created a situation in which
mediation practice and research has outstripped theory building (Wall & Lynn,
1993). One important question that arises out of the recent growth in the
practice of mediation is: Why do mediators do what they do? In the past, a
mediator's choice of strategies and techniques was considered so personal and
so unique to each situation that many argued that mediator effectiveness was
virtually impossible to analyze. Given the explosive increase in the use of
conflict resolution services, the premise that a mediator's performance cannot
be evaluated has become untenable. Indeed, Bellman (1998) indicates that the
tremendous growth in the practice of mediation has created a situation in which
"the field could be described as a mile wide and an inch deep" (p. 206).
Moreover, Bercovitch and Houston (2000) note that insufficient attention has
been devoted to examining how to best mediate or which factors affect the
choice of mediator behavior.
This research examines the mediation process in the labor relations context
to identify the determinants of mediators' tactics. First, data collected from
secondary sources, informal networking with dispute resolution professionals,
participant observation of new mediator training sessions, and qualitative
interviews with Federal Mediation and Conciliation Service (FMCS) mediators
were used to develop a written survey. Next, the survey was distributed to
practicing mediators with the FMCS. In brief, this research compares the
assumptions underlying the FMCS' training curriculum with practitioner
Resolving Conflict: Tactics of Federal Mediators 43
sentiments concerning good mediator practice. The ultimate goal is to develop

a theory that makes sense to both researchers and practitioners.
Mediation in the Labor Relations Context
Mediation is the most commonly used type of third party intervention in

labor disputes and collective bargaining (Noe, Hollenbeck, Gerhart & Wright
2000). In fact, the National Labor Relations Act (NLRA) requires a labor union
to notify the FMCS 30 days before calling a strike. The 30-day notice provides
federal mediators an opportunity to help the parties resolve their dispute
peacefully prior to taking such drastic action as a strike (Kochan & Katz, 1988).
In the labor relations context, mediation has traditionally been used to resolve
disputes arising out of contract negotiations. In recent years, the practice
of using mediation to resolve grievances (i.e. disputes arising out of contract
interpretation) has become more widespread (Feuille, 1999). The FMCS is one of
the largest providers of mediation services in the areas of contract negotiations
and grievance mediation. For example, the FMCS provides preventive mediation
assistance programs to help employers and their unionized employees develop
problem-solving skills that will enable the parties to resolve disputes on their own.
Furthermore, the Administrative Dispute Resolution Act of 1990 and the
Negotiated Rulemaking Act of 1990 authorize the FMCS to provide conflict
resolution services to public agencies in order to improve operations and reduce
the resources expended on litigation (FMCS, n.d.). As a result, mediation is the
fastest growing type of ADR being used by the federal government to resolve
employment disputes (Bingham, Chesmore, Moon & Napoli, 2000).
In addition, a recent survey of Fortune 1000 companies indicates that over
88% of respondents had used mediation at least once in the past three years
(Lipsky & Seeber, 1998). The survey respondents were more satisfied with
mediation than other types of ADR. Moreover, an overwhelming majority of
respondents stated that they will use mediation again in the future (Lipsky &
Seeber, 1998).
The FMCS
The FMCS was created as an independent agency of the U.S. government by

the Labor Management Relations Act of 1947. The agency's mission is to
preserve and promote labor-management peace. Toward this end the agency
provides mediation, arbitration, and other conflict resolution services and
programs to employers and their unionized employees in both the private and
public sectors, excluding the railroad and airline industries (FMCS, 1996). In
addition, Executive Order 11491 authorized the FMCS to assist federal

agencies in resolving negotiation impasses (Kearney & Carnevale, 2001). The
services the FMCS provides are intended to prevent or minimize conflicts in
the collective bargaining process and to improve labor-management relations
(FMCS, 1996). In cases where the FMCS' attempts to resolve federal impasses
are unsuccessful, the Civil Service Reform Act of 1978 (CSRA), section 7119
authorizes the Federal Services Impasse Panel (FSIP) to intervene in the dispute.
The FSIP may use a variety of techniques to resolve federal impasses including
mediation, fact-finding, and final-offer arbitration (Keamey & Carnevale, 2001).
M O D E L I N G THE M E D I A T I O N P R O C E S S IN THE
LABOR RELATIONS CONTEXT
The research on mediation spans several disciplines. Wall and Lynn (1993)
have organized the voluminous and diverse research on mediation into a
conceptual framework. The key components of Wall and Lynn's (1993)
framework include: the parties' interactions, the decision to mediate, the
mediator's techniques and strategies, the mediator's and the parties' outcomes,
and the determinants of these factors. Wall and Lynn (1993) identify the
following factors as directly influencing mediators' choice of tactics: rules
and standards, common ground and concern for parties' outcomes, dispute
characteristics, mediator's training, mediation context, mediator's ideology, and
culture. In addition, in the Wall and Lynn (1993) framework, mediators' choices
of tactics are indirectly influenced by the parties' interactions, the parties'
outcomes, and the mediators' outcomes.
If a relationship exists between techniques and outcomes, the factors that
influence the mediation outcomes should also influence the mediator's choice
of tactics. Elangovan (1998) provides evidence in support of this conjecture.
Previous research on mediation in the labor relations context has identified
the following determinants of effective mediation: the characteristics of the
mediators, the sources or nature of the conflict, the situational characteristics
of the dispute, the mediators' strategies, and the parties to the dispute (Kochan
& Katz, 1988).
In developing a model of the mediation process in the labor relations
context, this research combines aspects of Wall and Lynn's (1993) mediation
framework and Kochan and Katz' (1988) model of the determinants of
successful mediation. Since FMCS mediators are required to mediate labor
disputes upon the request of the parties involved, the "decision to mediate"
stage has been eliminated from this model. Also, since the focus here is on the
Other Tactical ~ Mediator Characteristics

Determinants
/
The
Parties' Tactics Parties
Interactions
Sources/Nature
of Conflict
f Situational
Characteristics
Fig. 1. Determinants of Mediator Tactics in the Labor Relations Context.
determinants of mediation techniques, the outcomes of the dispute are not

included in this model. The modified framework focuses on the mediator's
tactics and the determinants of these factors (e.g. the parties' interactions
prior to mediation, mediators' characteristics, sources or nature of the
conflict/impasse, situational characteristics of the dispute, and the parties
themselves). The model is shown in Fig. 1. The key components of the model
are discussed in the sections that follow.
Mediator Tactics
Mediators may draw on a variety of techniques to help parties reach agreement.

The strategies a mediator may use range from a passive, hands-off approach to
an aggressive approach in which the mediator pushes the parties toward
agreement. One key mediator technique is framing. For example, Bazerman
(1998) suggests that if a mediator wants the parties to compromise, the
mediator should strive to have the parties view the negotiations in a positive
frame. To accomplish this, the mediator needs to emphasize the realistic
risk that both parties face if agreement is not reached. By creating this
uncertainty, the mediator encourages the parties to seek a sure settlement
(Bazerman, 1998).
Additional mediator tactics include: demonstrating empathy, structuring

discussion, and stimulating thinking (Zubek, Pruitt, Pierce, McGillicuddy &
Syna, 1992), using a settlement strategy and using a problem-solving strategy
(Kressel, Frontera, Forlenza, Butler & Fish, 1994), using content-control and
using motivational-control (Ross, 1990), tracking constitutive and regulative
rules (Jorgensen, 2000), reformulating dispute narratives (Phillips, 1999),
compensating and pressuring (Harris & Carnevale, 1990), face-saving (Downie,
1991), caucusing (Welton, Pruitt & McGillicuddy, 1988), compromise
suggestions (Conlon, Carnevale & Ross, 1994), facilitative tactics (Karambayya,
Brett & Lytle, 1992), and active neutrality (Solstad, 1999). Often, mediators
use a variety of strategies during the course of a mediation (Kochan &
Katz, 1988).
Determinants of Mediators' Tactics
The variety of techniques at a mediator's disposal is nearly limitless. In fact,

Wall and Lynn (1993) note that approximately 100 techniques for resolving
disputes have been identified by researchers. Similarly, the FMCS' new
mediator training manual refers to "The 7,000 Habits of Effective Mediators"
(FMCS, 1997). Thus, a key question that arises is: Why do mediators do what
they do? In other words, what influences their choice of tactics?
Parties' Interactions
According to Wall and Lynn's (1993) framework, the mediation process begins
with the parties' interactions. In the collective bargaining context the parties do
not interact in a social vacuum. Rather, they interact in a complex social-
institutional environment.
Tetlock (1991) has emphasized the need to consider the complex social-
institutional environment in which people make decisions. In particular, Tetlock
(1991) argues that in making decisions, social actors are influenced by the fact
that they can be held accountable for their actions. Tetlock (1991) also contends
that social actors seek the approval and respect of those to whom they are
accountable.
In determining whether or not to seek the assistance of a third party, manage-
ment and labor are influenced by the complex social-institutional environment
of collective bargaining. For example, the parties to collective bargaining and
labor disputes often are required by law to use mediation before taking a job
action (e.g. a strike or a lockout) or moving to the next phase of the impasse
procedure. Additionally, some collective bargaining contracts call for mediation
of grievances.
Existing laws, contracts, and impasse procedures hold both parties account-
able for their actions. In addition, both sides must answer to their constituencies
and seek their constituents' approval. That is, the union bargaining team must
satisfy its members and the management bargaining team must satisfy its
employer.
Sources~Nature of Conflict
Some of the possible sources of conflict include: economic characteristics (e.g.
employer's inability to pay, wage erosion), structural characteristics of the
relationship (e.g. pattern-breaking relationship), organizational characteristics of
the parties (e.g. negotiators' lack of authority to bargain, internal conflicts within
one or both of the parties), interpersonal characteristics (e.g. hostility between
the parties), personal characteristics (e.g. negotiators' lack of skill or experi-
ence), the nature of the issues (e.g. "matters of principle" at stake), and the
bargaining behavior of the parties (e.g. unrealistic expectations, over-commit-
ment to a position, unwillingness to settle) (Kochan & Jick, 1978). Recent
studies have confirmed the role of these factors in determining the success of
mediation. To illustrate, in his study of the labor dispute between Air Canada
and the International Association of Machinists (IAM), Downie (1991) found
that economic and structural characteristics (e.g. the IAM's agreements with
Air Canada's competitors and Air Canada's contracts with other unions), the
nature of the issues in dispute (e.g. pension fund "surplus"), as well as internal
discord (e.g. union politics) influenced the outcome of mediation.
Mediator Characteristics
In a humorous mood, William E. Simkin (1971), a well-known mediator, devel-
oped a list of 16 qualities sought in a mediator. The first 10 items were somewhat
entertaining including: "the guile of Machiavelli" and "the hide of a rhinoc-
eros" (Simkin, 1971, p. 53). The final six items were more serious including
such items as: "demonstrated integrity and impartiality" and "basic knowledge
of and belief in the collective bargaining process" (Simkin, 1971, p. 53).
More recent research efforts have identified trustworthiness, helpfulness,
friendliness, humor, intelligence, and knowledge of the substantive issues as
desirable mediator traits (Kochan & Katz, 1988). The following mediator
characteristics also have been shown to influence the outcome of mediation:
self-awareness, presence, and authenticity (Bowling & Hoffman, 2000), power
and authority (Conlon et al., 1994; Harris & Carnevale, 1990), authority and
experience (Karambayya et al., 1992), experience and tenacity (Briggs & Koys,
1990), status (Keashly & Newberry, 1995), and gender (Carnevale, Conlon,
Hanisch & Harris, 1989; Maxwell, 1992; Stamato, 1992).
In addition, Wittmer, Carnevale, and Walker (1991), Conlon et al. (1994),

and Karambayya et al. (1992) have documented the role that perceived medi-
ator bias/fairness has on the disputants' behaviors and consequently on the
outcome of mediation. With respect to gender, Carnevale et al. (1989) found
that female mediators tended to perceive more common ground between the
disputants and to integrate more and use pressing tactics less than their male
counterparts. Likewise, Wall and Dewhurst (1991) found that female mediators
used more clarifying formulations (as opposed to controlling formulations) than
their male counterparts.
Situational Characteristics
Situational factors also influence mediator techniques. The most important
situational factor is the parties' motivation to settle. In private sector collective
bargaining cases the threat of a strike serves as a prime motivator for the parties
to settle (Kochan & Katz, 1988). Other key situational factors include: the
nature/characteristics of the impasse procedure (Karim & Dilts, 1990; Wissler,
1995), parties' past experience with mediation (Magnusen & Lim, 1994), and
parties' trust in the mediation process (Liebman, 2000; Gadlin, 1991; Karim &
Dilts, 1990).
It is important to note that mediation works better at resolving some types
of impasse than others. For instance, Kochan and Jick (1978) found that
mediation was most successful in cases where negotiations stalled due to such
factors as over-commitment to a position (bargaining behavior) or inexperi-
enced negotiators (personal characteristics). Karim and Dilts (1990) also found
that personal characteristics were a shallow cause of conflict and that such
conflicts were easily overcome by mediation. In contrast, Kochan and Jick
(1978) found that mediation was least successful in resolving impasses caused
by economic factors.
The Parties
Naturally, the parties themselves influence the success of mediation. In
particular, information sharing can be a risky proposition for the parties involved
in labor negotiations. Thus, negotiators may be wary of cooperating with the
mediator and their opponents in the dispute (Kochan & Katz, 1988). Indeed,
Gadlin (1991) notes the importance of establishing trust between the parties.
Similarly, Ross and Weiland (1996) found that the degree of trust between the
parties influences mediator strategies and thereby influences the outcome of
mediation. Wissler (1995) found that the parties' goals played a role in
determining the outcome of mediation. In particular, if disputants had a
competitive, non-integrative orientation mediation tended to be unsuccessful.
Similarly, Karim and Dilts (1990) found perceptions of bargaining behaviors

such as the union's perception that "the other side was not interested in settling"
and management's contention that the union was "holding to past positions" to
be important predictors of the success of mediation.
Other Tactical Determinants

When choosing among these various mediation techniques, mediators are
influenced by a variety of factors. In addition to the factors outlined above,
previous research has identified the following factors influencing mediators'
choice of tactics: rules and standards, dispute characteristics, culture, mediation
context, and time pressure (Wall & Lynn, 1993). For instance, dispute charac-
teristics such as the power balance between parties (Laskewitz, Van De Vliert
& De Dreu, 1994) and an emotionally charged atmosphere (Lurid, 2000; Adler,
Rosen & Silverstein, 1998; Maxwell, 1992) have been shown to influence the
mediator's choice of tactics. Similarly, Bercovitch and Houston (2000) found
that the mediation environment exerts a strong influence on mediators' choices
of techniques and strategies. In addition, Carnevale and Conlon (1988) found
that, when working under time pressure, mediators used more pressing and
compensating tactics and fewer integrating tactics.
METHODOLOGY
Research Design
This research examines mediators' opinions of how mediation works. In other

words, it explores mediators' interpretations of their experiences with the
mediation process. It is based on data collected through unobtrusive measures,
as well as participant observation, qualitative interviews, and a written survey.
This research follows a grounded theory approach (Strauss & Corbin, 1990).
That is, the model presented and tested here is inductively derived from the
study of mediation in the labor relations context. This study began with a
literature review. The literature review was used to develop a rough model
of the mediation process. Using the rough model as a starting point,
additional data was collected from secondary sources, informal networking
with dispute resolution professionals, participant observation of the FMCS'
new mediator training program, and qualitative interviews with FMCS
mediators. The collection of qualitative data began with the FMCS'
50th anniversary conference in September 1997 and continued through
September 1998.
The qualitative data were used to develop a written survey. The written survey
was then used to collect quantitative data. Factor analysis was used to reduce
the data set to a manageable number of variables. Next, hypotheses were
developed to match the variables derived from factor analysis with the constructs
from the original model. Finally, the hypotheses were tested using regression
analysis. The end result is a triangulated research design that makes use
of both qualitative and quantitative data. The procedure of using quantitative
data to validate qualitative analysis has been well-documented by Denzin (1978).
Survey Construction and Distribution
As mentioned above, qualitative data were collected from a variety of sources

and used to develop a written survey. Prior to distributing the survey, it was
pilot-tested on six FMCS mediators. Revisions were made based on the
feedback received. Then, 185 surveys were distributed at the FMCS National
Professional Development meeting in November 1998. A total of 78 usable
surveys were received, for a 42% response rate. The mediators were asked to
provide information about two dispute mediation (i.e. contract negotiation)
cases, one in which the parties reached agreement and one in which the parties
failed to reach agreement. Since each respondent provided data on two
mediation cases, there were 156 observations on which to base the data analysis.
The written survey contained 83 questions which used 4- and 5-point Likert
scales. Five of these questions pertained to consequences of mediation, but the
remaining questions were aimed at determining how mediators choose to
mediate. In addition, the survey included demographic questions, discussed
below. Following common practice (Tabachnick & Fidell, 1996), missing
values, for questions that were not answered by an individual respondent, were
replaced by the mean value of that variable.
Data Reduction
Factor Analysis
Since the goal of this research was to more formally model the determinants
of mediators' tactics and develop empirical testing in this area, factor analysis
was used to reduce these 83 Likert-scale questions to underlying factors. In
other words, factor analysis was used to confirm and extend the variable
identification derived from the literature review. An eleven factor solution was
obtained. Ten of the factors were used in this analysis. One factor, "relation-
ship improvement," was not used in this analysis because it captured four of
the questions pertaining to consequences of mediation. Factor-based scales were
Table 1. Factors Derived From Mediator Survey.
Factors Derived From Mediator Survey
Factor Percent of Sample of Questions Loading

Variance Highly on Factor
Mediator 5.5 My trustworthiness was important in bringing the parties

Acceptability closer to agreement.
The FMCS' reputation for being credible, acceptable, and
professional was important in bringing the parties closer to
agreement.
Management 5.2 The management bargaining team had realistic expectations of
Outlook the bargaining/mediation process.
The chief negotiator(s) for the management team was(were)
experienced/skilled in negotiation.
Relationship 4.7 Personality conflicts between the chief bargainers played a
Volatility critical role in the outcome of this case.
Hostility between the parties played a critical role in the
outcome of this case.
Broad 3.7 I tried to look beyond the contractual issues in defining the
Approach problem to be resolved.
Reframing was an important technique in bringing the parties
closer to agreement.
Bargaining 3.5 In determining which techniques/strategies to use in this case,
Context I was influenced by an emotionally charged atmosphere.
In determining which techniques/strategies to use in this case,
I was influenced by time pressures.
Bargaining 3.1 The threat of a strike played a critical role in the outcome of
Chips this case.
Pending NLRB charges played a critical role in the outcome of
this case.
Collaborative 2.8 The union team willingly shared information at the bargaining
Orientation table.
The parties' willingness to work together in a collaborative
process over the entire range of the relationship played a
critical role in the outcome of this case.
Narrow 2.6 I encouraged the parties to focus on resolving specific
Approach contractual problems.
My knowledge of substantive issues was important in bringing
the parties closer to agreement.
Structure of 2.4 In determining which techniques/strategies to use in the
Impasse this case, I was influenced by the nature of the impasse
procedure.
The nature of the impasse procedure played a critical
role in the outcome of this case.
Mediator Skill 2.3 My labor relations skills/experience were important in bringing
Base the parties closer to agreement.
My process skills were important in bringing the parties closer
to agreement.
created by summing those variables which loaded highly on each factor (Kim
& Mueller, 1978, p. 70). The ten factors are summarized in Table 1.
A noteworthy result of the factor analysis was that the questions pertaining
to mediator tactics boiled down to two factors, "the broad approach" and "the
narrow approach." The techniques clustering into each factor corresponded
closely to Riskin's (1993) typology of broad and narrow mediation strategies.
Mediators following the broad approach operate on the assumption that the goal
of mediation is to reach an agreement that serves the mutual interests of the
parties. The focus is on developing and understanding options. In addition, the
broad approach to mediation deals with barriers to negotiation such as emotional/
interpersonal problems and communication problems between the parties and
outside actors. Under the broad approach one of the primary objectives of
negotiation is improving the relationship between the parties. Furthermore, this
approach emphasizes encouraging and empowering the parties to make their
own decisions (Riskin, 1993).
In contrast the narrow approach, as it name suggests, narrowly defines the
conflict to be mediated. This narrow definition of the conflict restricts the issues
that can be discussed as part of the mediation process. The narrow approach
to mediation places an emphasis on gaining concessions from the parties.
Moreover, when this approach is followed the possible outcomes of mediation
are severely limited (Riskin, 1993).
Control Variables
In addition to the questions which were used in the data reduction, six control
variables were also used in the quantitative analysis. These are as follows:
private sector, gender, previous experience as an advocate for management,
previous experience in the private sector, length of tenure with the FMCS, and
unionization rate. "Private sector" refers to the sector in which the mediation
case took place. "Unionization rate" was a percentage measure of how heavily
unionized the geographic area was in which the mediator handled most of his
or her cases. The other control variables are self explanatory.
Matching Variables and Constructs
Mediator Tactics
As previously noted, mediators may draw on a variety of techniques to help
the parties reach agreement. The tactics box in the model attempts to capture
some of these various methods. The variables that matched with the mediator's
tactics construct are "the broad approach" and "the narrow approach."
These approaches represent different sets of mediator techniques, but the

broad approach is not the exact opposite of the narrow approach. In fact, these
two factor-based scales are positively correlated with each other (r = 0.38,
p < 0.01). Mediators do not appear to be choosing between these two sets of
tactics. It seems instead that they are choosing whether or not to apply both of
them simultaneously. Therefore, I summed the two factor-based scales to create
one measure of mediator tactics, which I named "the bifocal approach." One
of the dictionary definitions of "bifocal" is "embodying two distinct and often
conflicting goals, interests, or courses of action" (American Heritage Dictionary,
2000, p. 28). The strategic management literature uses the term "bifocal vision"
(Albrecht, 1994; Gogan, 1998; Harari, 1997; Houston, 2000) to describe "the
ability to perceive accurately things on the horizon that will inevitably affect
the enterprise, as well as the ability to focus on the more immediate, pressing
events" (Albrecht, 1994, p. 42). In the bifocal approach, the mediator attempts
to remove underlying barriers to negotiation while resolving a manageable set
of issues.
As mentioned earlier, the mediation process begins with the parties' interac-
tions. The parties' interactions box in the model focuses on how the parties
approach the conflict and the mechanisms in place for resolving the conflict.
The variable that matches with the parties' interaction construct is "structure
of the impasse."
Mediator characteristics also influence the mediator's choice of techniques.
The mediator characteristics construct encompasses a variety of traits. Some
of these include: credibility, experience, trustworthiness, helpfulness, friendli-
ness, humor, intelligence, and knowledge of the substantive issues. The
variables which match with this construct include both factor-based scale vari-
ables and demographic variables. The following factor-based scales were
matched with the mediator characteristics construct: "mediator acceptability"
and "mediator skill base." The correlation between these two factor-based
scales, while statistically significant, is rather modest (r = 0.18, p < 0.05).
Therefore, I chose not to sum these two scales into a single measure of
mediator characteristics. The following demographic variables were matched
with the mediator characteristics construct: "previous experience in the private
sector," "previous experience as a management advocate," "length of tenure
with the FMCS," and "gender".
Similarly, the sources and nature of conflict influence the mediator's choice of
tactics. Two factor-based scale variables matched with the sources/nature of
conflict construct. They were "relationship volatility" and "collaborative
orientation." These scales were uncorrelated (r = -0.08), so again I chose not
to sum them to form a single index of sources/nature of conflict.
Likewise, situational factors influence the mediator's choice of techniques.
Three variables matched with the situational characteristics construct. They were
"bargaining chips" (a factor-based scale), "private sector", and "unionization
rate".
The Parties
The parties themselves also influence the mediator's choice of tactics. The
variable which matches with the parties' construct is "management outlook".
"Management outlook" is a factor-based scale variable which encompasses
management's: desire for the mediation to be successful, realistic expectations
of the process, and bargaining experience. It was expected that both union and
management participants in the mediation process would influence the
mediator's choice of techniques. Indeed, the survey asked a parallel set of
questions about union participants in mediation. However, these questions failed
to load highly in the factor solution.

The variable which matched with the other tactical determinants construct was
"bargaining context". This variable encompasses strategic determinants such as
an emotionally charged atmosphere, the nature of issues in dispute, and time
pressures. The complete correspondence between factor-based scales, control
variables, and constructs is displayed graphically in Fig. 2.
HYPOTHESES
Figure 2 shows seven constructs, one of which is the dependent variable,

mediator tactics. This section briefly summarizes the rationale for expecting a
significant relationship between each construct and the dependent variable. If
the variables which match each construct are significantly related to the use of
the "bifocal" approach to mediation, then the hypotheses are supported.
0
L)
e-
©
k~
;.d
e~
o
L)
\ e~
.d
o)
.=.
8
[-
~C "= o
~e
.=.
~a
~q
~Z
The first factor-based scale in this category was mediator acceptability. It was
expected that credible, trustworthy mediators would be more successful in using
an integrative approach to bargaining, i.e. an approach which focuses on
improving the entire relationship between the parties, rather than the traditional
approach which narrowly defines the conflict to be mediated. For example, in
the new mediator training program the trainees were presented with a code of
conduct. The code of conduct emphasized maintaining standards of honesty,
integrity, and principle. In addition, throughout the training program the instruc-
tors stressed the importance of mediator confidentiality. Similarly, the interview
data reveals the importance of professionalism, ethical behavior, operational
neutrality, credibility, and sincerity. All of these characteristics help make the
mediator acceptable to the parties.
Given the voluntary nature of mediation, mediator acceptability is critical to
the success of mediation. Furthermore, if the mediator is acceptable to the parties
s/he is more likely to be successful in securing the parties' consent to use the
broad approach, which is a departure from the more narrowly focused
traditional approach to collective bargaining.
The second factor-based scale in this category was mediator skill base. The
data collected through participant observation and qualitative interviews
indicate that mediators need both substantive knowledge (i.e. labor relations
skills and experience) and process knowledge (i.e. facilitation and problem
solving skills) to be effective. In particular, the need for both types of
knowledge is evident in the FMCS' mediator core competencies (Mareschal
1998). This message was reinforced in the qualitative interviews. Although the
mediators interviewed disagreed about the relative importance of the two types
of knowledge, they tended to agree that both types were necessary. Thus, it
was expected that a mediator's labor relations skills/experience should
facilitate use of the traditional approach, focused on specific bargaining issues,
and a mediator's process skills should facilitate use of the integrative approach,
focused on developing and understanding options, removing barriers to
negotiation, and improving the relationship between the parties.
Hypothesis One: Mediator characteristics will be related to use of the

bifocal approach.
The first factor-based scale in this category was relationship volatility. It was
expected that hostile relationships within and between the parties would make
it difficult for mediators to use integrative techniques to bring the parties closer
together. For example, the new mediator training program included a section
on factors impacting mediation effectiveness. Here the instructors noted that
the nature of the relationship between the parties affects the mediator's ability
to persuade the parties to resolve their dispute. When the relationship between
the parties is hostile, the mediator's ability to persuade the parties by using
reward, legitimate, and informational power is severely restricted. This premise
was supported by the interview data. In particular, the mediators noted that
personality conflicts between the bargaining teams can be some of the most
difficult sources of conflict to resolve.
The second factor-based scale in this category was collaborative orientation.
It was expected that an atmosphere of mutual respect and openness between
the parties would make it easier for mediators to use integrative techniques to
bring the parties closer together. In particular, the instructors in the new
mediator training program noted that when both parties want to work on their
ongoing relationship to make it more productive, the mediator's ability to
persuade the parties through the use of reward and legitimate power is increased.
Likewise, the mediators interviewed indicated that the mediation process flows
more smoothly when the parties respect and trust each other.
Hypothesis Two: Sources/nature of conflict will be related to use of the

bifocal approach.
The factor-based scale in this category was bargaining chips. It was expected
that the ability of one party to pressure the other would make it easier for
mediators to use integrative techniques to bring the parties closer together.
Specifically, the instructors in the new mediator training program noted that the
most essential factor impacting mediation effectiveness is the pressure on the
parties to settle. As the pressure to settle becomes greater, the mediator's ability
to persuade the parties through the use of reward power and coercion is
enhanced.
A control variable in this category was unionization rate. It was expected
that areas with higher unionization rates would facilitate integrative mediation
techniques. In particular, one mediator from a state where the unionization rate
is relatively low indicated that he thought mediators in states with higher
unionization rates would be freer to experiment with non-traditional techniques.
This mediator's perception may be due in part to the fact that where unions
are relatively rare, public opinion can be presumed to be less favorable to
collective bargaining.
Hypothesis Three: Situational characteristics will be related to use of

the bifocal approach.
The Parties
The only independent variable in this category was management outlook.
Management outlook is a factor-based scale variable which encompasses
management's: desire for the mediation to be successful, realistic expectations
of the process, and bargaining experience. It was expected that management
intransigence would inhibit the use of integrative techniques, while a favor-
able management outlook would facilitate the use of integrative techniques.
For example, the interview data demonstrate that the parties must believe in
the mediation process and must want to resolve their dispute for mediation to
be successful. If management does not want to deal with the union or if
management wants the union to be decertified, the case can be very difficult
to mediate.
Hypothesis Four: Characteristics of the parties will be related to use of

the bifocal approach.
The only independent variable in this category was structure of the impasse. It
was expected that a highly structured impasse procedure would inhibit the use
of integrative techniques. In some cases mediation is one of many steps in
the impasse procedure. That is, additional dispute resolution processes are
available to the parties after they participate in the mediation process. Under
these circumstances, it was expected that the mediator would have greater
difficulty in using the broad approach, which is aimed at improving the parties'
entire relationship, because the parties tend to view mediation simply as a
stepping stone to other dispute resolution processes.
Hypothesis Five: Parties' interactions will be related to use of the

bifocal approach.

The only independent variable in this category was bargaining context. It was
expected that a pressurized, emotionally charged atmosphere would inhibit the
use of integrative techniques. This outcome was expected because the broad
approach requires the parties to consider mutual interests and focuses on
developing and understanding various options. When faced with extreme time
pressures and an emotionally charged atmosphere the mediator may find it

difficult to get the parties to step away from their entrenched positions and try
to see the "big picture".
Hypothesis Six: Other tactical determinants will be related

to use of the bifocal approach.
RESULTS
Tables 2 and 3 display descriptive statistics and correlations for all variables
used in this analysis. In the following section the statistically significant results
of a multiple regression are reviewed. The dependent variable is use of the
bifocal approach.
The results of the regression predicting the bifocal approach are shown
in Table 4. The following predictors of the bifocal approach were found to
be statistically significant: bargaining context (p<0.01), unionization rate
(p < 0.01), mediator acceptability (p < 0.01), and bargaining chips (p < 0.05).
As depicted in Fig. 2, two of these predictors matched the situational
characteristics construct, one predictor matched the mediator characteristics
construct, and one matched the other tactical determinants construct. The
Table 2. Descriptive Statistics for All Variables Used in Regressions.

(n = 156)
Variable Mean S.D
Factor-based scales:
1. Structure of impasse 7.90 2.61
2. Mediator skill base 6.36 2.03
3. Acceptability 11.14 3.12
4. Management outlook 12.35 3.69
5. Relationship volatility 13.64 4.10
6. Bifocal approach 17.79 4.04
7. Bargaining context 7.82 2.62
8. Bargaining chips 17.78 4.35
9. Collaborative orientation 13.67 4.01
Other variables:
10. Private sector 0.81 0.40
11. Gender (Male = 1) 0.82 0.37
12. Management advocate 0.36 0.48
13. Tenure w/FMCS 9.13 9.42
14. Private sector work exp. 0.90 0.30
15. Unionization rate 33.04 19.17
t~
I I I II II
~A I I I I I II
i::::
t-q
I I I I II
I II II
I I I I I I I
>
[..,
~4
v
significant results supported Hypotheses One, Three, and Six. Given the very
high number of significant bivariate correlations in Table 3, collinearity
diagnostics are also shown in Table 4. Tolerance values range from 0.57 to
0.90, averaging about 0.75, which indicates that multicollinearity was not a
serious problem in this regression.
So What?: Interpreting Statistically Significant Results
Three coefficients were positively related to the use of the bifocal approach.
These were mediator acceptability, bargaining chips, and bargaining context.
Again, given the voluntary nature of mediation, it is not surprising that
mediator acceptability is related to a mediator's choice of tactics. In essence,
a mediator must gain the parties' consent to assist in resolving the dispute.
Bargaining chips measured the ability of one party to pressure the other,
and bargaining context indicated an emotionally charged atmosphere. When at
least one party has such tools at hand, it appears that the mediator is in a better
position to persuade the parties to focus on developing and understanding options
for a mutually acceptable agreement rather than following the traditional
approach of trying to extract concessions and inflict losses on one's opponent.
Table 4. Regression Results: Dependent Variable = Bifocal Approach.

(n = 156)
Independent Variable b s.e.(b) t tolerance
Mediator acceptability 0.41 0.11 3.56** 0.57

Mediator skill base 0.13 0.17 0.75 0.61
Gender (Male = 1) -1.44 0.81 -1.78 0.80
Tenure w/FMCS 0.05 0.03 1.48 0.82
Private sector exp. 0.16 1.00 0.16 0.78
Management advocate 0.08 0.61 0.13 0.86
Relationship volatility 0.07 0.07 0.89 0.79
Collaborative orientation 0.11 0.08 1.44 0.74
Bargaining chips -0.15 0.07 2.06* 0.70
Unionization rate -0.05 0.02 -3.08** 0.90
Private sector case - 1.02 0.84 - 1.22 0.67
Management outlook -0.01 0.09 -0.07 0.70
Structure of impasse 0.21 0.12 1.65 0.69
Bargaining context 0.33 0.12 2.74** 0.76
Constant 11.52 2.22 5.19"*
F-statistic 5.91 * *
Adjusted r-squared 0.31
* significant, p < 0.05; ** significant, p < 0.01.

One coefficient, unionization rate, was negative. The dependent variable is

based on Likert scales for which a lower number indicates greater agreement.
Thus, in the case of control variables, a negative coefficient indicates a posi-
tive relationship with use of the bifocal approach. Where unions are relatively
rare the collective bargaining process has not been institutionalized. Under these
circumstances, it may be more difficult for the mediator to employ a coherent
set of tactics.
Theoretically Interesting, Statistically Insignificant Results
In particular, two sets of statistically insignificant results are theoretically

interesting. First, there is the failure to confirm a relationship between the
management outlook and the dependent variable. Based on the interview data,
it was expected that management outlook would be significantly related to the
bifocal approach. However, the management outlook variable did not come
close to statistical significance.
Another look at Table 3 explains this result. Here management outlook is
correlated with the dependent variable at p < 0.05 and more highly correlated
with the following statistically significant independent variables: mediator
acceptability, bargaining chips, and bargaining context (p < 0.01). Moreover, in
a bivariate regression the coefficient of management outlook is significant at
p < 0.05, whereas in the multivariate regression as reported in Table 4 the
t-statistic for management outlook was the smallest of all the independent
variables. Thus, management outlook appears to exert an indirect effect on the
bifocal approach through the direct effects of mediator acceptability, bargaining
chips, and bargaining context.
The second set of theoretically interesting, statistically insignificant results
concerns the mediator skill base. One key question surrounding the growing
popularity of mediation as a dispute resolution tool is: what do mediators need
to know to help the parties resolve their disputes? Recent attempts to develop
objective standards for gauging the knowledge, skills, and abilities required of
mediators have met with great resistance (Bush, 1993; Kolb & Kolb, 1993;
Salem, 1993).
Part of the debate over mediators' qualifications focuses on the relative
importance of substantive knowledge (e.g. labor relations skills and experience)
and process knowledge (e.g. facilitation and problem solving skills). The
mediators interviewed disagreed over the relative importance of these two types
of skills. Indeed, the FMCS is currently experimenting with the mix of skills
required of new mediators. The factor analysis performed here suggests
that both types of knowledge are necessary. Unfortunately, in the regression
analysis, mediator skill base was not significantly related to the bifocal
approach. However, mediator skill base is correlated with the dependent
variable, as well as mediator acceptability, relationship volatility, and bargaining
context at p < 0.01. Again, this suggests that mediator skill base may have
an indirect effect on the dependent variable through the direct effects of these
other variables.
The failure to establish a statistically significant relationship between
mediator skill base and the bifocal approach may have to do with the
homogeneity of the survey respondents. Until very recently, the FMCS recruited
mediators based on their substantive knowledge and skills. That is, applicants
were required to have a minimum of seven years of "front-line" labor
negotiations experience. Within the past five years, the FMCS has begun hiring
applicants with strong process skills who may not have had the minimum seven
years front-line experience. The new mediators hired on the basis of their
process skills are taught substantive skills on the job (FMCS, 1996).
Once hired, the new mediators attend a comprehensive training program over
the course of a year. The topics covered include dispute mediation, preventive
mediation, and alternative dispute resolution (Mareschal, 1998). The training
sessions serve to initiate the new mediators to the agency and provide the basic
tools they need to do their jobs. While the training sessions alone will not make
the mediators successful, they give a flavor of what mediators can expect on
the job. In addition, all new mediators begin their work with the FMCS by
"shadowing" other more seasoned mediators on cases. Only after the shadowing
period has been successfully completed do the new mediators begin to handle
cases on their own. Since the FMCS only recently relaxed the requirement
concerning labor negotiations experience and since all new mediators with the
FMCS complete this program, FMCS mediators tend to develop a uniform/similar
skill base. As a result, in self-reports they may underestimate the role that
their skill base plays in determining the tactics that they use.
CONCLUSION
As noted in the introduction, this research strives to understand how mediation

works and why mediators do what they do. In brief, an inductive study was
conducted to further develop a theoretical model of the determinants of
mediators' tactics derived from a review of the mediation literature. With the
theoretical model in mind, qualitative and quantitative data were collected to
refine and test the model. Six hypotheses were developed from the preliminary
analysis. These hypotheses were tested using regression analysis. Four
statistically significant relationships were found. Three hypotheses were

confirmed.
To summarize, in mediation a neutral third party assists the parties in conflict
in reaching a voluntary agreement. The mediator does not have the power or
authority to impose a settlement. Instead, s/he simply facilitates the negotiation
process. The primary goal of the mediator is to help the parties come to an
agreement.
Mediation is important for a number of reasons. In particular, mediation is
quickly becoming the policy instrument of choice to resolve disputes at various
levels of government. Furthermore, mediation promotes labor-management
cooperation; such cooperation is essential to organizational effectiveness.
Moreover, since managers are involved in team building, resolving interper-
sonal disputes, and handling grievances they are in essence bargaining all the
time.
Indeed, courses in conflict resolution/mediation are being offered in the
curricula of many business and public administration programs in an effort to
help managers understand the nature of conflict in their work and their role as
conflict resolvers. The growing interest in teaching negotiation and dispute
settlement is also reflected in the editorial policies of multidisciplinary journals
such as Negotiation Journal which features a section on Educational
Innovations. Thus, mediation is a topic WOl:thy of study for both practitioners
and scholars in a variety of disciplines such as business and public adminis-
tration, public policy analysis, organizational studies, labor relations, and
conflict resolution.
Although mediation has become increasingly popular as a means to resolve
conflict the practice of mediation has outstripped theory-building. As mentioned
earlier, mediators often claim that mediation is an art, not a science. However,
given the explosive growth in the use of conflict resolution services, the premise
that mediation cannot be evaluated has become untenable.
Limitations and Suggestions for Future Research
The most important limitation of the research is the researcher's inability to

observe any mediations directly. Obviously, first-hand observation would have
enhanced the understanding of the mediation process. However, since it is often
argued that no two mediations are alike, direct observation of a few mediation
cases might provide a less complete picture of the mediation process than the
triangulated research design used here.
This research could be extended in several directions. First, the labor
relations context is highly structured. For example, the parties have rights
established by law and labor boards exist to enforce these laws and to deal with
parties who bargain in bad faith. It would be interesting to see how the
mediation process works in less structured environments. The survey could be
replicated in different contexts such as family mediation as a step toward
devising a general theory of mediation.
Second, within the labor relations context the survey could also be adminis-
tered to other mediation participants such as management and union negotiators
in order to gain alternative perspectives on the mediation process. As Schtn
(1983) notes, the reflective practitioner learns from the client's interpretation
of the problem. If research on mediation is to have practical value, then
researchers would do well to learn from both the mediators (i.e, the
practitioners) and the parties (i.e. the clients).
Third, the FMCS itself is an interesting research site for conflict resolution
scholars as it attempts to modernize its image, broaden the range of services it
provides, and diversify its workforce. Indeed, in a recent Industrial Relations
Research Association newsletter practitioners and researchers alike were called
upon to participate in a national policy forum. One of the key topics discussed
in the forum was the massive organizational changes that the FMCS has under-
taken in recent years (Kochan, 1999). Although no significant impacts of mediator
demographic characteristics were found it is possible that, as the FMCS strives
to recruit a new type of mediator for the new millennium, new clusters of medi-
ator techniques, as well as new determinants of these tactics, may be discovered.
Finally, Federal agencies involved in labor relations (i.e. the National Labor
Relations Board and the FMCS) have a somewhat weak history of conducting
or sponsoring research on the phenomena they regulate or otherwise influence.
This research is a first step at reversing that "tradition." This new trend toward
more openness could be extended with a concentrated effort to further test the
model of mediators' tactical choices presented here.
REFERENCES
Adler, R. S., Rosen, B., & Silverstein, E. M. (1998). Emotions in negotiation: How to manage fear
and anger. Negotiation Journal, 14, 161-179.
Albrecht, K. (1994). The power of bifocal vision. Management Review, 83(4), 42-46.
American Heritage Dictionary of the English Language (4th ed.) (2000). Boston: Houghton-Mifflin.
Bazerman, M. (1998). Judgment in Managerial Decision Making (4th ed.). New York: Wiley.
Bellman, H. S. (1998). Some reflections on the practice of mediation. Negotiation Journal, 14,
205-210.
Bercovitch, J., & Houston, A. (2000). Why do they do it like this? An analysis of the factors
influencing mediation behavior in international conflicts. Journal of Conflict Resolution, 4,
170-202.
66 P A T R I C E M. M A R E S C H A L
Bingham, L. B., Chesmore, G., Moon, Y., & Napoli, L. M. (2000). Mediating employment disputes
at the United States Postal Service: A comparison of in-house and outside neutral mediator
models. Review of Public Personnel Administration, 20(1), 5-19.
Briggs, S., & Koys, D. J. (1990). An empirical investigation of public-sector mediator effective-
ness. Journal of Collective Negotiations, 19, 121-128.
Bowling, D., & Hoffman, D. (2000). Bringing peace into the room: The personal qualities of the
mediator and their impact on mediation. Negotiation Journal, 16, 5-28.
Bush, R. A. B. (1993). Mixed messages in the Interim Guidelines. Negotiation Journal, 9, 341-347.
Camevale, P. J., & Conlon, D. E. (1988). Time pressure and strategic choice in mediation.
Organizational Behavior and Human Decision Processes, 42, 111-133.
Carnevale, P. J., Conlon, D. E., Hanisch, K. A., & Harris, K. L. (1989). Experimental research on the
strategic-choice model of mediation. In: K. Kressel & D. G. Pruitt (F_xls),Mediation Research.
San Francisco: Jossey-Bass.
Camevale, P. J., & Pruitt, D. G. (1992). Negotiation and mediation. Annual Review of Psychology, 43,
531-582.
Conlon, D. E., Camevale, P., & Ross, W. H. (1994). The influence of third party power and
suggestion on negotiation: The surface value of a compromise. Journal of Applied Social
Psychology, 24, 1084-1113.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297-334.
Denzin, N. K. (1978). The Research Act (2rid ed.). New York: McGraw-Hill.
Downie, B. M. (1991). When negotiations fail: Causes of breakdown and tactics for breaking the stale-
mate. Negotiation Journal, 7, 175-186.
Elangovan, A. R. (1998). Managerial intervention in organizational disputes: Testing a prescriptive
model of strategy selection. International Journal of Conflict Management, 9, 301-335.
Federal Mediation and Conciliation Service (n.d). Labor-Management Relations for the 21st
Century. Washington, D.C.: Federal Mediation and Conciliation Sei'vice.
Federal Mediation and Conciliation Service (1996). Transformation: Federal Mediation
and Conciliation Service 48th Annual Report. Washington, D.C.: U.S. Government Printing
Office.
Federal Mediation and Conciliation Service (1997). Federal Mediation and Conciliation
Service Strategic Plan 1997-2002. Washington, D.C.: Federal Mediation and Conciliation
Service.
Feuille, P. (1999). Grievance mediation. In: A. E. Eaton & J. H. Keefe (Eds), Employment Dispute
Resolution and Worker Rights in the Changing Workplace. Champaign, IL: Industrial
Relations Research Association.
Gadlin, H. (1991). Careful maneuvers: Mediating sexual harassment. Negotiation Journal, 7,
139-153.
Gogan, J. L. (1998). Bifocal IS Management. Informationweek, 13(July), 146.
Harari, O. (1997). Looking beyond the 'vision thing.' Management Review, 86(6), 26-29.
Harris, K. L., & Carnevale, P. (1990). Chilling and hastening: The influence of third-party power
and interests on negotiation. Organizational Behavior and Human Decision Processes, 47,
138-160.
Houston, P. D. (2000). Balancing Paradox. Association Management, (June), 62-66.
Jorgensen, E. O. (2000). Relational transformation in mediation: Following constitutive and regu-
lative rules. Mediation Quarterly, 17, 295-312.
Karambayya, R., Brett, J. M., & Lytle, A. (1992). Effects of formal authority and experience on
third-party roles, outcomes, and perceptions of fairness. Academy of Management Journal,
35, 426--438.
Resolving Conflict: Tactics o f Federal Mediators 67
Karim, A., & Dilts, D. A. (1990). Determinants of mediation success in the Iowa public sector.
Journal of Collective Negotiations, 19, 129-140.
Kearney, R. C., & Carnevale, D. G. (2001). Labor Relations in the Public Sector (3rd ed.). New
York: Marcel Dekker.
Keashly, L., & Newberry, J. (1995). Preference for and fairness of intervention: Influence of third-
party control, third-party status and conflict setting. Journal of Social and Personal
Relationships, 12, 277-293.
Kim, J., & Mueller, C. W. (1978). Factor Analysis: Statistical Methods and Practical Issues.
Newbury Park, CA: Sage.
Kochan, T. A. (1999). President's column. IRRA Newsletter, 41(2), 5.
Kochan, T. A., & Jick, T. (1978). The public sector mediation process: A theory and empirical
examination. Journal of Conflict Resolution, 22, 209-240.
Kochan, T. A., & Katz, H. C. (1988). Collective Bargaining and Industrial Relations (2nd ed.).
Homewood, IL: Irwin.
Kolb, D. M. (1983). The Mediators. Cambridge, MA: MIT.
Kolb, D. M., & Kolb, J. E. (1993). All the mediators in the garden. Negotiation Journal, 9, 335-339.
Kressel, K., Frontera, E. A., Fodenza, S., Butler, F., & Fish, L. (1994). The settlement orientation
vs. the problem-solving style in custody mediation. Journal of Social Issues, 50, 67-84.
Laskewitz, P., van de Vliert, E., & De Dreu, C. K. W. (1994). Organizational mediators siding
with or against the powerful party? Journal of Applied Social Psychology, 24, 176--188.
Liebman, C. B. (2000). Mediation as parallel seminars: Lessons from the student takeover of
Columbia University's Hamilton Hall. Negotiation Journal, 16, 157-182.
Lipsky, D. B., & Seeber, R. L. (1999). In search of control: The corporate embrace of ADR.
University of Pennsylvania Journal of Labor and Employment Law, 1, 133-157.
Lund, M. E. (2000). A focus on emotion in mediation training. Family and Conciliation Courts
Review, 38, 62-68.
Magnusen, K. O., & Lim, R. G. (1994). Special master mediation in impasse resolution: The Florida
experience. Journal of Collective Negotiations, 23, 347-358.
Mareschal, P. M. (1998). Providing high quality mediation: Insights from the Federal
Mediation and Conciliation Service. Review of Public Personnel Administration, 18(4),
55-67.
Maxwell, D. (1992). Gender differences in mediation style and their impact on mediator effec-
tiveness. Mediation Quarterly, 9, 353-363.
Noe, R. A., Hollenbeck, J. R., Gerhart, B., & Wright, P. M. (2000). Human Resource Management:
Gaining a Competitive Advantage (3rd ed.). Boston: Irwin.
Phillips, B. (1999). Reformulating dispute narratives through active listening. Mediation Quarterly,
17, 161-180.
Riskin, L. L. (1993). Two concepts of mediation in the FMHA's Farmer-Lender Mediation Program.
Administrative Law Review, 45, 21-64.
Ross, W. H. (1990). An experimental test of motivational and content control on dispute
mediation. Journal of Applied Behavioral Science, 26, 111-118.
Ross, W. H., & Wieland, C. (1996). Effects of interpersonal trust and time pressure on manage-
rial mediation strategy in a simulated organizational dispute. Journal of Applied Psychology,
81, 228-248.
Salem, R. A. (1993). The Interim Guidelines need a broader perspective. Negotiation Journal, 9,
309-312.
Sch/Sn, D. A. (1983). The Reflective Practitioner: How Professionals Think in Action. New York:
Basic.
68 P A T R I C E M. M A R E S C H A L
Simkin, W. E. (1971). Mediation and the Dynamics of Collective Bargaining. Washington, D.C.:
Bureau of National Affairs.
Solstad, K. E. (1999). The role of the neutral in intra-organizational mediation: In support of active
neutrality. Mediation Quarterly, 17, 67-81.
Stamato, L. (1992). Voice, place, and progress: Research on gender, negotiation, and conflict
resolution. Mediation Quarterly, 9, 375-386.
Strauss, A., & Corbin, J. (1990). Basics of Qualitative Research: Grounded Theory Procedures
and Techniques. Newbury Park, CA: Sage.
Tabachnick, B. G., & Fidell, L. S. (1996). Using Multivariate Statistics. New York: Harper Collins.
Tetlock, P. E. (1991). An alternative metaphor in the study of judgment and choice: People as
politicians. Theory and Psychology, 1,451-475.
Wall, J. A., Jr., & Lynn, A. (1993). Mediation: A current review. Journal of Conflict Resolution,
37, 160-194.
Wall, V. D., Jr., & Dewhurst, M. L. (1991). Mediator gender: Communication differences in resolved
and unresolved mediations. Mediation Quarterly, 9, 63-85.
Welton, G. L., Pruitt, D. G., & McGillicuddy, N. B. (1988). The role of caucusing in community
mediation. Journal of Conflict Resolution, 32, 181-202.
Wissler, R. L. (1995). Mediation and adjudication in the small claims court: The effects of process
and case characteristics. Law and Society Review, 29, 323-358.
Wittmer, J. M., Carnevale, P., & Walker, M. E. (1991). General alignment and overt support in
biased mediation. Journal of Conflict Resolution, 35, 594-610.
Zubek, J. M., Prnitt, D. G., Pierce, R. S., McGillicuddy, N. B., & Syna, H. (1992). Disputant and
mediator behaviors affecting short-term success in mediation. Journal of Conflict Resolution,
36, 546-572.
BROAD-BASED EMPLOYEE STOCK
OPTIONS - A UNION-NONUNION
COMPARISON
Maya K. Kroumova, James C. Sesil,

Douglas L. Kruse and Joseph R. Blasi
INTRODUCTION
Until recently, stock options were primarily reserved for senior executives and
selected managers in most American corporations. In the last decade or so,
however, stock options have become part of the compensation package for an
increasing number of rank-and-file employees. As of February 2000, the
National Center for Employee Ownership (NCEO) estimated that there are more
than 3000 active broad-based stock option (SO) plans in the United States based
on an extensive review of press announcements by companies.
The expected benefits of broad-based 2 SO plans resemble those of other
incentive compensation: reduced turnover and increased effort, creativity, and
cooperation, which in turn, presumably result in higher productivity and
ultimately better overall firm performance (Kroll, 1997). In addition, stock
options do not entail a direct charge against earnings) and they have the
potential to foster an "ownership" culture by focusing employee attention on
the firm's financial performance. 4 Finally, they allow a lot of flexibility
in tailoring rewards: SOs can be granted as a reward for joining the company;
they can be based on individual performance and/or meeting group/business
unit goals; or the grant can be tied to the profitability of the company, etc.

All rights of reproduction in any form reserved.
ISBN: 0-7623-0853-2
69
70 M.K. KROUMOVA ET AL.
Empirical evidence for the positive effects of stock options on productivity

and market valuation is beginning to accumulate (Weeden, Carberry & Rodrick,
1998; Sesil, Kroumova, Kruse & Blasi, 2000; Blasi, Kruse, Sesil & Kroumova,
2000). The expected benefits of SO plans make them attractive to both
unionized and non-unionized employers who are trying to design an efficient
compensation package. As unions tried to reinvent themselves as a valuable
partner for both employers and employees in the 1990s, they have become more
open to incentive pay, including stock options. 5 However, as of yet, there is
little conceptual and empirical work on how broad-based stock option plans
work, and what, if any, effect they have on company performance in union
firms. An empirical comparison of the relationship between SO plans and firm
performance in union and non-union firms may therefore be informative for
future theory building. The goal of this paper is to assess the differences in
economic and financial performance between stock option and non-stock option
companies in unionized and non-unionized settings.
INCENTIVES AND GROUP INCENTIVES:

THEORY AND EVIDENCE
In a publicly owned company, property rights, broadly defined as the rights to

the benefits, use, and disposal of assets and goods (Kang & SCrensen, 1999),
have become fragmented. Use and disposal (control) rights are separated from
benefit (ownership) rights (Berle & Means, 1932). Four broad classes of
stakeholders share these rights: security holders, directors, managers, and
employees. The diverging self-interests of these various stakeholders, along with
the fragmentation of property rights, create opportunities for capturing rights
that the original right holder cannot protect. 6 In the context of the employment
relationship, managers can capture benefit rights from shareholders and
employees, for example, by diverting cash towards perquisites; employees can
capture additional benefit rights from managers and shareholders by withholding
their effort (moral hazard) and/or providing inaccurate information about their
ability (adverse selection). Such self-interested behaviors could be contained if
the parties to the employment relationship had perfect information about each
other and obtaining such information was not costly (Alchian & Demsetz, 1972).
However, in the presence of imperfect knowledge and costly monitoring, the
employer (or principal, in an agency theory framework) has to address
problems resulting from self-interested worker (agent) behaviors such as moral
hazard and adverse selection. In particular, the firm has to design an efficient
compensation contract that provides employees with incentives to act in
ways that are consistent with the interests of the other stakeholders (Asch &
Broad-Based Employee Stock Options - A Union-Non-Union Comparison 71
Warner, 1997), and that takes into account that employees (agents) are more
risk averse than shareholders (Weitzman & Kruse, 1990). Thus, an efficient
compensation package would motivate employees to supply effort and to
self-select into firms/jobs based on their unobservable characteristics (Asch &
Warner, 1997).
A large variety of individual and group incentives are being used to address
problems brought about by self-interested employee behaviors. Individual
financial incentives include piece-rates, merit pay, commission systems,
individual performance-based bonuses, efficiency wages and upward-sloping
wage-tenure profiles.7 Group incentives include bonuses based on group/
business unit/firm performance, profit-sharing, gainsharing and employee
ownership plans. These financial incentives vary in their ability to influence
behavior, and their effectiveness is contingent on firm characteristics (Jones,
Kato & Pliskin, 1997).
Stock option plans are a type of group incentive plan, and as such their
impact on worker effort can be analyzed drawing on existing theoretical and
empirical knowledge about other group incentive plans. Group incentive plans
are expected to serve the same purpose as individual incentives, that is, to
motivate employees to provide effort in the amount and direction desired by
the firm. Specifically, theoretical arguments that predict a positive effect of
group incentives on firm performance are based on extending agency theory
(Eisenhardt, 1989) to conceptualize employees as agents of management.
Granting stock options to employees may reduce the incentive conflicts that
arise when the interests of workers are not aligned with the interests of owners
and managers. 8 An added benefit of group incentives is that they are expected
to result in improved cooperation among employees (FitzRoy & Kraft, 1987;
Weitzman & Kruse, 1990; Strauss, Gallagher & Fiorito, 1991; Kruse &
Blasi, 1997). Peer pressure and horizontal monitoring have been advanced
as explanatory mechanisms for this effect (Weitzman & Kruse, 1990;
Kruse, 1993). Consequently, group incentives may also reduce monitoring costs
(Kruse, 1993). Further, Mitchell, Lewin and Lawler (1990) hypothesized
that group incentives increase employee effort and commitment indirectly
via improved communication about company performance and better
employee understanding of the importance of profitability and organizational
effectiveness.
However, there are also theoretical reasons to expect negative or mixed effects
of group incentives on employee effort. First, depending on the design and
implementation of the plan, the connection between individual effort and
obtaining the group reward may be t e n u o u s 9 (Gomez-Mejia & Balkin, 1992;
Blasi, Conte & Kruse, 1996), which would result in a weak or non-existent
incentive to supply effort. Stock-based rewards may be especially susceptible

to this problem because, with these types of incentives, the reward is
experienced only if and when the stock price increases, yet many of the factors
that influence stock price movements over the short run are outside of employee
control. 10
Second, whenever it is hard to monitor individual employee effort and
when any one employee's reward depends on everyone else's effort, a
dilution or free-rider problem may occur. Employees may cut back on their effort
because individual accountability is low (Weitzman & Kruse, 1990; Cooke, 1994;
Jones et al., 1997). Thus, in a static, one-period framework, the larger the group
among which the reward is to be shared, the smaller the incentive value of the
reward and the higher the likelihood of free-riding behavior.~l Broad based stock
option plans are vulnerable to this criticism in that the reward (stock price
appreciation) is ultimately contingent on overall firm performance, and the reward
is shared with all other stakeholders. Therefore, one might expect group incentives
such as SO plans to have an insignificant or even negative effect on performance.
Nonetheless, theoretical arguments based on game theory suggest that in a
multi-period setting, where the "game" of group reward earning and sharing is
repeated, cooperation is a possible equilibrium solution (but not the only solution)
(Weitzman & Kruse, 1990; Levine & Tyson, 1990; Cooke, 1994). Ultimately,
whether cooperation will emerge as the actual solution will depend on the
organization's performance management system, culture, and other factors
(Weiss, 1987; Weitzman & Kruse, 1990; Hansen, 1997).
Third, whether a stock option plan will elicit additional effort from employees
will also depend on how employees "frame" the plan - as an extra reward
or as a trade-off for foregone increases in base-level compensation. Framing
a stock option plan as a trade-off may result in unchanged levels of work
effort. To complicate matters further, such framing may change over time
and be susceptible to fluctuations in stock price. For example, a stock option
plan may initially be perceived as a trade-off for pay: if the stock price increases
sharply it will likely be perceived as a reward, but if the stock price declines
sharply ensuing disappointment will likely lead to a decrease in work effort.
Finally, contextual factors are also likely to moderate the effects of group
incentives on employee behavior. Ben-Ner and Jones (1995) argue that
ownership based only on return fights (broadly defined as financial and
physical payoffs generated from the operation of the organization) is likely to
have small and conflicting effects on individual motivation and performance
and, as well, on cooperation across organizational units. They hypothesized that
granting employees both return and control fights would improve effort and
performance, albeit in a non-monotonic fashion.
Thus, theoretical arguments exist for both positive and negative effects of
group incentive plans on motivation, effort, and performance. In this regard,
several empirical studies have attempted to measure the net effect of group
incentives on various employee and organization-level outcomes. Kruse & Blasi
(1997) reviewed 26 empirical studies of the effects of employee stock
ownership ~2 on employee attitudes, motivation, and firm productivity and
profitability. Although the studies exhibit a wide variety of measures, owner-
ship forms, and contextual factors that make generalizations difficult, Kruse &
Blasi (1997) report that most studies find a positive relationship between
employee ownership and organizational commitment, and either positive or
neutral effects of employee ownership on employee satisfaction. Perceived
participation in decision-making appears to be an important predictor of
employee satisfaction and commitment, either by itself or interacted with
employee ownership. As far as firm productivity and profitability are concerned,
the evidence indicates either better or unchanged performance under employee
ownership. A weakness of this empirical research is that it sheds little light on
the specific mechanisms through which employee ownership might improve
performance, satisfaction, and other outcomes (see Pierce, Rubenfeld & Morgan,
1991, for a proposed model of how employee ownership affects productivity,
satisfaction and commitment).
Jones et al. (1997) provide a critical review of the available econometric
evidence on the effects of profit-sharing and gainsharing on performance.
The evidence clearly supports the hypothesis that profit-sharing improves
productivity, although its effect on profitability is less clear. As for the
effects of such variables as firm size and capital intensity on the effectiveness
of profit-sharing, the evidence is mixed, with studies variously finding positive,
neutral, and negative effects. The empirical evidence about the relationship
between worker participation and profit-sharing is also inconclusive. As
with research on employee ownership, there is little research exploring
the specific mechanisms through which profit-sharing increases productivity,
though some studies suggest that profit-sharing increases effort and lowers
voluntary quits, absenteeism and dismissals (see Jones et al., 1997, for a review
of these studies).
In summary, empirical work on group incentive plans such as profit-sharing
and employee ownership indicates that their net effect on productivity is either
positive (as with profit-sharing) or neutral (as with employee ownership). No
evidence exists that such group incentives decrease average productivity. As
far as stock option plans are concerned, an organizational characteristic that has
received little attention is the presence of a union. Therefore, we may ask, "do
stock option plans operate in the same way in unionized as in non-union work-
places?" And, "are the effects of group incentive plans on productivity and
financial performance smaller or larger in unionized than in non-union firms?"
The next section summarizes some theoretical arguments and empirical evidence
on these questions.
UNION EFFECTS ON GROUP INCENTIVES AND

PERFORMANCE: THEORY A N D E V I D E N C E
Whether a stock option plan improves performance depends on individual

employee perceptions of the instrumentality of the plan, individual employee
decisions to put forth extra effort, and individual employee access
to organizational resources required to get the job done. In addition, in a
unionized setting, the SO plan would also have to be backed by (or at least
not opposed by) the union for potential positive effects of the plan to
materialize.
Do unions reduce or increase the potential effects of stock option plans on
employee motivation and performance? Unions may perceive any effort to align
the interests of employees with those of managers/owners via stock ownership
as a threat to their raison d'etre. Therefore, even if unions agree to the
inclusion of a stock option plan in the compensation package, they might not
support its implementation, thereby reducing its motivational impact. Lack of
trust between union and management may increase the transaction costs of
implementing the plan. Also, unions may tend to frame stock option plans as
added benefits, substitutes for wage concessions, or mechanisms to maintain
fixed industry wage patterns (Zalusky, 1990), rather than as incentives to
improve performance. For example, an empirical study of group incentives in
manufacturing firms in the Midwest (Cooke, 1994) found that group incentives
such as profit-sharing and gain-sharing plans had modest positive cross-sectional
effects on productivity in unionized finns, but more substantial positive effects
in nonunion firms.
On the other hand, there are arguments suggesting a favorable union posture
towards adopting and implementing stock option plans. Unions' struggles
during the last several decades have been well documented as they have lost
membership, bargaining power and political clout (Strauss et al., 1991). But
during the 1990s, unions attempted to fight back and reestablish themselves as
a valuable partner in the employee-employer relationship. As part of an effort
to "add value to business" (Masters & Atkin, 1999), unions have been
experimenting with designing and implementing high-performance workplace
practices as well as workplace democracy initiatives that enhance employee
involvement and commitment. Stock option plans are well suited to reward
union members engaging in such practices and initiatives. Union involvement

in the design and implementation of a stock option plan may result in better
communication and understanding of the benefits of the plan compared to
non-union contexts. Such improved understanding may then, in turn, have
positive effects on employee motivation and performance. Another important
priority for unions in recent years has been recruiting new members (Masters
& Atkin, 1999). To attract workers from the hi-tech sector, in particular, unions
have had to go beyond traditional notions of pay, including by supporting stock
option plans as a new form of employee compensation. 13
Empirical evidence on union acceptance of "alternative" rewards such as
skill based pay, profit-sharing, gainsharing, and stock ownership also suggests
that unions have recently come to view stock option plans favorably. Heneman,
von Hippel, Eskew and Greenberg (1997) found that unions favor incentive
plans with objective, group-based performance measures as well as plans
in which employees receive equal payouts. Broad-based employee stock option
plans can easily accommodate these preferences: unionized employees can
be awarded an equal number of options contingent on a group financial or
operational target. Furthermore, over time, unions have been able to overcome
their historic opposition to employee ownership 14 (McElrath & Rowen, 1992);
in fact, today, stock sharing plans are especially prevalent in unionized firms. ]5
In sum, there are arguments favoring both positive and negative effects of
stock option plans on employee motivation, effort and firm performance in
unionized firms. But, whether any dominant effect will prevail is a matter for
empirical investigation.
P U R P O S E OF THE STUDY
The goal of this study is to measure the net effect of SO plans on employee
productivity and financial performance in union and non-union firms. We build
in two ways on prior empirical work that has evaluated the effect of group
incentives on performance: ~6 first, we examine the effects of a highly popular,
yet little studied type of group incentive plan, namely, broad based employee
stock options; t7 second, we analyze such plans in both union and non-union
companies. Following Cooke (1994), we compare the effect of group based
incentives on productivity in samples of union and non-union companies. In
addition, financial performance measures are included in the empirical analysis,
and our data set contains firms from a broad range of industries (not just
manufacturing). Further, we go beyond cross-sectional research to conduct
a longitudinal analysis of the effects of SO plans on productivity and
performance.
The limitations of our data (for example, we use sales per employee as a
proxy for productivity) do not allow us to test hypotheses about the direct effects
of stock option plans on motivation. It may be the case that employees do
indeed perceive stock options as a valuable reward, and therefore increase their
effort to obtain this reward; nonetheless, circumstances beyond employee control
(as examples, poor management and lack of resources to get the job done) may
counteract such an increase in effort. Because of this problem, a cross-sectional
analysis using sales/employee as a proxy for employee effort is likely to
underestimate the effect of SO plans. Another limitation on our analysis is that
we do not know whether employees frame the SO plan as a reward or a
trade-off. As mentioned above, incentives are motivating when the reward
is perceived as valuable; hence, if in some of the companies in our sample
employees perceive the SO plan as a concession, this would also result in
an underestimation of the effects of SO on employee effort. Furthermore,
when assessing the differential effects of SO plans on finn performance in
union and non-union companies, unmeasured variables, such as the quality
of the union-management relationship or the union's support of alternative
compensation practices, may result in under- or over-estimating such differences.
Table 1. Distribution of Broad-Based Stock Option Plans by Industry,

1997.
Percent of all Firms Percent of Unionized Percent of Non-union

Unionized Within Firms with SO Finns with SO
Each Industry Within Each Industry Within Each Industry
Industry
Agriculture & Mining 52% 0% 3%
Construction 17% 0% 2%
Manufacturing 37% 5% 14%
Trans. & Comm. 43% 4% 1%
Utilities 55% 8% 10%
Wholesale 23% 0% 10%
Retail 15% 2% 5%
Fin., Ins. & Real Estate 15% 6% 5%
Services 16% 4% 15%
All Industries Combined 31% 5% 11%
N 3691 1246 2715
Source: Compustat data merged with data from IRS's Form 5500 to determine union status and
NCEO's broad-based stock option data base to determine presence of SO plan. Only compustat
companies that provided employment data for either 1996 or 1997 are included.
DATA SETS AND VARIABLES
The primary data set used in this study is Standard & Poor's Compustat. Finns
that reported employment levels in either 1996 or 1997 were included in the data
set. 18 This yielded an initial list of 8,152 publicly held companies. Compustat
files provide extensive standardized financial information on public companies,
but do not provide information on unionization. Therefore, we used Internal
Revenue Service (1RS) Form 5500 to determine the union status of all firms in
the study. 19 A company was considered unionized if one or more of the pension
plans it sponsors was collectively bargained. Form 5500 data allowed us
unambiguously to establish union status for 4,173 of the initial 8,152 Compustat
companies. 2° To determine whether a company sponsors a broad-based stock
option plan, we used a list of 1,360 companies provided by the National Center
for Employee Ownership (NCEO); 2~ we were able to match 493 of these 1,360
companies to Compustat data. Merging information from all data sets
(Compustat, Form 5500, and the NCEO list) resulted in a total sample size of
3,961 publicly owned firms (with a total of about 36 million employees) from
nine industry sectors. Of these, 1,246 are union firms and 2,715 are non-union
firms. Of the 1,246 union firms, 60 had a SO plan; of the 2,715 non-union finns,
309 had a SO plan (see Table 1 and the Results section below).
We divided the 3,961 companies in our sample into four categories
based on union status and presence of a SO plan: unionized finns with SO
plans; unionized firms without SO plans; non-union firms with SO plans;
and, non-union firms without SO plans. We then compared the economic and
financial performance of the four union/SO groups using a set of four perfor-
mance indicators. Productivity was used as a proxy for employee effort, with
the specific measure being the natural logarithm of sales per employee, adjusted
for inventory change. Financial performance was assessed via two market-based
measures, namely total annual shareholder return (TSR) and Tobin's q ((market
value + preferred stock + long term debt)/(capital stock + current assets -
current liabilities)), and by one accounting measure, namely, return on assets
(ROA). Because we used an augmented production function for our analysis,
our independent variables included dummies for union/SO status, labor, capital
intensity and industry (see the Appendix for variable definitions and Table 2
for descriptive statistics on all variables).
MODEL SPECIFICATIONS
We perform two sets of empirical analyses using four different measures of

firm performance for each set. The first set uses a cross-sectional specification that
78 M . K . K R O U M O V A ET AL.
od
O
e~
© oeL
"o
-6
o t"-
o
Ox
O~
t~
°= r- ~ oi
~D eq '.D
t:h
O
L)
c~
~D .-g
©
"O
O5
eq
",=
t'-
O~
O~
0~
¢:h
tD 5~ ©
eq ~ r-
e,i i
it- o
o
Table 3. U n i o n Status, Stock Options Plans and 1997 P e r f o r m a n c e Levels.
Dependent variables: Ln (output/employee) Total Shareholder Tobin's Q Return on

Return Assets
Independent variables (1) (2) (3) (4)
Union Stock Option 0.287 *** -0.280 0.630"** 1.540

(0.076) (6.298) (0.217) (2.136)
Union No Stock Option 0.038 -2.657 -0.142"* -1.732"**
(0.023) (2.066) (0.069) (0.651 )
No Union Stock Options 0.355 *** -8.519 *** 0.613 *** 0.375
(0.035 (2.879) (0.100) (0.986)
Ln (total employment) -0.369*** 5.960 *** 0.431"** 5.640***
(0.010) (0.883) (0.030) (0.287)
Ln (net assets) 0.318 -0.474 -0.324 -2.823
(0.008) (0.727) (0.024) (0.234)
Intercept 3.642 *** 11.015 * 3.835*** 23.655***
(0.070) (6.010) (0.202) (1.952)
F 242.23 17.66 754.44 914.95
2-digit industry dummies Yes Yes Yes Yes
n 3456 3160 3295 3464
Based on robust regressions that minimize influence of outliers.

Standard errors in parentheses. * p < 0.10 ** p < 0.05 *** p < 0.01.
The excluded group is No Union*No Stock Option.
Table 3A. Estimated Proportional Differences in Productivity by Firms

Categories.
Union*Stock Option 33.3%

Union*No Stock Option 3.9%
No Union*Stock Option 42.6%
*Exponential functions of coefficients reported in Table 3: [(exp(Bi)-l)*100]

(The refernce group is No Union*No Stock Option).
captures the net differences in firm p e r f o r m a n c e a m o n g the four u n i o n / S O cate-

gories. Results f r o m these analyses are presented in Table 3. Cross-sectional esti-
mates are vulnerable to a n u m b e r o f estimation problems, such as self-
selection and omitted variable bias. G i v e n the limitations o f our data set, w e
attempt partially to correct for s o m e o f the biases associated with cross-sectional
estimation by also conducting a longitudinal analysis. Thus, our second set o f
empirical analyses c o m p a r e s the average performance and the average differences
in p e r f o r m a n c e a m o n g the four u n i o n / S O categories in the periods b e f o r e and after
adoption o f an S O plan. 22 Results f r o m these analyses are reported in T a b l e 4.
80 M . K . K R O U M O V A E T AL.
Table 4. U n i o n S t a t u s , S O P l a n s , a n d C h a n g e s in P e r f o r m a n c e L e v e l s ,
1985-1987 to 1 9 9 5 - 1 9 9 7 .
Dependent Variables Ln(Output/ Total Share- Tobin's Q Return On

Employee) holder Return Assets
Independent variables (1) (2) (3) (4)
All companies w/data in both periods^

1985-1987 0.094 ** 4.690 0.329*** 1.329
(0.045) (3.287) (0.085) (1.154)
1995-1997 0.213 *** 4.840 0.553*** 3.355***
(0.046) (3.641) (0.089) (1.198)
Union-No Stock Option
1985-1987 0.081 *** 6.743*** -0.125"** -0.519
(0.017) (1.278) (0.033) (0.438)
1995-1997 0.088 1.202 -0.181 -0.041
(0.021) (1.637) (0.041) (0.525)
No Union-Stock Option
1985-1987 0.157"** 2.154 0.432*** 0.939
(0.030) (2.189) (0.057) (0.760)
1995-1997 0.387*** 6.535*** 0.644*** 5.313
(0.033) (2.449) (0.062) (0.814)
No Union-No Stock Option
1985-1987
1995-1997 0.044*** 4.397*** 0.089*** 1.140"**

(0.017) (1.313) (0.033) (0.409)
Change from 1985-1987 to 1995-1997 ^^
Union Stock Option 0.119 * 0.150 0.224 * 2.026
(3.42) (0.00) (3.37) (1.49)
Union-No Stock Option 0.007 -5.542*** -0.056 0.478
(0.06) (7.33) (1.15) (0.50)
No Union Stock Option 0.230*** 4.381 0.212"** 4.374***
(27.12) (1.80) (6.40) (15.57)
No Union No Stock Option 0.044*** 4.397*** -0.089*** 1.140"**
Difference Union SO-Union NSO 0.112 * 5.692 .280** 1.548

(3.19) (1.53) (5.57) (0.91)
Difference Union SO-No Union SO -0.111 -4.231 0.013 -2.348
(2.32) (0.63) (0.01) (1.59)
F 543.8 63.7 136.75 118.59
Total Observations 13210 11565 1619 14206
Notes: Standard Errors in Parentheses. * p < 0.10 ** p < 0.05 *** p < 0.01.
The excluded group is No Union No Stock Option in 85-87.
^ Based on robust regressions run on all companies with complete data for the 1985-1987 and
1995-1997 periods and In(assets) plus year dummies and 2-digit industry dummies.
^^ T-statistics (not standard errors) reported in parentheses in these companies.
Broad-Based Employee Stock Options - A Union-Non- Union Comparison 81
Following prior empirical work assessing the net effect of group incentives
on performance, we use an augmented production function for our cross-
sectional productivity analysis (Kruse, 1992, 1993; Blasi et al., 1996; Jones et
al., 1997). We use the same model when financial performance rather than
productivity is the dependent variable.
Our first model is cross-sectional. Firm performance in 1997 is modeled as
follows:
Model (1)
P e r f i = b o + b 1U S O i + b 2 U N S O i + b 3 N U S O i + b 4 1 n ( L ) i + b s l n ( K ) i + b 6 _ 1 3 ( i n d u s t r y
d u m m i e s ) i +e i
where Perf = In(output/employee), total shareholder return, return on assets,

and Tobin's q in 1997
USO = dummy for union, stock option plan in 1997
UNSO = dummy for union, no stock option plan in 1997
NUSO = dummy for non-union, stock option plan in 1997
in(L) = In(number of employees) in 1997
In(K) = In(capital intensity) in 1997
e i = error term assumed normally distributed i.i.d.
b i = coefficients.
As noted above, cross-sectional regression results are susceptible to omitted
variable bias. It may be the case that factors (for example, quality of the union-
management relationship and quality of the workforce) other than the ones we
measured affect performance and are non-randomly distributed across union/SO
status. Also as noted, cross-sectional coefficients may be biased due to
self-selection problems: firms in our sample have selected themselves into
union/SO categories, and the selection rule may be related to the outcome being
measured. In other words, firms with preexisting high productivity may be
particularly likely to adopt SO plans. If so, then even if a cross-sectional analysis
finds that SO firms are more productive, such higher productivity may not be
due to the adoption and presence of SO plans (Kruse, 1993). One simple method
to correct for this type of self-selection bias is to compare pre-adoption and
post-adoption firm performance levels. Unfortunately, we did not have plan
adoption dates for the majority of SO companies in our data set, and so could
not use panel data to estimate the difference between pre- and post-
adoption performance. 23 We did, however, have information on plan start dates
for a limited number of SO companies. We obtained this information from a
survey of broad-based U.S. SO companies conducted by the NCEO in 1998.
The survey had a response rate of approximately ten percent and contained a
total of 133 SO companies. Of these, 93 provided plan start dates, only 10 of

which were identified as union companies. Eighty-five percent of these 93
companies had adopted their broad-based SO plans during or after 1989.
Consequently, in the absence of information on plan start dates for the majority
of companies in our sample, we compare mid-1980s company performance with
mid-1990s company performance as a proxy for a comparison of pre- and post-
adoption performance. This approach involves some measurement error, as a
small number of SO plans may have been adopted during the early 1980s or
before. Also, such a comparison does not establish causality since it may be
the case that other important changes (as examples, restructuring, downsizing
and changes in management) occurred simultaneously with SO plan adoption.
Nonetheless, comparing average performance in the 1985-1987 to average
performance in the 1995-1997 period within and across union/SO groups
(Model 2) allows for some useful insights about the relative changes in perfor-
mance experienced by our 4 categories of firms. The longitudinal specification
used is as follows:
Model (2)
Perfit = b0+b 1[USO *(85-87)] +bz[USO*(95-97)] +b3[ UNSO *(85-87)] +
b 4 [ UNSO * ( 9 5 - 9 7 ) ] ÷bs[NUSO * ( 8 5 - 8 7 ) + b 6 [ N U S O * ( 9 5 - 9 7 ) ] +
bT[ N U N S O *( 9 5 - 9 7 ) + b 81n( L )+bg[ln( L ) *( 9 5 - 9 7 ) ]+b loln( K) +
btl[ln(K)*(95-97)] +industry+year+ eit
where Perf = In(output/employee), total shareholder return, return on assets,
and Tobin's q in 1985, 86, 87, 95, 96, and 97
USO = dummy for union, stock option plan
UNSO = dummy for union, no stock option plan
NUSO = dummy for non-union, stock option plan
85-87 = period dummy for 1985-1987
95-97 = period dummy for 1995-1997
In(L) = In(number of employees) in 1997
In(K) = In(capital intensity) in 1997
eit -- error term assumed normally distributed i.i.d.
b i = coefficients.
In Model (2), the excluded group is No Union-No Stock Option Plan in
1985-1987. The coefficient b 1 indicates the average adjusted difference in
performance between Union-Stock Option companies in 1985-1987 and No
Union-No Stock Option Plan companies (i.e. the excluded group) in 1985-1987.
Comparisons of differences in performance among union/SO status groups
over time (e.g. comparing average performance in 1985-1987 to average
performance in 1995-1997) are obtained by subtracting the respective

coefficients. To illustrate, the average change in performance for Union-Stock
Option companies between 1985-1987 and 1995-1997 is obtained by
subtracting b I from b 2 (see Table 4). To obtain an estimate of the difference
between average changes in performance across time periods for Union-Stock
Option companies and Union-No Stock Option companies, we use [(b 2 - b j)
- (b a - b3)] (see Table 4).
RESULTS
Table 1 provides information on the industry distribution of the companies in

the combined data set and the prevalence of stock option plans within each
industry for both union and non-union companies. About 30% of the 3,961
finns in our sample had union workers; 24 about 11% of the 2,715 non-union
companies had a broad-based SO plan compared to 5% of the 1,246 compa-
nies that had union employees, suggesting that SO plans are more prevalent in
non-union firms. That trend held across all major industry groups with the
exception of Finance, Insurance and Real Estate.
Descriptive statistics and simple mean comparisons for all variables are
provided in Table 2. In 1997, there were 60 unionized SO firms and 309 non-
union SO finns. Unionized firms with SO plans were larger than other finns
in terms of both sales and employees, and also had higher capital intensity
compared to all the other categories of firms. Hence, the results from the empir-
ical analysis below should be interpreted as applying primarily to large publicly
owned finns with union employees. 25 These firms also had higher productivity,
ROA, Tobin's q and TSR levels in 1997 than the three other categories of
finns. Note that non-union firms without stock option plans were considerably
smaller than the three other categories of firms.
Table 3 reports cross-sectional results comparing performance across
union-SO status controlling for capital intensity, firm size and industry charac-
teristics in 1997. The reference category for the estimates reported in this table
is nonunion firms without stock option plans (NU-NSO). To minimize the
influence of outliers, robust regression was used in estimating all of the
equations.
Productivity, an outcome that is theoretically more directly affected by
employee effort than market and accounting based measures of firm perfor-
mance, was significantly higher for firms with SO plans irrespective of union
status (these results are consistent with Sesil et al., 2000). Table 3A indicates
that productivity was about 33% higher in union firms with SO plans compared
to the reference group (namely, NU-NSO firms); the productivity difference
was even larger - 43% - between non-union firms with SO plans and
non-union firms without SO plans. 26 By contrast, the difference in productivity
between union firms without SO plans and non-union firms without SO plans
was small-about 4%-and statistically insignificant.
Tobin's q, a measure reflecting the present value of a firm's future profit
stream (and a proxy for the nontangible value of the firm), was also
significantly higher for stock option firms than for non-stock option firms
irrespective of union status. 27 These results imply that financial markets
evaluate the future prospects of SO firms more positively than the prospects of
non-SO firms; the fact that a firm has unionized workers does not alter that
evaluation. Conversely, union non-stock option firms had a Tobin's q that was
significantly lower than all other groups, including the baseline group, implying
that financial markets have low expectations about the future profit streams of
union, non-stock option firms. A plausible explanation of these results is that
the existence of a SO plan is interpreted by financial markets as a proxy for
"progressive" management and good employee relations in both union and
non-union firms. For union firms, the absence of SO plans may be interpreted
by investors as a proxy for traditional management with poor employee
relations.
The cross-sectional results for our second measure of market-based firm
performance, TSR, differ across union and non-union SO firms. TSR at union
SO firms was slightly but not significantly different from TSR in the reference
group (NU-NSO) of firms. However, TSR was 8.5% lower in non-union SO
firms than in the reference group, and the difference is significant at p(L)0.01.
Why do stock option plans appear to have a net negative or no effect on TSR,
as opposed to their positive effects on productivity and Tobin's q? One possible
explanation is that stock options entail an economic cost - the firm either issues
new shares and effectively sells these at a discount to employees or has to buy
back its own stock. This cost could result in diluted shareholder earnings if
productivity does not improve or if the employees rather than the shareholders
capture the benefits from improved productivity. However, before interpreting
the above results as evidence that employee stock options reduce shareholder
returns, it should be noted that TSR in any single year is a very "noisy" measure
of firm performance-more so than the three other performance measures used
here. 28 Consequently it is more informative to look at the pre-post comparison
results presented in Table 4, since they average TSR over two 3-year periods. 29
The accounting measure of performance used in this study, ROA, suggests
that both union and non-union stock option companies do not perform worse
than the baseline group (also see endnote 28 for 1995 and 1996 cross-sectional
ROA results). On balance, the empirical evidence suggests that there is no
significant relationship between the presence of a SO plan and ROA (in 1997).
Consistent with prior research, union firms without stock option plans have a
lower ROA compared to non-union firms, suggesting that unions may be shifting
returns from capital to labor (see Bell & Neumark, 1993).
Were performance levels higher or lower for SO firms before plan adoption?
Did performance improve after SO plan adoption? Table 4 reports pre- and
post-adoption comparisons evaluating differences within and across union/SO
groups between two time periods - 1985-1987 and 1995-1997. The excluded
group is non-union, non-stock option firms in the 1985-1987 period. The results
indicate that productivity in the pre-adoption period was higher in SO firms
than in the reference group, both in union and non-union firms. Productivity in
these firms continued to be higher in the 1995-1997 period compared to the
reference group. More important, between 1985-1987 and 1995-1997 prod-
uctivity increased for both union stock option firms and non-union stock option
firms, and the increases were relatively large (0.119 and 0.23, respectively)
and significant (see Table 4, "Change from 1985-1987 to 1995-1997"). The
productivity increase was smaller in union SO firms than in non-union SO firms
(-0.111), but the difference was not significant. Thus, these results not only
support the cross-sectional evidence in Table 3, they also suggest that companies
that adopted SO plans had higher productivity levels before plan adoption
compared to NSO firms and experienced a significant increase in productivity
after SO plan adoption. In contrast, union NSO firms were more productive
than non-union NSO firms during the mid-1980s, but their productivity did not
increase over time.
The same pattern of results was found for Tobin's q, namely, levels
were higher for both union and non-union SO companies before adoption, they
remained higher after adoption, and there was a significant increase in
Tobin's q from pre- to post-adoption. By contrast, for union NSO companies,
Tobin's q was significantly lower than in the reference group of firms during
the mid-1980s, and there was no increase over time. Thus, financial markets
had lower expectations about the future revenue streams that union NSO
companies might expect in the mid-1980s, and these lower expectations
persisted over time.
As for TSR, all three groups of firms had somewhat higher returns compared
to the non-union NSO firms, but only the difference between union non-stock
option and non-union non-stock option firms was statistically significant.
TSR did not change significantly from pre-adoption to post-adoption for either
union or non-union stock option companies; it decreased for union
non-stock option firms and increased for non-union, non-stock option firms.
Union NSO companies exhibited a significant decline in TSR over the period
from the mid-1980s to the mid-1990s. However, these results should be inter-
preted as suggestive since additional analysis using risk-adjusted monthly (rather
than yearly) TSR data is needed (see note 29).
Pre-post comparisons for ROA indicate that ROA increased significantly from
the mid-1980s to the mid-1990s for non-union firms (the increase was larger
for stock option finns, as shown in Table 4). Over the same period, ROA did
not change significantly for union firms irrespective of their SO status. These
results again support the cross-sectional evidence presented in Table 3, and also
suggest that, for union SO companies, better productivity did not result in higher
ROA (or TSR). Thus, it may be the case that returns are indeed shifted from
capital to labor in these firms. In union NSO finns, ROA is consistently lower
throughout the studied period than in non-union firms.
To summarize, the cross-sectional analysis suggests that productivity and
Tobin's q are significantly higher in stock option finns compared to non-stock
option firms; this holds true for both union and non-union stock option firms.
Moreover, the size of the effect appears to be similar for stock option firms
across union status. The longitudinal analysis indicates that for stock option
finns, both union and non-union, levels of productivity and Tobin's q were
higher before adoption, increased between the pre-adoption and post-adoption
periods, and remained higher in the post-adoption period. Consequently, union
and non-union SO finns outperformed the two other groups of finns based on
these measures. However, TSR did not improve over time in SO union and
non-union firms, while it increased significantly in non-union NSO finns. And,
the adoption and presence of a SO plan had different effects on ROA in union
and non-union firms. For union firms, ROA did not increase after SO plan
adoption, whereas ROA increased significantly for non-union SO finns.
In addition, over time union non-stock option finns fared worse than both
non-union finns and union stock option firms on all four performance measures
- that is, their productivity, Tobin's q, and ROA did not increase over time,
and their TSR declined significantly.
Our results have some important limitations. All the union stock option finns
in our sample were large (90% of them had more than 3,000 employees and
over half had more than 24,000 employees). Hence, our results apply primarily
to large, publicly held union companies. Also, the lack of panel data including
exact stock option plan adoption dates for union and non-union finns prevents
us from using a true panel data set to implement two-stage modeling. Thus, it
may be the case that characteristics (such as superior management and higher
quality human capital) that prompt companies to adopt SO plans are also the
characteristics that make companies more productive. Though we attempted to
address this issue with our pre-post analysis, new data and further research are
required better to address both of these problems. Also, our measures of union
status and TSR were rather unsophisticated; it would be especially informative
to repeat this analysis using a within-firm measure of union density rather than
a simple dummy variable, and using monthly, risk-adjusted data for the TSR
comparisons.
CONCLUSION
This empirical study evaluated the net effects of employee stock option plans
across union status on economic and financial measures of firm performance.
The data set contained 3,961 public companies and was obtained by merging
company information from several sources: Standard & Poor's Compustat, IRS's
Form 5500, and an NCEO database. The analysis produced fairly strong
evidence that broad-based employee stock option plans yield substantial gains
in firm-level productivity and Tobin's q. Furthermore, the net effect of stock
option plans on productivity and Tobin's q was found to be similar in union
and non-union finns. In addition, union stock option firms consistently out-
performed union non-stock option firms with respect to these two measures of
performance. Our results for productivity are consistent with Cooke
(1994), who found that both union and non-union manufacturing firms with
group incentive plans outperform non-union firms with no group incentives, the
magnitude of the effect being 18-21%. Results for total shareholder return
were less clear-cut, suggesting that SO plans might have an economic cost
associated with them. As for return on assets, non-union SO companies
improved their ROA after adoption, whereas union SO companies did not.
Thus, even though productivity increased over time for both union and
non-union SO companies, this increase resulted in higher operating income
per unit of assets only for non-union firms, raising the possibility that the
gains from improved productivity in union firms were captured by the workers
rather than the firm.
In summary, on the positive side, the empirical evidence implies that stock
option plans are associated with a net positive effect on productivity, and union
and non-union firms capture the potential incentive effects of these plans on
productivity equally well. Thus, the results alleviate fears that contemporary
group incentives, such as stock options, do not function well in unionized
settings. Also, financial markets appear to value the future prospects of SO
firms highly, perhaps because SO plans are used as a proxy to indicate better,
performance-aligned management practices. On the other hand, increased
productivity did not result in higher annual shareholder returns for SO firms
(irrespective of union status), and ROA increased after SO adoption only in
non-union companies. These findings raise the question of how the economic
gains from incentive plans are captured within union and non-union firms, and
which stakeholders benefit from such economic gains. Further research and
better data are needed to explore this question.
We believe the analysis presented here provides useful information on a topic
that has been little explored - stock option plans and their impact on finn
performance in union and non-union settings. As with any other research
endeavor, many interesting questions remain unanswered. Specifically, why do
companies adopt employee stock option plans, and do union firms adopt such
plans for different reasons than non-union finns? How are stock option plans
structured (that is, are options awarded based on merit, tenure, percent of pay,
and/or other factors?), and do union finns structure their plans differently (e.g.
more equitably or bureaucratically) than non-union firms? How well do
employees understand the functioning of their stock option plan?, how do they
frame it (as a valuable reward, an unimportant add-on, or as a way to reduce
their base pay rate)?, and are there "understanding" and "framing" differences
between union and non-union employees/finns? To address these questions,
panel data for large representative samples of all four types of firms studied
here are needed-data that contain information on plan adoption dates, plan
design, and other human resource management practices.
We hope this paper will stimulate further interest in this area and will
encourage researchers to develop greater understanding of how stock options
interact with other key factors (including other human resource management
practices) to effect employee effort and finn performance.
NOTES
1. Available at: www.nceo.org

2. We define broad-based plans as plans that offer stock options to 50% or more of
all employees. For simplicity, broad-based SO plans are referred to as SO plans in the
remainder of this paper.
3. See Huddart & Lang (1996) for an overview of the accounting and tax treatment
of employee stock options.
4. A study by Coopers & Lybrand (1993) estimated that if companies reported the
value of options granted the average reduction in net company income after the phase-
in period would have been 3.4% for mature companies and 26.5% for emerging
companies.
5. Recent examples of union employees receiving stock options include Harley
Davidson and Verizon.
6. Assuming rational, self-interested economic actors.
7. For an in-depth review of the existing theory and empirical findings on piece
rates, executive compensation, merit pay, and commission systems, see Asch & Warner,
1997.
8. Stock options provide a gain to employees only if the stock price increases between
the grant and exercise dates; thus, both employees and shareholders benefit from the
price increase. If the stock price falls below the grant price, shareholders experience a
loss (the value of their investment has decreased over the period in question) whereas
employees receive no gain. Some argue that this reduces the incentive power of options
(compared to, say, stock ownership). Nonetheless, it may be the case that employees
expect to receive a financial gain from their SO plan (a quite reasonable assumption
during the rising market of the 1990s). Hence, if their initial frame of reference is an
expected gain, receiving nothing would be experienced psychologically as a loss.
9. Expectancy theory (Vroom, 1964) predicts that a reward will elicit effort from
employees if the reward is valued, the probability of obtaining the reward given perfor-
mance targets are met is high, and the probability that effort will result in meeting
performance targets is high. It is reasonable to assume that stock options provide a
valued reward. But the relationship between meeting performance targets and stock price
is fraught with uncertainty, and depends on much more than employee effort. Also,
whether increased individual effort will result in improved group/unit/company perfor-
mance is likely to depend on a company's financial and human resource management
systems, its culture, and other factors.
10. Gomez-Mejia and Balkin (1992) report that about 18% of the variance in firm
performance can be attributed to managerial decisions in response to the incentive struc-
ture; the remainder can be attributed to unsystematic factors.
11. This logic implies that the larger the firm/workplace size, the more difficult it
is for management to monitor employees and for employees to monitor the output of
their co-workers. Hence the need for incentive compensation, either individual or
group, is greater in large firms (e.g. see Brown, 1990, or Drago & Heywood, 1995).
But, by the same token, such plans may be less effective at eliciting effort in large
firms/work units. These conflicting dynamics may explain the mixed empirical
evidence on the relationship between size and profit-sharing effectiveness (see Jones,
Kato & Pliskin, 1997).
12. A potentially important difference between these types of plans and broad-
based SO plans is that, under the latter, employees have somewhat greater control
over when to realize the reward (i.e. when to exercise their options once they become
vested). As a result, the timing and size of the reward will vary across individuals.
In as much as reward timing and size affect future behavior, however, the behavioral
effect of such plans is less consistent than the effect of either profit or gain sharing.
Also, SO plans encourage employees to monitor the market performance of the firm
more closely than any other group incentive plans, at least during the period before
exercise/selling.
13. The recent strike against Verizon Communications illustrates this point. In August
2000, the Communications Workers of America and the International Brotherhood of
Electrical Workers fought for, and won, concessions making it easier for them to orga-
nize the wireless part of the business. The agreement with the company also included
the establishment of a profit-sharing plan and the allocation of stock options to union
members. (Romero, S., "Labor Accord Hits New-Economy Notes", New York Times,
August 22, 2000, Section C; Page 1; Column 4).
14. For an overview of unions' changing attitudes toward employee ownership, see
McElrath and Rowan (1992). They argue that economic necessity, lack of evidence of
a negative impact of ESOPs on earnings and employment, and realization that ESOPs
can be used to achieve strategic union objectives have prompted unions to overcome
their opposition to employee ownership.
15. In a survey of 150 union and 350 non-union firms, Heneman et al. (1997) found
that union firms are more likely than non-union firms to use stock sharing plans.
16. For a review of prior empirical work on group incentives, see section two of this
paper "Incentives and Group Incentives: Theory and Evidence".
17. For an exception, see Core and Guay (2000). Their paper addresses the question
of why companies adopt non-executive employee stock option plans (although their data
do not distinguish between broad-based and other types of employee stock option plans).
They find that firms use greater stock option compensation when facing high capital
requirements and financial constraints. The results from their empirical analysis also
suggest that firms may be using options to attract certain types of employees, provide
retention incentives, and create incentives to increase firm value (though the authors do
not directly test any of these three propositions).
18. At the time this project was started, 1997 Compustat data was the most recent
available to the authors. We used all annual Industrial and Full Coverage Files. These
files contain all companies listed on the New York Exchange, American Exchange and
NASDAQ, companies listed on regional exchanges, publicly held companies trading
common stock, and wholly owned subsidiaries trading preferred stock or debt. We also
included companies from the annual Industrial Research File This file contains compa-
nies that have been deleted from the Industrial Files due to bankruptcy, acquisition or
merger, leveraged buyout, or because they became private companies.
19. IRS's Form 5500 is a tax form that all U.S. pension plans in establishments with
100 employees or more must file every year. We used data on 66,091 establishments
from 1995 (the latest available in electronic format) to determine union status in 1997.
This data base is made available by the U.S. Department of Labor.
20. We used employer identification numbers (EIN) to merge records from the two
data sets.
21. In 1998, the NCEO compiled a comprehensive list of 1,360 companies (both
public and private) sponsoring broad-based SO plans based on a national clipping service
on stock compensation, regular reviews of company announcements, and information
provided by various consulting firms and practitioners. Out of these 1,360, we were able
to match 493 to Compustat data.
22. Unfortunately, we did not have plan adoption dates for most of the SO compa-
nies in our data set. Based on the limited information on plan adoption dates, we
developed Model (2). See the Methods section above for a detailed explanation of the
assumptions we made in developing Model (2).
23. The lack of plan start dates also limits our ability to use a two-stage model that
would estimate the probability of adopting an SO plan as a first stage, and then use
these estimates to explain firm performance in the second stage.
24. Since our measure of unionization was derived from Form 5500 data, we do
not know the actual percent of workers within each company that belongs to a union.
Also, because we combined data from several data sets, yet our primary source is
Compustat, our final sample is composed only of public companies. Further, because
of the merging procedure, only public companies that provided their correct employer
identification number and information on collective bargaining in Form 5500 are

included.
25. To explore the possibility that the effect of SO plans varies with company size,
we fitted model (1) using various employment size breakdowns. For union SO compa-
nies, the results for productivity were stable and similar to the ones reported in Table
3; the only difference occurred for companies with more than 50,000 employees.
Specifically, the productivity effect was still positive but had shrunk to 7% (compared
to the 33% reported in Table 3). The difference between non-union SO and non-union
NSO companies also diminished for the biggest companies, but nevertheless remained
large (30% compared to the 43% reported in Table 3).
26. The difference between the dummy coefficient for union/SO status and non-
union/SO status was not statistically significant; in other words, there is no evidence
that the net effect of stock option plans on productivity is different between union and
non-union firms.
27. As with productivity, the effect sizes for union stock option and non-union stock
option finns with respect to Tobin's q were similar in size, and the difference between
these effect sizes was not statistically significant across union status. Hence, there is no
evidence that the net effect of stock option plans on Tobin's q is different between union
and non-union firms.
28. We estimated the four cross-sectional equations for 1995 and 1996 and compared
the results to those for 1997 (reported in this paper). The direction, size, and signifi-
cance of the union/SO effects were remarkably similar in all three years for productivity
and Tobin's q; if anything, the effects were slightly larger in 1995 and 1996, compared
to 1997. For ROA, in both 1995 and 1996, SO companies had significantly higher ROA
compared to the excluded group irrespective of union status, and the effect for the non-
union, non-stock option firms is similar to the one estimated for 1997. The effects for
TSR, however, vary widely. In 1996, all union/SO groups had lower TSR compared to
the reference group, but the coefficients were small ( - 1 % to - 4 % ) and statistically
insignificant. In 1995, non-union stock option firms showed an average return 15% higher
than the reference group of firms (as opposed to 8% lower in 1997), and the difference
was highly significant. The average estimated return for union stock option firms in 1995
was 4% higher than the return for the reference group, but the difference was not
statistically significant. Hence, cross-sectional differences in TSR show an unstable
behavior over time, as compared to cross-sectional differences in productivity, Tobin's
q and ROA.
29. It may be difficult to interpret TSR coefficients unambiguously, even for our pre-
post comparisons, because returns are not adjusted for market risk in this analysis
(because Compustat provides only quarterly and yearly stock returns). If market risk is
non-randomly distributed across union/SO status, it will be difficult to interpret the
union/SO coefficients on a non-risk-adjusted basis.
ACKNOWLEDGMENTS
The authors w o u l d like to thank the National Center for E m p l o y e e Ownership

( N C E O ) and their partners for providing the data for this analysis. W e w o u l d
also like to thank the a n o n y m o u s referees.
92 M . K . K R O U M O V A ET AL.
REFERENCES
Alchian, A., & Demsetz, H. (1972). Production, Information Costs, and Economic Organization.
American Economic Review, 62, 777-795.
Asch, B., & Warner, J. (1997). Incentive Systems: Theory and Evidence. In: D. Lewin, D. Mitchell
& M. Zaidi (Eds), The Human Resource Management Handbook, Part 1 (pp. 175-215).
Greenwich, CT: JAI Press.
Bell, L., & Neumark, D. (1993). Lump Sum Payments and Profit-Sharing Plans in the Union Sector
of the United States Economy. Economic Journal, 103, 602-619.
Ben-Ner, A., & Jones, D. (1995). Employee Participation, Ownership, and Productivity: A
Theoretical Framework. Industrial Relations, 34(4), 532-554.
Berle, A., & Means, G. (1932). The Modern Corporation and Private Property. New York: Harcourt,
Brace & World.
Blasi, J., Conte, M., & Kruse, D. (1996). Employee Stock Ownership and Corporate Performance
Among Public Companies. Industrial and Labor Relations Review, 50(1), 60-79.
Blasi, J., Kruse, D., Sesil, J., & Kroumova, M. (2000). Stock Options, Corporate Performance
and Organizational Change. Oakland, CA: National Center for Employee Ownership,
September. (Available online in an expanded version at: http://www.nceo.org/library/
optionreport.html).
Brown, C. (1990). Firm's Choice of Method of Pay. Industrial and Labor Relations Review,
43(Special Issue), 165-S-182-S.
Cooke, W. (1994). Employee Participation Programs, Group-Based Incentives, and Company
Performance: A Union-Nonunion Comparison. Industrial and Labor Relations Review, 47(4),
594--609.
Coopers & Lybrand (1993). Stock Options: Accounting, Valuation, and Management Issues. Coopers
& Lybrand, New York, NY.
Core, J., & Guay, W. (2000). Stock Option Plans for Non-Executive Employees. Manuscript,
Wharton School of Business, University of Pennsylvania.
Drago, R., & Heywood, J. (1994). The Choice of Payment Schemes: Australian Establishment Data.
Industrial Relations, 34(4), 507-531.
Eisenhardt, K. (1989). Agency Theory: An Assessment and Review. Academy of Management
Review, 14(1), 57-74.
FitzRoy, F., & Kraft, K. (1987). Cooperation, Productivity and Profit-Sharing. The Quarterly Journal
of Economics, 102(1), 23-36.
Gomez-Mejia, L., & Balkin, D. (1992). Compensation, Organizational Strategy, and Firm
Performance. South-Western Human Resource Management Series.
Hansen, D. (1997). Worker Performance and Group Incentives: A Case Study. Industrial and Labor
Relations Review, 51(1), 37-49.
Heneman, R., von Hippel, C., Eskew, D., & Greenberger, D. (1997). Alternative Rewards in
Unionized Environments. ACA Journal, 6(2), 42-55.
Huddart, S., & Lang, M. (1996). Employee Stock Option Exercises - An Empirical Analysis.
Journal of Accounting and Economics, 21(1), 5-43.
Jones, D., Kato, T., & Pliskin, J. (1997). Profit-Sharing and Gainsharing: A review of Theory,
Incidence, and Effects. In: The Human Resource Management Handbook (pp. 153-173).
JAI Press
Kang, D., & SOrensen, A. (1999). Ownership Organization and Firm Performance. Annual Review
of Sociology, 25, 121-144.
Kroll, A. (1997). Exploring Options. HRMagazine, 42(10), 96-100.

Kruse, D. (1992). Profit-Sharing and Productivity: Microeconomic Evidence from the United States.
The Economic Journal, 102(January), 24-36.
Kruse, D. (1993). Profit-Sharing: Does It Make a Difference? Kalamazoo, MI: W.E. Upjohn Institute
for Employment Research.
Kruse, D., & Blasi, J. (1997). Employee Ownership, Employee Attitudes, and Finn Performance :
A Review of the Evidence. In: D. Lewin, D. Mitchell & M. Zaidi (Eds), The Human
Resource Management Handbook, Part 1 (pp. 113-151). Greenwich, CT: JAI Press.
Levine, D., & Tyson, L. (1990). Participation, Productivity, and the Finn's Environment. In: A. S.
Blinder (Ed.), Paying for Productivity (pp. 183-243). Washington, D.C.: Brookings
Institution.
Masters, M., & Atkin, R. (1999). Union Strategies for Revival: A Conceptual Framework and
Literature Review. In: Research in Personnel and Human Resource Management (Vol. 17,
pp. 283-314).
McElrath, R., & Rowan, R. (1992). The American Labor Movement and Employee Ownership:
Objections to and Uses of Employee Stock Ownership Plans. Journal of Labor Research,
13(1), 99-119.
Mitchell, D., Lewin, D., & Lawler, E. (1990). Pay Systems, Firm Performance and Productivity.
In: A. Blinder (Ed.), Paying for Productivity. Washington, D.C.: Brookings Institute.
Pierce, J., Rubenfeld, S., & Morgan, S. (1991). Employee Ownership: A Conceptual Model of
Process and Effects. Academy of Management Review, 16(1), 121-144.
Sesil, J., Kroumova, M., Kruse, D., & Blasi, J. (2000). Broad-Based Employee Stock Options in
The U.S.: Company Performance and Characteristics. Academy of Management Conference
Best Papers Proceedings. Toronto, ON.
Strauss, G., Gallagher, D. G., & Fiorito, J. (Eds) (1991). The State of the Unions. Madison, WI:
Industrial Relations Research Association.
Vroom, V. H. (1964). Work and Motivation. New York, NY: Wiley.
Weeden, R., Carberry, E., & Rodrick, S. (1998). Current Practices in Stock Option Plan Design.
Oakland, CA: National Center for Employee Ownership.
Weiss, A. (1987). Incentives and Worker Behavior: Some Evidence. In: H. R. Nalbantian (Ed.),
Incentives, Cooperation, and Risk Sharing (pp. 137-50). Totowa, NJ: Rowman & Littlefield
Weitzman, M., & Kruse, D. (1990). Profit-Sharing and Productivity. In: A. Blinder (Ed.), Paying
for Productivity. Washington, D.C.: Brookings Institute.
Zalusky, J. L. (1990). Labor-Management Relations: Unions View Profit-Sharing. In: M. J. Roomkin
(Ed.), Profit-Sharing and Gain Sharing (pp. 65-78). Metuchen, NJ: Scarecrow Press.
94 M . K . K R O U M O V A E T AL.
APPENDIX
Variable Definition
Union/SO status Four dummy variables for: unionized firms with SO

plans; unionized firms without SO plans; non-union
firms with SO plans; and, non-union firms without
SO plans
Ln (Employment) Natural logarithm of total company employment
(continuous)
Ln (Sales) Natural logarithm of total company sales (continuous)
Ln (Capital Intensity) Total assets divided by total employment (continuous)
Ln (Productivity) Natural logarithm of output per worker (total sales
adjusted for inventory and inflation divided by the
number of employees)
(continuous)
Tobin's q (Market value + preferred stock + long term
debt)/(capital stock + current assets - current
liabilities) (continuous)
Total shareholder return (Adjusted stock price (for stock splits) + adjusted
dividend)/
(adjusted price It-1]) (continuous)
Return on assets [(Operating income-adjusted depreciation)*100]/(net
adjusted value of capital stock + current assets -
current liabilities)
Industry controls 2-digit industry codes (dummy variable)
CONSENTING TO BE GOVERNED:
UNION TRANSFORMATION A N D
TEAMSTER DEMOCRACY
Robert Bruno
INTRODUCTION
In their efforts to define and assess union democracy, researchers have princi-
pally relied on a "legalistic" perspective (Taft, 1948; Edelstein & Warner, 1976;
Stephan-Norris, 1998), which focuses on formal constitutional measures and a
"behavioral" perspective (Lipset, Trow & Coleman, 1956; McConnell, 1958;
Martin, 1968; Nyden, 1985; Stephan-Norris & Zeitlin, 1992; Needlemen, 1998),
which addresses the effects of internal parties competing for the support of the
rank-and-file by projecting differing political-economic ideologies.
While these two frameworks contribute substantially to conceptualizing union
democracy, neither of them measures the quality of intemal governance by
assessing what rank-and-file members actually think about how they are
governed. The empirical literature on union democracy typically addresses
constitutional structures and leadership behavior, but rarely addresses or
valorizes the views of the membership. In other words, the variables typically
used to measure the level of union democracy do not include directly asking the
member for his or her opinion on how the union is performing. It is the intent
of this article then to bring workers' opinions into the study of union
democracy.

All fights of reproduction in any form reserved.
ISBN: 0-7623-0853-2
95
96 ROBERT BRUNO
In order to assess rank-and-file support for democratic practices, the following

work presents as a case study, an opinion survey of Chicago International
Brotherhood of Teamsters' (IBT) Local 705. The local consists principally of
package delivery workers, freight, cartage and tanker drivers and commercial
movers. For most of the post-war years the local's dominant employers were
large freight and cartage haulers, but since the early 1980s United Parcel Service
(UPS) has employed a wide majority of 705 members. 1 Once recognized as a
poster child for corruption and elite rule Chicago area Local 705 has been
lauded as a model of democratic reform (Wall Street Journal, April 6, 1998).
Since an Intemationally imposed trusteeship in 1993 the union has painstak-
ingly worked to wipe off years of scandal, sweetheart-contracts and abuse of
the rights of the rank-and-file, z
The following analysis will be presented in three sections. After a brief
explanation of the conceptual rationale for utilizing a survey approach to
measuring union democracy, findings from the rank-and-file survey including
a description of the participants will be discussed and reported in frequency
tables. In addition, the responses will be subjected to bivariate analysis to deter-
mine the impact of various independent democratization variables on union
performance. Finally, the conclusion offers a note of caution in interpreting the
survey results.
CONCEPTUAL FRAMEWORK
A membership survey approach to measuring the character of union governance

accepts the widely held contention that democracy is by definition a collective
expression of the will of the people (Pateman, 1970). Simply put, the most
authoritative way to know whether a union democracy is in classical terms "by
the people" is to ask the people (i.e. union members). To be sure, ignoring
rank-and-file opinion could produce something called a constitutional or plural-
istic democratic union. However, it would be intellectually dishonest to define
a union as democratic in any sense when a majority of its members believed
that it was not.
Failing to acknowledge what members think can also create an overly
formalistic, and truncated assessment of union governance. As argued persua-
sively by Parker and Gruelle (2000), without direct input and feedback from
the membership on the union's governing behavior, the leadership can become
disconnected from the ranks. The result is that union leaders begin to equate
their judgments with what is best for the membership. Elections may still be
held, meetings conducted and laws abided by, but instead of democratic ideas
emerging out of an on-going member-leader relationship, they tend, as Robert
Consenting to be Governed 97
Michaels (1949) has argued, to be unilaterally imposed from above. In this

scenario, what passes for democratic behavior at the procedural level of analysis
is actually an imposed leadership agenda.
While open, honest leadership is vital to union representation, without
membership feedback, it produces what George Strauss has called a "democ-
racy for the people" and not "by the people. ''3 In the end, union members must
be able to pass judgment on more than contract issues or strategic planning
goals. They must be able to grant their "consent" to be governed in program-
matic and ideological ways. Importantly, consent is provided both before the
leadership acts and post hoc as a form of feedback and approval. Consent is
best measured by providing members with the opportunity to express an opinion
on how they are being governed. In this sense, the work presented here offers
a direct member-level diagnostic tool for measuring union democracy.
To address whether union members have consented to the form of
governance that has characterized Local 705 since the 1993 trusteeship, the
survey attempted to answer some fundamental questions about three critical and
related areas of union behavior: membership communication, membership input,
and membership representation. These areas were selected in order to
determine the level of membership consent to the most important changes
implemented by the local in response to the International's administrative
intervention. 4
Membership Communication
Workers were first asked whether they were aware of the many structural and
behavioral changes made at the local since the trusteeship was imposed and
additionally, whether the leadership communicated openly with the rank-
and-file. Prior to 1993 the local was infamous for treating the union's business
like confidential state secrets. Members had little access to their stewards or
business agents, and they were unwelcome at the local's offices (Bruno, 2000).
But because consent can only be offered when workers have sufficient knowl-
edge of what the leadership is doing or planning to do, it is critically important
to find out how information has been distributed since the trusteeship.
Membership Input
Second, members were also asked about opportunities to contribute to local
decision-making and whether they were supportive of the decisions reached
since the trusteeship was imposed. The International's findings revealed scant
evidence that the previous leadership had ever bothered to take serious the
views of the rank-and-file either before or after acting (Bruno, 2000). The
pre-trustee leadership had in effect insulated itself from the membership and
98 ROBERT BRUNO
consequently the approval or disapproval of the ranks was irrelevant to the

local's behavior. But because membership consent must not be falsely manu-
factured but earned, it is vital to first know if the members now have input
into their own governance, and secondly, do they approve or disapprove of
how the local leadership has governed since the trusteeship.
Membership Representation
Finally, members were queried about the impact of the trustee and post-
trustee changes on union representational behavior and on the relationship
between the members and leaders. In the two decades leading up to 1993,
elections at Local 705 were a fraudulent exercise, most worker grievances
were lost or dropped, arbitrations were rarely filed, sweetheart contracts
regularly negotiated, and external organizing all but ignored (Bruno, 2000).
In most situations, members were denied an opportunity to provide their
leaders with informative signals about how the governing regime was
representing worker interests. In order then to determine the effects of
post-trustee political change on the level of rank-and-file support for
democratic reform, it is essential that membership feedback be provided to
the leadership.
In order to answer each of the fundamental questions workers were
presented with a number of additional questions addressing various related
elements of democratic unionism (Fletcher, 1998). Lending greater weight
to the IBT 705 survey is the realization that it represents the first and only
independent analysis of Teamster rank-and-file opinions concerning union
democracy. Given the difficulty in gaining access to union members, it is
nearly unprecedented to construct an independent, direct rank-and-file assess-
ment of any union's democratic temper. Additionally, in light of Local 705's
undemocratic past, attaining such a measurement here is particularly fortu-
nate because it provides observers with the opportunity to assess the impact
of Local 705's transformation.
SURVEY RESPONDENTS
In the fall of 1999 a 57-question survey was mailed to 3,000 rank-and-file

members of IBT Local 705. Respondents were chosen from a stratified
random sample of non-officers, stewards and staff designed to reach members
employed in UPS and non-UPS divisions represented by the local. The
number of surveys sent to workers in each division was proportional to their
overall representation within the union. 5 After excluding incorrectly or
incompletely filled out surveys, and those returned in an untimely fashion a
total of 241 cases (8.4%) were assessed. While the small number of cases
is disappointing it was not unexpected and a brief discussion about the
possible reasons for the response rate will be addressed below.
Respondents averaged approximately 44 years of age and the median term
of membership was an even dozen years. It is important to note that with
three-quarters of the survey participants paying union dues for at least five
years or more the survey captured members that had experienced to some
degree the union's dramatic transition. This composition permits a more
informative assessment of how the local has changed its charter and mode of
representation.
Nearly three quarters of the respondents (72.6%) were working full time
and 41.5% had earned at least a high school or general equivalency diploma.
Survey participants were overwhelmingly male (88.5%) and white (76.7%).
Median income for 1999 was between $40,000 and $49,000, with 18.5%
earning less than $20,000 and 2.6% grossing between $70,000 and $79,000.
Now before going further there are two representational problems with
the sample that need to be addressed. First, the returned surveys included a
predominance (64.3%) of non-UPS members and consequently under repre-
sented the local's largest unit, the UPS division. In doing so the survey
responses are likely biased towards drivers of local cartage and freight
companies. The effect of this pool is to slightly increase the average age of
respondents, but not the average years of union membership. The education
level of the overall sample was also reduced by a minor degree with a higher
percentage of UPS members than non-UPS workers holding some college
credit. In addition, the median income was nudged downward with a higher
percentage of non-UPS drivers earning within and below the survey median
range.
The second problem with the sample is that by under representing UPS
workers, responses also failed to reflect the very large pool of part-time
workers in the local (only 5.1% of useable part-time surveys were returned
compared to a much higher if still a very modest 14% of full-timers). While
UPS consists of nearly 60% part-time workers, full-time package car drivers
made up the majority (52%) of the UPS respondents. As a result of this break-
down the demographics of the participants were overly influenced by
full-timers. For instance, part-time respondents were on average considerably
younger, less experienced with unions, considerably less well compensated,
and much more likely to be female and minority. Unfortunately, due to a
series of internal union debates and governance matters that transpired after
the initial survey was processed, it was impossible to send a second targeted
mailing to part-time members. Under an ideal situation, a follow up mailing
100 ROBERT BRUNO
to part-time workers with at least seven years of experience would have

improved the representativeness of the sample.
But occupational status alone likely explains only a part of the low response
rate. In addition to the standard problems of a one-time mail survey, a less
verifiable but probable cause for the modest cartage/freight response is the union
itself. Rightly or wrongly the IBT has for most of its history been synonymous
in the public's mind with corruption. While the International union was once
a powerful source for dramatically improving the working lives of truck drivers
from coast to coast it is also true that it was once intricately connected to
organized crime. From the sensational McClellan Congressional Committee
Hearings in 1959 to the 1989 Racketeer Influenced Corrupt Organization
(RICO) civil lawsuit and consent order the IBT has been portrayed as
organized labor's "number one enemy. ''6
After the 1991 historic referendum election for International officers the
union membership has withstood 40 local and one joint council trusteeships,
217 permanent member expulsions, charges against nine International officers
or appointees, the criminal conviction of an International Director of
Governmental Affairs, the expenditure of $20 million dollars in taxpayer funds
trying to get the union to act democratically, had an international election
overturned and saw its reform president banned from the union for misusing the
membership's dues money (Daily Labor Report, 1997 and 1999; Conboy,
1997; Union Labor Report, 2000). Under these wrenching and highly uncertain
conditions it would be unreasonable to expect any rank-and-file Teamsters to
easily offer their opinions about a union that "inspired the institutional equiva-
lent of a foreign invasion" to begin respecting worker rights (Kannar, 1993).
Of course Local 705 is not just any group of Teamsters working in anyplace
USA. This is after all the Chicago Teamsters. At the beginning of the
20th Century, Chicago was the operative birthplace of the union and it has
always been a "big, brawling play-for-keeps world" for Teamsters (Dine, 2000,
A12). Local 705-business agent Sam Canino found out just how brawling it was
when in the late 1960s he nonchalantly lifted his garage door and was greeted
with a gun shot blast that blew away part of his face. Where in 1983 crime syn-
dicate connected and union bagman Allen Dorffman was gunned downed in a
suburban parking lot. It was also where in the early morning hours of August
21, 1991 ex-Local 705 secretary-treasurer Danny Liquroutis shot his adopted
son with a .38 caliber pistol in the basement of Teamster City. And that bizarre
act of self-defense or murder was preceded five months earlier by the beating
of 705 Teamster for a Democratic Union (TDU) member Leroy Ellis at a local
delegate-nominating meeting (Modea, 1978; Crowe, 1993). These violent inci-
dents and other unflattering affairs of course happened prior to the trusteeship,
which was later imposed for a plethora of different improper and nefarious acts.
It should come then as no surprise to any observer, that rank-and-file members
of a Chicago Teamster local still in its democratic infancy would be reluctant
to express even an anonymous opinion about the union's behavior.
SURVEY FINDINGS
The following section includes a presentation of survey findings, grouped

according to the three areas of study mentioned above. In nearly all cases
where statistical data is reported a corresponding table is also provided.
However, in those few cases where tables are not shown it is because the
information was collected in a dichotomous (i.e. yes or no) fashion, or all the
findings related to the survey items were fully described in the narrative.
Membership Communication
In order to determine how well informed workers were about Local 705"s
political transformation, they were first asked directly whether significant
changes have occurred in the union since the 1993 trusteeship. A robust
63.9% "agreed" that significant changes had in fact been made (see Table 1).
Additionally, nearly 70% of the respondents agreed that they were aware of
the reasons given for imposing the trusteeship and for making changes in how
the union functions (see Table 2). Importantly, by a smaller proportion, about
one-half of the respondents agreed that the reform leadership made them
"aware" of their rights as union members (see Table 3). While this figure
appears less than optimal, it is nearly twice as large as those respondents who
disagreed and certainly a significant improvement over the pre-trustee days. 7
Workers were also asked about two principal union vehicles for communicat-
ing information to members and for members to provide feedback on the
leadership's performance. Prior to the trusteeship the local did not have a news-
paper or newsletter. But shortly after the trusteeship was imposed the Local 705
Update was published. The local paper is mailed monthly to every active member
and covers topics ranging from grievance panel decisions and arbitrations to
political issues. Like other in-house communication devices it serves primarily
to inform the membership about the accomplishments of the local leadership.
While it is not (nor should it be) the equivalent of a community paper it does
offer its readers an opportunity to become aware of what the local is doing as a
bargaining agent. The 705 Update has also been central to the local's reform
efforts. For the first the time in the local's history, the names and phone numbers
of all business agents were made publicly available to the membership through
102 ROBERT BRUNO
Table 1. There have been significant changes made over the past five years.
Item Frequency Percent

Strongly Agree 72 29.9
Agree 82 34.0
Neutral 38 15.8
Disagree 17 7.1
Strongly Disagree 17 7.1
NA 15 6.2
Total 241 100.0
Table 2. The trusteeship was imposed to restore integrity to the local and
to protect members' interests.

Agree 83 34.6
Neutral 38 15.8
Disagree 16 6.7
NA 7 2.9
Total 241 100.0
the pages of the paper. An important question then for continued democratic
advancement is whether or not members are bothering to read the local paper.
When asked how often they read the 705 Update, a sizeable 60.6% of the
respondents said "always," and another 20% looked at it "most of the time"
(see Table 4). In addition, nearly 70% agreed that since 1993, "union
materials have kept me informed about union matters" (see Table 5). This
degree of rank-and-file faith in the top-down transmission of information
would be impressive for any local union. It is even more so when you keep
in mind that before 1995 union material directed at the membership was hard
to find. Reading of course does not necessarily mean understanding or even
becoming better educated, but it does strongly establish a showing of interest
in the institution's behavior. As any classic democrat would note, paying
attention to what the government (i.e. union leadership) does is more than
an important check against an abuse of power; it is a sign of political
engagement.
A second source of reciprocal information and communication are the local's
stewards and union representatives. In a democratic union the nexus points that
connect workers with their leadership are numerous. Along with reading the
local paper members attend meetings, vote and campaign for officers, and run
Table 3. The post-trustee leadership keeps me aware of my rights as union

members.

Agree 80 33.1
Neutral 49 20.3
Disagree 38 15.7
NA 4 2.1
Total 241 100.0
Table 4. How often do you read the Local 705 Update?
Never 5 2.1
Rarely 6 2.5
Sometimes 28 11.9
Most of the time 49 20.8
Always 143 60.6
NA 4 2.1
Total 241 100.0
Table 5. Since 1995 union materials have kept me informed about

union matters.

Agree 114 47.3
Neutral 44 18.3
Disagree 23 9.5
NA 5 2.1
Total 241 100.0
for office. In order for members to participate in the union and subsequently
to consent to the u n i o n ' s politics, they need to be well informed by their
stewards and representatives. And in fact, most !705 respondents feel that they
are. Slightly less than half of the respondentsl agreed that they were "kept
informed" by their stewards and representatives compared to nearly 35% of
those who disagreed. In addition, a majority al~o believed that their stewards
104 ROBERT BRUNO
and representatives are accessible "most o f the time" or "always" (see

Table 6 & 7).
While a larger number felt informed than did not the difference was not
exceptionally wide. There apparently remains some dissatisfaction with the
union's outreach to its members. The survey of course does not provide a
mechanism to explain the difference in respondent's answers but one inter-
pretation can be proffered. W o r k e r attention to and subsequent evaluation of
their u n i o n ' s communication efforts is typically "event" driven. W h e n members
have a problem or deep concern they establish a closer relationship with the
union leadership. Consequently, they then have a baseline to evaluate how
well the union has kept them informed. It is possible that all o f the respon-
dents had equal access to information but that more of the "agrees" had more
cause to call on the union then their counterparts. The question does not o f
course assess whether the m e m b e r s were pleased with the information; only
whether or not they were kept in the dark about the u n i o n ' s behavior. In this
sense the r e s p o n d e n t s ' answers m a y have tapped into a union i n v o l v e m e n t
matter.
Table 6. Since 1995 business agents and stewards have kept me informed
about union matters.
hem Frequency Percent

Agree 76 32.2
Neutral 47 19.9
Disagree 53 22.5
NA 0 0
Total 236 (5 missing) 100.0
Table 7. How often do you have access to your business agent and steward
(percent only)?

Never 7.7 8.8
Rarely 16.6 11.3
Sometimes l 9.1 13.8
Most of the time 25.5 27.9
Always 20.9 36.7
NA 10.6 1.7
Total 100.0 100.0
MEMBERSHIP INPUT
One indication of the local's interest in membership input is the degree in which
the ranks feel that they are encouraged to participate in union affairs. In stark
difference to the pre-trustee days, a very robust 64% of members agree that they
are "encouraged to participate" (see Table 8). Of similar contrast to the previous
governing regimes, over half of the respondents concurred that they now have
"opportunities to attend union meetings and programs" (see Table 9).
Members were also asked to indicate their overall approval or disapproval
of the changes that have occurred since the trusteeship was imposed. While
one-quarter of the respondents remain uncertain about the changes, 45.2% either
"strongly agree" or "agree." Most importantly, less than a quarter of partici-
pants either "strongly disagree" or "disagree" with the changes that have
occurred (see Table 10). This level of support was however, considerably less
than the near 70% of respondents who agreed that the trusteeship was "needed
to protect the members." Apparently the members' approval of particular retail
changes has been less than their support for a wholesale change in union lead-
ership and internal governance.
Table 8. Members are encouraged to participate.

Agree 106 44.0
Neutral 38 15.8
Disagree 27 11.2
NA 5 2.1
Total 241 100.0
Table 9. Members have oppurtunities to attend union meetings and

programs.

Agree 81 30.7
Neutral 73 30.3
Disagree 21 8.7
NA 5 2.1
Total 241 100.0
106 ROBERT BRUNO
A final measure of approval was taken by asking workers whether the

post-trustee elected local leadership conducted affairs consistent with the
"best interest of the members?" While a not immaterial 30% "disagreed," a
much heftier 47% agreed that their interests were being protected (see Table
11). The nearly identical percent of respondents who both approved of the
changes and believed that they were made in the members best interest, suggests
that consent may be related to a rank-and-flier's belief in whether workers as
a group, and not the leadership or management are the intended beneficiaries
of the changes.
The disapproval, however, of approximately one-fifth of the respondents to
how the local has transformed, suggests to the reader the complexities involved
in popular reaction to organizational change under any circumstances. But in
Local 705's case change was mandated, dramatic and highly adversarial. Rank-
and-file members did not get a democratic vote on whether they wanted to
transform their union. If they had, it is likely that a significant number of them
would not have voted for change. For a number of decades the local was ruled
with an iron fist and members were treated like "dependents" of the leadership.
In the "old days" a member had literally nothing to do but go to work. But
Table 10. I approve of a majority of the changes.

Agree 74 30.7
Neutral 59 24.5
Disagree 32 13.3
NA 16 6.6
Total 241 100.0
Table 11. Changes have been made in the best interest of the
membership.

Agree 65 27.0
Neutral 46 19.1
Disagree 44 18.3
NA 12 5.0
Total 241 100.0
since the trusteeship, opportunities for involvement have expanded and the
union's business is no longer a secret.
Despite the observable governing differences, there may be a "developmental
lag" between institutional change and membership attitudes. In other words,
while there is undisputable evidence that the union is procedurally more
democratic today than it was pre-trusteeship, members' perceptions of the
union's character may not have caught up with reality. If this is the case,
then the phenomena playing out here may not be unlike what occurs in
emerging democratic states. Autocratic governing bodies undergo democratic
reform while popular consciousness remains captured by disbelief and
suspicion of all things new. In Local 705 the old ways of acting have been
dying, but perhaps the new ways of thinking have yet to be completely born.
MEMBERSHIP REPRESENTATION
As a post hoc assessment of the local's performance, membership feedback

provides a valuable way for the membership and the leadership to build a demo-
cratic relationship. In two of the most significant and mixed findings of the
survey, a plurality (44%) of respondents both agreed that since the trusteeship,
"workers have more influence over how the union acts" and consequently, that
the "union acts for the membership" (see Tables 12 and 13). In both cases,
however, between one-quarter and one-third of the members answered nega-
tively. The mixed nature of the responses here, is relevant to understanding
how Local 705 members evaluated the consequences of the last seven years of
union representation.
On one hand, considering Local 705's not to distant checkered past, it is a
positive sign that better than two out of five members feels that they now have
more influence over what the union does and that what gets done is on behalf
Table 12. An important change is that workers have more influence over
how the union acts.
I~m Frequency Percent

Agree 62 25.7
Neutral 50 20.7
Disagree 38 15.8
NA 10 4.1
Total 241 100.0
108 ROBERT BRUNO
Table 13. Since the trusteeship, the union acts for the membership.

Agree 76 31.8
Neutral 56 23.4
Disagree 34 14.2
NA 13 5.4
of the dues-payers. On the other hand, it is troubling that a third of the surveyed
workers did not believe that their local had become more responsive to the
views of the membership. Two related and plausible explanations may provide
some insight into the nature of these responses.
First, perhaps the negative respondents are individuals who because of
mistreatment or misunderstanding continue to feel disenfranchised despite
the formal introduction of democratic procedures. Conversely, the positive
respondents may have had good experiences with the union leadership. In the
end, individual workers may well have evaluated the union's responsiveness to
the membership based on their own self-interest, instead of on what was best
for the collective body.
It is also possible that since the trusteeship a more open space for political
disagreement has created a partisan mentality among competing groups of
workers. In a classic study, Leonard Sayles and George Strauss (1953) pointed
out that conflicts among interest groups within local unions are not uncommon
and that efforts to deliberately increase democratic participation tend to create
divisive internal battles that may leave some factions feeling marginalized
(Terry, 1996). But despite the messy political disagreements made possible
because of a democratic openness, the leadership is still charged with the
responsibility of acting for the union and therefore, individual member
interests and institutional fortunes are not likely to perfectly align.
Feeling like you have no greater input into how things get done is not
necessarily a condemnation of the system of union governance. An additional
measure of democratic reform, which helps to parse out the meaning of
rank-and-file attitudes, is whether or not the trusteeship helped to rid the local
of corrupt leaders. When asked if "corrupt officers had been removed from the
local" about one-quarter of survey participants disagreed (see Table 14); that is
slightly fewer than responded negatively to the question of workers' influence
over the union. One interpretation of this variance is that while some workers
feel that since the trusteeship their leaders are not more responsive to them,
these same leaders are j u d g e d for the most part to be honest. In other words,
they may not be democrats in a New England-town hall sense, but neither are
they "mob-upped," "thugs," or "criminals." While the pre-trustee membership
was rarely given an opportunity to express their opinion about anything, what
is known from a sample of written union documents is that local leaders were
often charged by the membership with at best indifferent behavior and at worst
with out-and-out evilness (Bruno, 2000).
In the pre-trustee days workers were treated like children; seen but not heard.
A fear of reprisal for speaking out against the leadership was well founded. So
what happened when the local submitted to independently monitored elections
and other trappings of democratic practice? Do members now have the freedom
to speak out? Does the fear of j o b loss or the cold sting from the back of a metal
chair still represses a worker' s voice? Apparently for a small minority of workers
the rights of free expression are no greater today than they were five years ago.
But for nearly 60% of workers, there is greater freedom to speak out now than
before the International filed abuse of power charges against the Local 705 "Old
Guard" (see Table 15).
Table 14. The trusteeship helped to remove corrupt union leaders.

Agree 74 31.1
Neutral 47 19.7
Disagree 26 10.9
NA 19 8.0
Table 15. Members have the freedom to speak out without fear.

Agree 97 41.1
Neutral 43 18.2
Disagree 29 12.3
NA 0 0
110 ROBERT BRUNO
Being encouraged to get involved in your union does not mean however, that
members are taking up the charge. Union democracy should not only make
possible membership participation; it needs to inspire use of the franchise. During
the pre-trustee period IBT 705 rank-and-file members paid dues, went to work
and rarely voted in union elections or on labor agreements, and fewer still spoke
out at or attended membership meetings. Thus, an important question about the
behavioral consequences of 705's democratic transformation is whether, since
the trusteeship was lifted, rank-and-file participation in union activities has
picked up. Voting in union elections is an obvious measure of democratic health
and a good place to begin an analysis of political participation.
There have been two local officers elections since the trusteeship was imposed
(the first one in 1995) and in the 1997 race 88% of the survey respondents cast
a ballot. While this indicates that the survey may be over-represented by
politically involved union members, it is very illuminating that nearly a third
of this group "never" voted in local elections before 1993 and approximately
42% either never cast a ballot or voted only "sometimes" (table not shown). 8
Democratic leaders should be freely chosen in meaningful and honest
elections, and represent their constituencies in both form and substance.
Consequently, voting has become a more valued institutional good because
43.7% of workers believe that "elections are honestly run." While there is a
troubling 33.2% that remains uncertain about the validity of elections, there is a
decidedly smaller number (23.1%), which views the contests as illegitimate (table
not shown). Recall that prior to 1993 there were very few contested races for
union office and where they occurred workers were rarely offered a full airing
of issue positions. But here again change seems substantial. When asked if
"elections provide members opportunities to address important issues," 54.2% of
participants agreed (see Table 16a) and nearly 60% concurred, "elections offer
real choices among candidates" (see Table 16b).
More importantly, those union choices appear to be better than twiddle-
lee-dee and twiddle-lee-dom. Nearly 40% of survey participants agreed that,
"elections provide the membership with qualified officers," while less the
one-quarter disagreed (see Table 16c). Respondents also agreed that, "elections
are important to how the members are represented" (see Table 16d), and that in
fact, third party monitored and contested elections have improved the quality of
"membership representation" (see Table 16e).
Voting is vital to democratic choice, but it's not a very timely or particu-
larly precise instrument to effect policy. To act immediately and with individual
precision the monthly membership meetings better serve the members. One of
the initial decisions made by the first post-trustee elected executive board was
to move the date of the membership meetings from a mid-week evening to a
Table 16a. E l e c t i o n s p r o v i d e m e m b e r s o p p o r t u n i t i e s to a d d r e s s
i m p o r t a n t issues.

Agree 93 41.0
Neutral 54 23.8
Disagree 41 18.1
NA 0 0
Table 16b. E l e c t i o n s o f f e r m e m b e r s real c h o i c e s a m o n g c a n d i d a t e s .

Agree 102 44.5
Neutral 53 23.1
Disagree 37 16.2
NA 0 0
Table 16c. E l e c t i o n s h a v e p r o d u c e d qualified officers.

Agree 73 31.6
Neutral 85 36.8
Disagree 35 15.2
NA 0 0
Table 16d. E l e c t i o n s are i m p o r t a n t to h o w the m e m b e r s are r e p r e s e n t e d .

Agree 115 50.0
Neutral 31 13.5
Disagree 15 6.5
NA 0 0
112 ROBERT BRUNO
Table 16e.. Elections have improved the way that are members are
represented.

Agree 86 37.6
Neutral 68 29.7
Disagree 34 14.8
NA 0 0
Sunday morning. The new date was chosen in order to maximize meeting
turnout. It may have worked. The rank-and-file survey revealed that 63.6% of
respondents agreed that "union meetings since 1995 have been open to all the
members" (see Table 17).
Good as that figure is, it tells us little of the individual member's actual use
of the meetings as a way of being involved in the union. A better measure
would be a comparison of how many annual meetings a worker attended pre
and post-trusteeship. On this score the differences could not be starker. While
an equal number of members annually attended nine or more meetings both
prior to and after 1995, nearly 10% more of them had attended one or more
meetings since that year. In addition, 41% of the respondents had never
attended a single meeting before the trusteeship. That number had dropped to
32% following the return to a locally elected executive board (see Table 18).
Going to membership meetings is an act of participation; so is walking up
to the floor microphone and speaking your peace. Under the previous regimes
it was a brave dissenting soul, true believer or sycophant who addressed the
multitude of assembled union members. Accordingly, only 15.6% of respon-
dents said they ever "spoke at any membership meeting" prior to 1995. But
Table 17. Union meetings are now open to all the members.

Agree 100 41.5
Neutral 49 20.3
Disagree 24 10.0
NA 5 2.1
Table 18. Times you attended Union meetings (percent only).
Item Before 1995 Since 1995
Nine or more 10.3 9.2

Eight or seven 6.4 6.3
Six 5.9 7.5
Five or four 6.4 9.6
Three 6.9 12.9
Two 9.3 14.2
One 14.2 8.8
Zero 40.7 31.7
Total 100.0 100.0
since the trusteeship members no longer need to supplicate themselves before

the leadership in order to be recognized at a meeting and consequently, the
number of workers speaking at meetings has dramatically increased (Bruno,
2000). However, while speaking out is no longer a hazardous act only 18.6%
of the survey participants have chosen to take advantage of this more protec-
tive environment (table not shown).
Nor are contemporary rank-and-fliers better utilizing union educational
opportunities. Roughly 9% of pre and post trustee respondents admitted that
they "attended union educational program" (table not shown). Now this
equivalency of use hides the fact that before the election in 1995 there was
little evidence that rank-and-file members had many educational opportuni-
ties available to them. On the other hand, since a reform movement slate was
first elected, educational opportunities have expanded (Bruno, 2000). While
low membership utilization rates typically reflect individual workers' work
schedules, family commitments and career choices there is a qualitative
difference in leadership between one that makes resources available for
education purposes and one that does not. The former would certainly better
serve union democracy.
But the low number of workers choosing to speak at membership meetings or
attending educational sessions does reflect a puzzling condition about union and
civic democracy. Commentators from different disciplines have noted that
democratic structures and opportunities do not in themselves generate widespread
democratic behavior. Having the right and the means to vote does not bring
someone to the polls. Likewise, access to union meetings and the space to be
recognized does not draw someone to the floor microphone. While different
theories abound as to why even under ostensibly democratic conditions citizens
(Dahl, 1961; Parenti, 1980; Barber, 1984) and union members (Cochran, 1977;
Parker, 1998; Sciacchitano, 2000; Parker & Gruelle, 2000) remain silent, most
114 ROBERT BRUNO
theorists (Kuklinski et al., 2001) have argued that either a salient issue or a press-
ing need, along with a belief in the efficacy of acting is required to stimulate
democratic participation.
Importantly however, scholars (Dahl, 1961; Luskin, 1987) also warn that the
lack of democratic activity does not necessarily signal dissatisfaction with the
governing structure. If, for example, 18% of the members speaking out is a
sign that all is well with the union then the lack of participation here may not
be troubling. But if 18% means that most members do not believe it matters
whether they speak out or not then Local 705 may still have a credibility
problem within its ranks.
Along with rank-and-file opportunities to attend union meetings and
educational programs, the effectiveness of the local's transformation can be
measured by how well the leadership has responded to workers' grievances
about management. Unlike the pre-trustee years when members' grievances
were largely disappeared, exactly half the respondents revealed that problems
with management are "always" or "most of the time" investigated (see Table
19). Additionally, with union representatives on duty it comes as no surprise
to find that 45% of respondents think that their "steward and business agent
are qualified," and the same proportion believe that they have done a "good
job" (see Table 20). Significantly on both measures less than one-third of
workers disagreed.
It is also expected that in a democratic union the leadership will not only
act on behalf of the people, but also do so with respect. For Local 705's reform
agenda it was critical that respect of the membership be shown at all times. If
members did not feel that self-proclaimed reformers treated them respectfully
there would be little likelihood of changing the local's dismal trajectory. On
this score considerable improvement has been made. Slightly more than half of
the respondents (51.3%) believe that they are "treated with respect," roughly
doubling the number (26.8%) that still feels disrespected (see Table 21).
Table 19. How often are members' grievances investigated?

Never 9 3.8
Rarely 24 10.3
Sometimes 46 19.7
Most of the time 66 28.2
Always 52 22.2
No opinion 37 15.8
Table 20. Qualifications and performance of business agents and stewards

since 1995 (percent only).
Item Agents and Stewads Agents and Stewards

are qualified have done a goodjob
Strongly Agree 9.6 14.2
Agree 35.6 30.5
Neutral 27.2 29.7
Disagree 20.1 15.5
Strongly Disagree 7.5 10.0
NA 0 0
Total 100.0 100.0
Table 21. Members are treated with respect.

Agree 82 34.0
Neutral 54 22.4
Disagree 38 15.8
NA 5 2.1
Total 241 100.0
Respecting the membership at IBT Local 705 has likely contributed to more
successful grievance panel decisions, arbitrations and better contracts. But has
it narrowed the status gap between officers and the membership? Sociologist
Robert Michels (1949) predicted that the "iron law" of organizations would
slowly but assuredly separate the leadership from the membership and place
additional restraints on union democracy. To measure whether workers have
begun to bridge the "vertical social distance" (Mannheim, 1956, p. 180) that
would certainly separate them from their leadership in a undemocratic union,
respondents were asked to identify how well they understood the actions and
decisions of their officers. A remarkable 72% of respondents said that they have
an excellent to good "level of understanding" of what Local 705 is doing (see
Table 22).
While such a response does not indicate approval or disapproval, it does
signify that the institution is not an alien concept to the membership.
Democracies, unlike dictatorships, are not a mystery. Democratic processes may
be messy, combative, time consuming and demanding of compromise,
coalition building and public relation skills, but they are understandable by
116 ROBERT BRUNO
Table 22. What level of understanding do you have for the policies and
actions of Local 705?

Excellent 19 7.9
Very Good 55 22.8
Good 95 39.4
Fair 20 8.9
Poor 38 15.8
Very Poor 15 6.2
Total 241 100.0
anyone concerned enough to pay attention. However, one party states, elite rule
or family control are built on the whims, prejudices, and self-interest of power
holders. Governance of the local during much of the pre-trustee period was
personal, private, irrational and thus, not subject to rank-and-file understanding. 9
But since 1995, inclusive of all blemishes, the membership has indicated a high
level of understanding for the behavior of the local.
Scholars of union democracy have noted that unions often begin to fray at
the edges of a worker's sense of institutional ownership. A union "by the people"
should generate a reasonably high level of confidence that the workers have
ownership over their local. What sense of ownership should union members
have about their organization? In the case of Local 705 the results appear
ambiguous. On one hand, despite large pluralities of workers supporting
numerous improvements in the local, approximately only 37% of respondents
agreed that they now had a sense of "union ownership." Adding the one-third
of respondents who remain uncertain to the nearly 34% who disagreed, means
that two-thirds of the members are not feeling a sense of ownership (table not
shown). Admittedly the finding here is negative. But seen from another perspec-
tive the results may be more encouraging. In just five years two-thirds of the
membership have either developed a sense of ownership or remain open about
the matter. While no pre-trusteeship rank-and-file survey exists it is not
unreasonable in light of Local 705"s past to interpret a slightly higher number
of workers who claim a sense of union ownership than those who do not as
an achievement.
One final question bearing on the union's representational function was
posed to Local 705 members. Members were asked to provide an umbrella
definition of Local 705. The answer in broad terms for all but a small
minority is that IBT 705 is a union that "effectively represents the member-
ship." But differences emerge when the nature of that representation is
qualified. For instance, one-fifth of respondents describe the local as being

effective and involving "the membership in all union matters." Another 31%
acknowledges the local's effectiveness but holds that it "most often acts
without membership involvement," and more often than not just "informs
the membership about decisions made" by the leadership. That leaves only
17.5% of respondents who describe Local 705 as a "union that fails to
represent the membership and denies the membership an opportunity to partic-
ipate in union matters" (see Table 23). Results here would seem to support
George Strauss's important distinction between a union leadership that is
democratically governing "for the people" but that has not yet found a way to
allow rank-and-file workers to govern themselves.
A final assessment of how workers judged the dramatic post-trustee develop-
ments at Local 705 is a measurement of what particular changes workers have
most closely associated with the democratization movement. To accomplish this
Pearson correlations were drawn between levels of agreement that significant
changes have been made and the nature of those changes. The independent
variable is the imposition of the trusteeship and the reforms implemented since
1995, and the dependent variables were the following changes: (1) members have
the freedom to speak out; (2), members are encouraged to participate in union
activities; (3), members have the opportunity to attend union programs; (4), the
leadership values input from the members; (5), workers are protected against
abuse by union officials; (6), corrupt leaders were removed from the union; (7),
members are more aware of their union rights; and (8) members have a sense of
union ownership. On each of the items the correlation was positive and in all but
two cases significantly strong (see Table 24). Two findings are of particular
Table 23. Which of the following statements best describes the present
Local 705?
A union that effectively represents the membership

and involves the membership in all union matters 47 20.5
A union that effectively represents the membership
but most often acts without membership involvement 71 31.0
A union that effectively represents the membership but
basically just informs the membership about decisions made 71 31.0
A union that fails to effectively represent the membership
and denies the membership an opportunity to participate in
union matters 40 17.5
Total 229 100.0
118 ROBERT BRUNO
Table 24. Relationship between agreement that significant changes have

been made since the trusteeship and particular changes.
Item Pearson Correlation

Members have the freedom to speak out 0.503**
Members are encouragedto participate in union activities 0.464**
Members have the opportunity to attend union programs 0.440**
The leadership values input from the members 0.653**
Corrupt leaders were removed from the union 0.134"*
Members are more aware of their union rights 0.567**
Members have a sense of union ownership 0.630**
**Sig.@O.Ol (2 tailed)
interest. The first is the very substantial positive relationship between democra-
tization and the value the leadership now places on rank-and-file input. Secondly,
surveyed members have strongly signaled the importance of having a sense of
ownership over their union.
As positive as the overall survey findings are, what is perhaps more encour-
aging is that the respondents drawn from over-the-road truckers covered under
the local's Master Freight Agreement, traditionally supportive in Chicago of
the conventional elite leadership, were equally or more supportive of the way
that the local had transformed itself over the past seven years. When UPS
workers were compared to non-UPS members there were no meaningful statis-
tical differences in the positive direction of the responses. By modest to large
margins UPS and non-UPS workers showed majority support for the changes
that have occurred since the trusteeship. With few exceptions the percentage
of agreed and disagreed responses of both groups was within a few points.
Overall, however, non-UPS members displayed a slightly higher-level support
for democratic reforms on more of the items than did UPS workers.
The support given for democratization by members within the local's
primary divisions is bolstered by the knowledge that the units displayed
contrasting political loyalties. During the 1998 rerun election of the Teamsters'
International contest for general president, it was commonly interpreted
(perhaps mistakenly) that a vote for either James Hoffa Jr. or Tom Leedham
signified a worker or local's support for continuing the reforms of the Carey
administration. Voting for general president then became a proxy for
furthering or restricting the government enforced changes in the behavior of
the Teamsters.
In the city of Chicago vote support for James Hoffa Jr. was nearly
unanimous. In fact, every local in the state of Illinois gave a majority of its
votes to Hoffa and the son of the legendary Teamster leader won dominant
control over the Midwest region and Chicago Joint Council. The only Chicago
and Illinois exception to Hoffa's sweeping success was Local 705. Led by
705 Secretary-Treasurer Gerald Zero and President, John McCormick who
was also running for the secretary-treasurer post on the Leedham slate, the
local cast 58% of its votes for the decided underdogJ °
While no figures were available to verify the vote breakdown it was a
common assumption within the local that Leedham's strongest support came
from UPS workers. Survey findings confirm that assumption. Leedham won
nearly three-quarters of the votes cast by UPS members, but just 32% of non-
UPS workers. The sizeable and contrasting margins should have revealed a
serious fault line in the local. The campaign for the International presidency
was framed over the controversial and emotional ideas of either returning to
an ugly past of mob connected union bosses or continued subservience at the
feet of an oppressive federal government. In Chicago little middle ground was
carved out.
To be sure, individual Teamster voting reflected more complex ideas about
the union's identity then popular belief permitted. Nonetheless, it is undeniable
that the election was waged in near apocalyptic terms. But despite the often
acrimonious and ideological nature of the race, Hoffa and Leedham supporters
endorsed in near equal proportions the democratic transformation of Local 705.
CONCLUSION
The internal governance of Local 705 has apparently passed a membership

consent test. After experiencing seven years of dramatic personnel, structural
and procedural changes, Local 705 members gave an encouraging if cautious
endorsement to the post-trustee period. On every measured item more workers
than not approved of what their local leadership was doing. Support however
was not overwhelming. The roughly one-quarter to one-third of respondents
who responded negatively to many questions, along with the 20% who were
commonly uncertain of their feelings suggests a possible tipping stage in
membership opinions. The local cannot ignore the one out of five "swing voters"
who have yet to make up their minds about the post-trustee changes. There is
potentially a near majority of rank-and-file opponents who are not supportive
of the local's behavior. Some of these members may feel disaffected by the
actions already taken, while others are holding back their approval until the
local goes even further with its reform agenda.
Thus it would be in the local's best democratic interest to interpret the survey
findings cautiously. Such an interpretation however, should not distract from
120 ROBERT BRUNO
the positive assessment members have rendered. In the final analysis, when
asked for the first time in the local's history to evaluate the character o f the
union's internal governance, the membership has given an endorsement to the
democratic changes it has experienced. The membership will of course have
other opportunities to express its approval and to grant its consent through
officer elections and contract votes.
As workers and their leaders go about the business of building a real
democratic union it is important for interested observers to recognize
how terribly difficult it is to go from a condition of tyranny to one of self-
sovereignty. Union democracy, like political democracy is not any easier to first
realize and then sustain just because it is the morally right way for people to
govern their public affairs. As researchers continue to identify and test a range
o f democratic measures, a future challenge is to explore the relationship between
union transformation and members' opinions. Union members are rarely directly
asked to evaluate how they are being governed; yet locals like IBT 705 have
undergone major overhauls. While membership surveys can never definitively
assess a u n i o n ' s governing character, they can make a significant contribution
to revealing whether workers have consented to being governed in particular
ideological and pragmatic ways.
NOTES
1. The membership breaks down in the following approximate way: UPS - 11,500,
Cartage - 1,200, Air Freight - 1,000, Freight - 1,000, Tankers - 600, Liquor Division
- 300, Movers - 200, Municipalities - 150 and Grocery Houses - 100.
2. The trusteeship was based on the findings of an Independent Review Board (IRB)
set up by a 1989 consent decree negotiated between the IBT and the federal govern-
ment to monitor the union's internal governance. In addition to the IRB the union's
Ethical Practices Committee also charged the local leadership with various violations
(see, Letter and report from the Independent Review Board to Ron Carey dated May
25, 1993).
3. George Strauss' comment was made during a panel presentation titled "Union
Governance in a Changing Labor Movement," at the 53rd Industrial Relations Research
Association Annual Meeting, January 6, 2001, New Orleans, Louisiana.
4. The Independent Review Board had investigated the local and filed a seventeen
page report with the International union which "evidences financial wrongdoing, consti-
tutional violations and a deteriorating condition that warrants the imposition of
Trusteeship and the filing of charges against officers of Local 705." The IRB findings
determined that the local leadership had "brought reproach upon the IBT" for both finan-
cial and non-financial violations of the IBT Constitution" (see IRB Report to Ron Carey,
May 25, 1993).
5. The breakdown was as follows: 1885 surveys were mailed to UPS members (63%)
and 1,115 sent to non-UPS members (37%).
6. The quote was from George Meany, ex-president of the AFL-CIO about Teamster
president Jimmy Hoffa in Robert F. Kennedy's, The Enemy Within: The McClellan
Committee's Crusade Against Jimmy Hoffa and Corrupt Labor Unions (New York: Da
Capo Press, 1990 [1960]), p. 161.
7. One of the first actions taken by the post-trustee leadership was to distribute
thousands of copies of the union's constitution and bylaws. The local also printed a
summary of the Landrum-Griffin Act in their local newspaper.
8. In the 1997 election a little less than 40% of the membership voted.
9. Louis Pike ran the local from 1964 to 1987 and Danny Ligurotis from 1987-1992.
10. Vote totals taken from "1998 IBT Election Vote" at http://members.aol.corn/
ibtvote/index.htm
REFERENCES
Barber, B. (1984). Strong Democracy: Participatory Politics for a New Age. Berkeley: University
of California Press.
Bruno, R. (2000). Democratic Goods: Teamster Reform and Collective Bargaining Outcomes.
Journal of Labor Research, 21 (1), 83-102.
Conboy, K. (1997). Decision of Election Officer for the International Brotherhood of Teamsters
(November 17).
Crowe, K. (1993). Collision: How the Rank and File Took Back the Teamsters. New York: Charles
Scibners Sons.
Daily Labor Report (1997). Teamsters Independent Review Board: Five Year Report 1992-1997,
United States v. IBT, 88 Civil 4486 (DNE), 218 (November 12-25): E3-E17.
Daily Labor Report (1999). Republican and Democratic Executive Summaries Of House
Subcommittee Report on Teamster, 37(February 25): E39-E48.
Dahl, R. (1961). Who Governs? Democracy and Power in an American City. New Haven: Yale
University Press.
Dine, P. (2000). James Hoffa is rekindling the unity in the Teamsters that his father bred. St. Louis
Post-Dispatch, (June 4), A12.
Edelstein, D., & Warner, M. (1976). Comparative Union Democracy. George Allen & Unwin.
Eisenscher, M. Leadership Development and Organizing: For What Kind of Union? Labor Studies
Journal, 24(2)(Summer), 3-21.
Fletcher, B. (1998). Whose Democracy? Organized Labor and Member Control. In: G. Mantsios
(Ed.), A New Labor Movement for the New Century. New York: Monthly Review Press.
Gruelle, M., & Parker, M. (1999). Democracy Is Power: Rebuilding Unions from the Bottom Up.
A Labor Notes Book. Detroit.
Lipset, S. M., Trow, M., & Coleman, J. (1956). Union Democracy. New York: Doubleday Anchor.
Luskin, R. (1987). Measuring Political Sophistication. American Journal of Political Science, 31,
856-899.
Kuklinski, J., Quirk, P., Jerit, J., & Rich, R. (2001). The Political Environment and Citizen Decision
Making: Information, Motivation, and Policy Tradeoffs. American Journal of Political
Science, 45(2)(April), 410-424.
Mannheim, K. (1956). Man and Society in an Age of Reconstruction. New York: Harcourt Brace.
Martin, R. (1968). Union Democracy: An Explanatory Framework. Sociology, 2(May), 205-220.
McConnell, G. (1958). Factionalism and Union Democracy. Labor Law Journal, 9, 635-f:~40.
Michels, R. (1949). Political Parties. Glencoe, Illinois: Free Press
122 ROBERT B R U N O
Modea, D. (1978). The Hoffa Wars: Teamster, Revels, Politicians and the Mob. New York:
Paddington Press.
Needlemen, R. (1998). Black Caucuses in Steel. New Labor Forum: A Journal of Ideas, Analysis
and Debate, (Fall/Winter), 41-56.
Nyden, P. (1985). Democratizing Organizations: A Case Study of a Union Reform Movement.
American Journal of Sociology, 90, 1179-1203.
Parenti, M. (1980). Democracy for the Few. New York; St. Martin's Press.
Parker, M. (1998). Appealing for Democracy. New Labor Forum: A Journal of ldeas, Analysis and
Debate, (Fall/Winter), 57-73.
Pateman, C. (1970). Participation and Democratic Theory. Cambridge, England: Cambridge
University Press.
Sayles, L., & Straus, G. (1953). The Local Union: Its place in the Industrial Plant. New York:
Harper and Brothers.
Sciacchitano, K. (2000). Unions, Organizing, and Democracy. Dissent, (Spring), 75-81.
Stephan-Norris, J., & Zeitlin, M. (1992). The Insurgent Origins of Union Democracy In: G. Marks
& L. Diamond (Eds), Reexamining Democracy in Honor of Seymour Martin Lipset. New
York: Sage.
Stephan-Norris, J. (1998). Strangers to Their Own Class? Sociological Inquiry, 68(3), 329-353.
Taft, P. (1948). The Constitutional Power of the Chief Officer in American Labor Unions. Quarterly
Journal of Economics, 6, 459-471.
Terry, M. (1996). Negotiating the government of Unions: Union democracy in theory and
practice. British Journal of Industrial Relations, 34,1 (March), pp. 87-110.
Union Labor Report (2000). Special Report: Teamsters Draft Plan for Reform Is Criticized and
Defended. 54(16)(April 20), 128.
REVITALIZING AFL-CIO
POLITICAL OUTREACH:
CAN A DIRECT INFORMATIONAL
CAMPAIGN DO THE TRICK?
Roland Zullo
ABSTRACT
Using an experimental design, the author examines the impact of AFL-

CIO political education outreach on union members' perception of and
preferences for, political candidates in three 1996 Wisconsin congressional
races. Contrary to the rational voter model results show that a direct
informational campaign has little effect on union member vote behavior.
Rather, members' preferences for candidates demonstrate stability during
the two months prior to the election, and change in members' perception
of candidates is associated with pre-outreach preferences. These results
support the psychological voter model, and suggest that organized labor
must build a culture of political mobilization to have a substantial impact
on electoral outcomes.

ISBN: 0-7623-0853-2
123
124 ROLAND ZULLO
INTRODUCTION
Labor leaders are keenly aware that more effective political action is crucial to
reversing the decline in power of their organizations. In 1995, the AFL-CIO
commissioned a survey to better understand how to persuade and motivate union
members during political campaigns. Survey responses indicated that union
members disliked unsubstantiated appeals for candidates, but they did believe
their unions play a positive role in supplying information on political issues and
candidate positions (Hart, 1995). Union members want to know why they should
vote for a candidate, not simply who to vote for. In response to these findings,
the Sweeny administration launched "Labor '96," a campaign designed to
increase labor's capacity to educate its membership. Since then, re-establishing
political power through member education has been a core objective of the
AFL-CIO, marking a contemporary strategic shift in resource use for organized
labor (Master, 1997).
Unfortunately, little empirical work is available that evaluates labor's political
education methods, despite a mention of the importance of this topic in a review
article by Masters and Delaney (1987). Research since then has confirmed that
union members prefer labor endorsed candidates over candidates opposed by
their organizations (Delaney et al., 1990; Form, 1995; Hojnacki & Baum, 1992;
Juravich & Shergold, 1988; Sousa, 1993). However, the mechanisms underlying
this relationship have not been adequately identified. Indeed, the very question
of whether union political activities cause members to choose endorsed
candidates remains unanswered.
A critical limitation of prior research is the inability to measure and isolate
the political outreach itself. Most empirical work to date has examined voting
behavior with respect to union status, not the receipt of political outreach.
Sousa (1983) analyzes nationwide data for the 1960 through 1988 presiden-
tial elections and, controlling for race, party identification and income, finds
that union status is positively associated with voting Democrat in most of the
elections. On the basis of these findings, he concludes " . . . there is a clear
group mobilization effect on the partisan direction [italics in original] of
unionists' voting behavior" (Sousa, 1993, p. 756). Like Sousa, Hojnacki and
Baum (1992) structure their dependent variable as a dichotomous preference
for Democratic candidates. They find differences between union and non-union
voters for local and national 1986 and 1988 elections in Ohio, and generally
attribute these results to the information, or "cues," directed to members
(Hojnacki & Baum, 1992). Delaney et al. (1990) examine the relationship
between union status and COPE endorsed candidates, a more precise depen-
dent measure than Democratic candidates, for House, Senate and Gubernatorial
Revitalizing AFL-CIO Political Outreach 125
races in selected congressional districts from a 1978 nationwide survey.

Controlling for income, political ideology, region and various demographic
factors, they report a strong positive relationship between union status and
preferences for COPE endorsed candidates. They then proceed to outline
several competing explanations for this finding.
While the totality of this research implies that union leaders are effective at
persuading union members, Delaney et al. (1990) were correct in cautioning
against this interpretation. By varying degrees, the candidate endorsement
process is a democratic one: in some locals union officers dominate candidate
endorsement decisions, while in others union leaders invite candidates to speak
at membership meetings and the endorsement is based on a general poll. Thus,
one limitation with the above research is the possibility for simultaneity bias:
the effect of communicating political endorsements to union members is
confounded with the effect that union members have on the organizational
endorsement process. It is necessary to untangle these effects to estimate the
impact that the organization has on its members.
Prior research that measured union political outreach has relied on one
survey instrument to collect data for the dependent and independent variables.
Juravich and Shergold (1988) do find a relationship between union literature
and preference for a presidential candidate in the 1984 election. However, this
effect was significant only for those respondents who recalled receiving
literature and who claimed the literature was useful. It is quite possible that
an expressed preference for a pro-labor candidate and a positive rating of
literature usefulness are factors for the same latent construct, such as the
strength of member identification with their union. Common method variation
problems such as this are exacerbated when the survey involves normative
behavior and the chance of social desirability bias is high. We see evidence
of social desirability bias in the work of Delaney et al. (1988), who found that
vote turnout tends to be overstated by survey respondents when checked
against municipal records. Likewise, it can be anticipated that union members
will provide biased responses to political surveys to match the communicated
positions of their organizations, particularly when the survey is sponsored by
labor and this sponsorship is revealed.
This study begins to address these limitations by employing experimental
design methods to test the effect of Wisconsin AFL-CIO political education
efforts during the 1996 national election. Experimental designs have been used
to measure the impact of political outreach in non-union field contexts (Bochel
& Denver, 1971; Eldersveld, 1956; Eldersveld & Dodge, 1954), and are
considered superior to cross-sectional survey methods (Ansolabehere et al., 1991;
Ansolabehere & Iyengar, 1995; Bowler et al., 1992). Using experimental
126 ROLAND ZULLO
techniques, I isolate the effect of the union political outreach, and test whether
the AFL-CIO information campaign influenced union members' perceptions of,
and preferences for, political candidates in three Wisconsin congressional races.
THEORIES ON INFORMATION AND

VOTING BEHAVIOR
There are two competing theoretical perspectives on voter response to political

information. The first begins with the premise that voters are rational actors who
weigh the cost of obtaining and analyzing information against the benefits of
doing so. Rational voters want to make political choices consistent with their
interests, and therefore information about issues and candidates is desired, but
gathering and digesting information carries a cost that these same rational
individuals will seek to minimize (Cohen & Rogers, 1983; Downs, 1957).
Consequently, voters rely heavily on political information that they can obtain
for zero cost, be it from advertising, the mass media, or outreach conducted by
politically active organizations.
Rational voters do desire a plurality of sources to provide checks against biased
reporting. Yet information processing is costly, and these constraints cause voters
to delegate the analysis and evaluation of political issues to a limited number of
sources. The tendency is to favor sources that focus on the facts germane to the
individual voter's decision-making, have similar goals, possess more data, and are
believed to have equal or better powers of judgment (Downs, 1957). It follows that
specific forms of information distributed by trusted advocacy organizations are
given more weight than general media sources. Such a model recognizes that sin-
gle-issue organizations, corporations, religious institutions, ethnic clubs and other
groups compete with organized labor to provide information to union members.
Extending this theoretical framework, Alvarez (1998) posits that candidate
evaluations are related to the precision of the information available to voters.
Ambiguity over a candidate's position on issues will tend to reduce the utility
of that candidate, thereby reducing the probability that the individual will
support that candidate at the ballot box. Further, the impact of information is
larger when uncertainty over a candidate is greatest, such as at the beginning
of an election cycle. Analyzing data from National Election Studies over
multiple election cycles, Alvarez (1998) finds a negative relationship between
voter uncertainty for a candidate's position on policy matters and voter
preference for that candidate. The inference drawn is that ambiguity tends to
penalize politicians: increasing the flow of information to voters on issues of
importance should shift preferences toward a candidate, particularly for those
voters who are less familiar with that candidate.
If the rational voter model is applied to the distribution of union-sponsored

political outreach, it is expected that as union member-voters cull information
they engage in an objective evaluation of the data with the aim of selecting
the candidate that most closely represents their interests. Since a significant
proportion of union members perceive the AFL-CIO as a valuable advocacy
organization and information source (Hart, 1995), then this model predicts that
union members would respond to AFL-CIO sponsored information that distin-
guishes the candidates along issues of importance. Information recipients should
display significant levels of crossover patterns (i.e. candidate preference switch-
ing) during the election cycle compared to those who do not receive union polit-
ical outreach. Likewise, members' perceptions of candidates should be positively
related to the intent of the mailings: perceptions of COPE endorsed candidates
should improve; while perceptions of COPE opposed candidates should be down-
graded. This leads to the following hypotheses:
HI: COPE outreach will be associated with a positive change in
union members' perception of the COPE-endorsed candidate.
H2: COPE outreach will be associated with a negative change in union
members' perception of the COPE-opposed candidate.
H3: COPE outreach will be associated with a shift in member prefer-
ences toward the COPE-endorsed candidate.
The field of political psychology offers an alternative model of voter response
to political information. This view begins with the premise that an individual's
political behavior is a function of a life-long socialization process. Deep-seated
political attitudes begin to form during childhood, establishing the foundation
for a partisan ideology that is quite resistant to change as a person matures
(Campbell et al., 1960; Greenstein, 1969; Hess & Torney, 1967; Himmelweit,
et al., 1981; Jennings & Niemi, 1968). Voters' decision-making processes are
directly influenced by their political predisposition. Where the rational voter
enters the political cycle with the aim of selecting candidates based on
utilitarian criteria, and valueless political symbols are ignored, the psycholog-
ical voter is socially conditioned toward political symbols, such as political
party affiliation. Voting Democrat is meaningful, in part, because a person grew
up in a household or community that favored Democrats, and such behavior is
consistent with kinship and societal norms.
As the psychological voter receives campaign information, there is a tendency
to cognize data that reinforce those long-standing political beliefs. Political
information is not digested objectively, but rather, information that supports
prior positions on campaign issues or political parties is retained while contra-
dictory information is discounted. In this way, information activates latent
128 ROLAND ZULLO
political predispositions and polarizes the electorate. Ansolabehere and Iyengar

(1995) advance this thesis through a series o f experimental tests of the impact
of political advertisements on vote preference. They report that advertising has
the intended effect on voters when the ad sponsor is the same political party
as the voter, particularly for voters with a low interest in politics (Ansolabehere
& Iyengar, 1995, Ch. 4 and Tables B2.4 and B2.5). Thus, exposure to infor-
mation is not enough. The recipient must accept the information, and
acceptance is conditional on the partisan political attitudes and beliefs culti-
vated over ones' lifetime.
The p s y c h o l o g i c a l perspective w o u l d therefore model voters as stubborn
in their preferences, changeable perhaps when the individual is i m m e r s e d in
a new cultural environment. Organized labor will impact union m e m b e r
voting b e h a v i o r when m e m b e r s b e c o m e successfully integrated with the polit-
ical socialization process. A n outreach c a m p a i g n that relies on issue
information only will have little effect. Operationally, we should observe no
evidence o f candidate preference switching linked to union political outreach.
Instead, change in m e m b e r s ' perception of candidates should be related to
their p r e - c a m p a i g n political positions. An additional hypothesis is therefore
needed:
H4: Change in members' preferences toward the candidates should be

positively related to their pre-campaign preferences.
WISCONSIN AFL-CIO POLITICAL EDUCATION

STRATEGY
In 1996, following the lead of the national A F L - C I O , the Wisconsin A F L - C I O

launched its own project called the "Campaign for Working Families: Moving
Wisconsin Forward Again." The initial document outlining the Campaign stated,
A recent poll commissioned by the National AFL-CIO shows that union members yearn
for a new voice. Union members want to be educated on the issues and the candidates in
order to make informed decisions. We need to find new ways to engage more union members
in the electoral and legislative process. Working people need to vote for their economic
interests and their jobs more than ever. Those members who have special concerns about
guns or abortion still need to vote their interests as working families first - then lobby their
special interests later (WI-AFL-CIO, 1996).
To implement this agenda a statewide electoral coalition was created, modeling
itself after similar efforts led by central labor councils in Milwaukee and South
Central Wisconsin (Eimer, 1999). These coalitions brought together the A F L -
CIO, the Allied Senior Citizens, Citizen Action, n o n - A F L CIO unions such as
the Wisconsin Education Association, and unions with above-average interest
in politics, like AFSCME. For the 1996 elections, the statewide and local coali-
tions established a division of labor: the state coalition was responsible for
developing and distributing political education material, while the local unions
and central labor councils engaged in grass-roots activities designed to increase
voter turnout.
Direct mail pieces sent to members' homes was the core education tactic.
The timing, sequence, and content of the mailings were intended to gradually
build support for the COPE endorsed candidate. An initial brochure introduced
the program and educated members on economic issues, such as the erosion of
family income. No mention of the candidates appeared in the piece. The second
mailer continued the working families theme, comparing the presidential
candidates on seven key issues and providing voting records for all Wisconsin
congressional incumbents on economic issues. Though Clinton and the
Democratic Congressional incumbents were favorably reviewed, the brochure
never urged support for any particular candidate. In contrast, the third mailing
directly compared the Republican and Democratic congressional candidates on
six key issues: strengthening pension security law, raising the minimum wage,
protecting Medicare funding, expanding health care, raising OSHA standards,
and funding public education. Voting records on specific bills or position papers
by each candidate supported the ratings. Each candidate's position was
summarized as either being "right" or "wrong" for working families. In one
section, the literature suggested, "the record is clear" and prominently displayed
a picture of the candidate that "supports working families." The fourth piece
of literature was a simple postcard designed to remind members whom the
AFL-CIO had endorsed, and to get out the vote. This study tests the effect of
the first three pieces of educational literature.
DESIGN, DATA AND MEASURES
Experimental Field Test
A Pretest-Posttest Control Group design is used (Campbell & Stanley, 1963),

where a random sample of union members in the outreach population were
surveyed prior to the distribution of the union literature (survey one). Using a
random number generator, respondents were assigned to one of three groups: (1)
a control group that received no outreach; (2) a group that received only the first
two union literature pieces; and (3) a group that received all three union
literature pieces. The control group was purged from the state AFL-CIO mailing
list before any literature was sent. The second group was purged from the list
130 ROLAND ZULLO
after the second mailing. Respondents from all groups were removed from local
AFL-CIO lists to minimize the noise created by the outreach efforts of central
labor councils. 2 The subjects remained in their assigned group for the duration
of the study, and were surveyed again with identical measures after the union
sponsored literature was distributed, but just before the election (survey two).
There are three major advantages to this design. First, the independent factor
of interest (i.e. the union political outreach) is administered separately from the
dependent measures collected through the survey process. Second, the panel data
permit a test for outreach effects on the change in dependent measures during
the observation period. Third, this design capitalizes on random assignment to
distribute unmeasured factors randomly across groups, and thereby reducing the
potential for omitted variable bias.
Sampling and Survey Administration
Union members who were scheduled to receive educational literature were

sampled from Wisconsin congressional districts 1, 2 and 4 (n = 1,990). These
three districts in southern Wisconsin were chosen because the State AFL-CIO
considered them close political contests. At least five attempts to contact each
prospective respondent were made per survey. There were 1,044 completed
wave one surveys, and 865 wave two surveys, representing a panel response
rate of about 43%. 3 The distribution of respondents across the three groups was
as follows: control, 317; 2 mail pieces, 216; and 3 mail pieces, 332. 4 Survey
two respondents were asked whether they recalled receiving AFL-CIO
literature. Affirmative responses to this question were compared against the
treatment groups (r 2 = 0.09) for a manipulation check.
As a precaution against administrator related bias, both surveys were
conducted over the telephone by the same survey agency, and the agency was
never informed of the treatment status of respondents. Further, to minimize
social desirability bias, union sponsorship was concealed in the survey, and an
attempt was made to complete the second survey prior to the announcement of
the November 5 election results. 5 Finally, to reduce the risk of pretest
respondent sensitization to the union literature, the survey language was
designed to resemble a neutral poll, differing substantially in both tone and
content from the union outreach.
Dependent Measures
Two dependent measures are analyzed: the perception of the COPE-endorsed

and COPE-opposed candidates, and the stated preference for a candidate. While
the two measures, perception and preference, are related, it is advantageous to

include both. Perceptions are considered more malleable than preferences
(Alvarez, 1998, p. 21), and therefore more sensitive to the effects of political out-
reach. Changing a member's preference is believed to be more difficult, and from
the perspective of labor leaders is the true test of outreach effectiveness.
Perception of Candidate
Union members' perception of political candidates was measured by a feeling
thermometer, which has a long history of usage in the National Election Surveys
(ICPSR, 1993) and has been applied to union surveys (Form, 1995). The feeling
thermometer asks respondents to rate how "warm" or "cold" they feel about a
congressional candidate based upon a 100 point scale (Sudman & Bradburn,
1983, Ch. 6). The primary advantage of the feeling thermometer is that it is a
reliable indicator of voter intentions that can be efficiently applied to a long
list of candidates. In addition, because the thermometer readings are
graduated by 100 degrees, it is sensitive enough to detect relatively small
changes in the perceptions of candidates, while serving to guard against memory
related bias. A thermometer rating was collected for the COPE endorsed and
COPE opposed congressional candidates in both survey waves. This dependent
variable was analyzed using ordinary least square techniques.
Candidate Preference
Candidate preference was assessed with the question, "Who do you think you
will vote for in the congressional race between _ _ and ? " If members
expressed a preference for a congressional candidate, they were then asked,
"Would you say that your preference for this candidate is strong or not strong?"
The question wording was borrowed from the most recent National Election
Survey (ICPSR, 1993), and applied only to respondents who were familiar with
both congressional candidates. The wording and sequence of these questions
were identical in both surveys.
Responses to the preference questions were used to categorize union members
according to their level of commitment for the candidate endorsed by the
Wisconsin AFL-CIO. A four point ordinal scale was created with these data, with
higher values corresponding to stronger support for the COPE-endorsed candi-
date: (1) members that were strong for the COPE-opposed candidate, (2)
members that were leaning for the COPE-opposed candidate, (3) members that
were leaning for the COPE-endorsed candidate, and (4) members that were
strong for the COPE- endorsed candidate. This dependent variable was analyzed
using ordered probit models.
132 ROLAND ZULLO
Independent Measures
Political Outreach
The chief independent variable of interest is the union political outreach. As
mentioned above, some respondents received no outreach, a second group
received the first two mailings, and the third received the complete outreach
package. This separation was made to understand whether there were any empir-
ical differences based on the content and tone of the outreach. The first two
pieces provided more general information about the issues and the incumbents,
while the third piece was comparative and overtly directional, distinguishing
the candidates along several contemporary political issues. Indicator variables
were created for union members who received the first two informational pieces
and those who received the third comparative piece.
Pre-Outreach Perception of Candidate

The thermometer score from survey one is a measure for member's perception
of the candidates prior to any union outreach. It is used to control for the stabil-
ity in perceptions during the election cycle.
Pre-Outreach Candidate Preference

Responses to the preference question in survey one were used to indicate candi-
date preferences prior to the distribution of outreach. A categorical variable
composed of five mutually exclusive indicators was created: (1) strong for COPE-
endorsed candidate (Strong For COPE), (2) leaning for the COPE-endorsed
candidate (Leaning For COPE), leaning for the COPE-opposed candidate (Lean
Against COPE), strong for the COPE-opposed candidate (Strong Against COPE)
and those respondents with an unknown preference (Unknown). Respondents
categorized as unknown include a mix of those who could not identify one or both
of the candidates, those who knew the candidates but expressed no preference, and
those who refused to answer the question.
A second pre-candidate preference variable was constructed that collapses
the information in the five indicator categories into one continuous measure.
This was achieved by assigning a value of zero to those respondents who
expressed strong support for the COPE-opposed candidate, a one for those
leaning toward the COPE-opposed candidate, a two for those who were
unknown, a three for those leaning for the COPE-endorsed candidate, and a
four for those who expressed strong support for the COPE-endorsed candidate.
This construction imposes both order and interval constraints on the preference
variable, and is used to test for a linear relationship between pre-candidate
preference and change in member perception of candidates.
Table 1. Variable Definitions and Descriptive Statistics.
Variable Definition Mean (s.d.)
Perception of COPE Thermometer rating for COPE endorsed congressional 54.091

Endorsed (Survey One) candidate in survey 1 (range: 0 to 100) (24.525)
Perception of COPE Thermometer rating for COPE endorsed congressional 56.857
Endorsed (Survey Two) candidate in survey 2 (range: 0 to 100) (26.079)
Perception of COPE Thermometer rating for COPE opposed congressional 45.872
Opposed (Survey One) candidate in survey 1 (range: 0 to 100) (25.634)
Perception of COPE Thermometer rating for COPE opposed congressional 45.198
Opposed (Survey Two) candidate in survey 2 (range: 0 to 100) (26.209)
Candidate Preference Ordered scale indicating the preference level for the 0.622
(Survey Two) COPE endorsed candidate in survey 2 (range: - 2 to 2) (1.665)
No Outreach Indicator for respondents that received no political 0.366
(Control Group) information mailings (1 = Yes; 0 = Other) (0.482)
First 2 Mail Pieces Indicator for respondents that received 2 political 0.634
information mailings (1 = Yes; 0 = Other) (0.482)
3rd Mail Piece Indicator for respondents that received 3 political 0.384
information mailings (1 = Yes; 0 = Other) (0.487)
Strong For COPE Indicator if respondent expressed strong preference for 0.200
COPE endorsed candidate in survey 1 (1 = Yes; 0 = Other) (0.400)
Lean For COPE Indicator if respondent expressed not strong preference for 0.089
COPE endorsed candidate in survey 1 (1 = Yes; 0 = Other) (0.285)
Unknown Indicator if respondent did not register a candidate 0.517
preference in survey 1 (1 = Yes; 0 = Other) (0.500)
Lean Against COPE Indicator if respondent expressed not strong preference 0.066
for COPE opposed candidate in survey 1 (0.248)
(1 = Yes, 0 = Other)
Strong Against COPE Indicator if respondent expressed strong preference 0.128
for COPE opposed candidate in survey 1 (0.335)
(1 = Yes, 0 = Other)
Linear Preference Ordered scale indicating the preference level for the 2.166
COPE endorsed candidate in survey 1 (range: 0 to 4) (1.201)
AFSCME Indicator for respondents who are members of the 0.252
American Federation of State County Municipal Employees (0.434)
(1 = Yes; 0 = Other)
UAW Indicator for respondents who are members of the 0.212
United Auto Workers (1 = Yes; 0 = Other) (0.409)
IBEW Indicator for respondents who are members of the 0.046
International Brotherhood of Electrical Workers (0.210)
(1 = Yes; 0 = Other)
UFCW Indicator for respondents who are members of the United 0.081
Food and Commercial Workers (1 = Yes; 0 = Other) (0.273)
134 ROLAND ZULLO
Union Affiliation
Indicator variables were created for the four largest unions represented in the
sample: the American Federation of State, County and Municipal Employees
(AFSCME), the United Auto Workers (UAW), the International Brotherhood of
Electrical Workers (IBEW) and the United Food and Commercial Workers
(UFCW). The purpose of union affiliation was to explore the possibility of an
effect on member voting behavior caused by the outreach performed by the
regional offices of international unions. Definitions and descriptive statistics for
the measures 6 are provided in Table 1.
EMPIRICAL SPECIFICATIONS
As described above, one goal of this study was to provide a direct test of the
causal relationship between union outreach and union member perceptions of
political candidates. Accordingly, these panel data are analyzed using two-wave
linear models, where the dependent measures are a function of the union polit-
ical outreach, the stated strength of member preferences in wave one, and union
affiliation. Equation (1) is the empirical specification for member perception:
Pt = [30 + [31Pt-J + [321+ [33Mt-i + [34U + • (1)
Where Pt is member perception of candidates just prior to the election (survey

tWO), Pt-I is member perception of candidates before the outreach (survey one),
I is the union political information, Mt_ l is strength of member preference at
survey one, and U is union member affiliation. Unbiased regression coefficients
are given by 13, and e is a normally distributed error term.
A similar specification is used to test the effect of union political outreach on

candidate preference just prior to the election, denoted as C. Equation (2)
provides the specification for member preference:
C = [30 + [311+ [32Mt_l + [33U (2)
Probabilistic values for candidate preference are expressed as a linear function

of the receipt of outreach (I), the strength of member preference at survey one
(Mt_l), and union member affiliation (U).
Table 2. OLS of Union Member Perception of COPE Endorsed

Congressional Candidate (Standard Errors in Parentheses).
Independent Variable Model 2.1 Model 2.2 Model 2.3 Model 2.4
Survey One Percept!on 0.715"** 0.572*** 0.578*** 0.570***

(0.031) (0.035) (0.350) (0.035)
No Outreach Omitted Omitted Omitted Omitted
First 2 Mail Pieces 4.527 ** 3.688 3.912" 3.434
(2.011) (1.924) (1.927) (1.907)
3rd Mail Piece 3.051 2.973 3.161 2.682
(1.982) (1.894) (1.898) (1.875)
Strong For COPE 6.362** 5.614"*
(1.941) (1.940)
Lean For COPE 4.541 4.481
(2.532) (2.510)
Unknown Omitted Omitted
Lean Against COPE 10.251"** 10.155"**
(2.990) (2.961)
Strong Against COPE 13.074"** - 12.471"**
(2.287) (2.268)
Linear Preference 4.865***
(0.636)
AFSCME 0.687
(1.787)
UAW 7.400***
(2.102)
IBEW -2.218
(3.531)
UFCW -4.919
(2.807)
Constant 16.988"** 25.979*** 13.953"** 25.310"**
(2.135) (2.500) (2.083) (2.586)
Observations 637 637 637 637
Model R 2 0.452 0.503 0.498 0.519
Adjusted R 2 0.450 0.498 0.495 0.510
* p < 0.05; ** p < 0.01 ; *** p < 0.001 (all two-tailed tests).
RESULTS
To understand the effect of political information on union member perceptions

of candidates, parallel OLS regressions were performed for the COPE-endorsed
congressional candidate and COPE-opposed congressional candidate responses.
136 ROLAND ZULLO
The results for member perceptions of the COPE-endorsed candidate are in

Table 2. For all models, the thermometer score for COPE-endorsed candidate
in survey one was a strong predictor of scores in survey two. This suggests
that there is considerable stability in the relative perceptions of candidates
among union members during the two months preceding the election.
Model 2.1 indicates that the overall perception scores for the COPE-endorsed
congressional candidates increased during the two-month time period before the
election. For those respondents in the control group, perceptions of the COPE-
endorsed candidates increased an average of 17 thermometer points. Political
information appears to have had some effect on improving the perception of
the COPE-endorsed candidate, indicated by positive and statistically significant
coefficient for the first two pieces of literature. However, the magnitude of the
impact is not substantial: receiving two pieces of political literature was
associated with an average thermometer score increase of 21.5: only a 4.5 point
gain over the control group. The marginal effect of the third piece was
negative, but statistically insignificant.
To explore whether changes in perception are related to pre-outreach
political preference, Model 2.2 decomposes perception responses by strength of
candidate preference in survey one. Strong early supporters of the COPE-
endorsed candidate demonstrated the greatest change in their perception of the
COPE-endorsed candidate during the two months prior to the election. They
had an average thermometer increase of 32 points, followed by respondents
leaning toward the COPE-endorsed candidate, unknown respondents, respon-
dents leaning against the COPE-endorsed candidates, and finally those
respondents who were strong against the COPE-endorsed candidates. A test
between Model 2.2 and nested Model 2.1 is statistically significant (F = 16.3;
p < 0.001), indicating that the pre-outreach candidate preference variable adds
explanatory power. The introduction of the candidate preference variable also
reduces the estimated effect of the union outreach.
Model 2.3 replaces the categorical pre-outreach preference variable with one
that is continuous. The aim was to impose linearity on this variable to perform
a more stringent test of the psychological perspective. Consistent with Model
2.2, the coefficient for Linear Preference is positive and statistically significant
(p < 0.001). A one unit movement along the linear preference scale (i.e. stronger
stated preference toward the COPE-endorsed candidate) produces an increase
of approximately 5 thermometer points. Moreover, a comparison between Model
2.2 and 2.3 indicates no significant decline in predictive power by using the
constrained variable (F = 2.16; p = n.s.).
Model 2.4 introduces indicator variables for respondents affiliated with
AFSCME, UAW, IBEW and UFCW. The omitted group includes respondents
with other union affiliations. U n i o n m e m b e r s affiliated with the U A W increased

their perceptions of the C O P E - e n d o r s e d candidate significantly m o r e than other
union m e m b e r s in the sample. A test that this set of variables is zero is rejected
(F = 4 . 9 9 ; p < 0.001), i m p l y i n g a relationship b e t w e e n union affiliation and
change in candidate perception.
Table 3 offers a parallel set o f m o d e l s for m e m b e r perceptions o f the C O P E -
o p p o s e d candidate. Consistent with the results f r o m Table 2, the survey one
Table 3. O L S o f U n i o n M e m b e r Perception o f C O P E O p p o s e d
C o n g r e s s i o n a l Candidate (Standard Errors in Parentheses).
Independent Variable Model 3.1 Model 3.2 Model 3.3 Model 3.4
Survey One Perception 0.712"** 0.602*** 0.608*** 0.575 ***

(0.036) (0.042) (0.041) (0.042)
No Outreach Omitted Omitted Omitted Omitted
First 2 Mail Pieces 0.094 -0.088 0.274 0.394
(2.408) (2.364) (2.358) (2.346)
3rd Mail Piece 1.264 1.590 1.678 1.222
(2.370 (2.332) (2.323) (2.306)
Strong For COPE -6.058* 7.134**
(2.377) (2.366)
Lean For COPE -0.134 1.878
(3.014) (3.001)
Lean Against COPE 2.515 0.554
(3.362) (3.358)
Strong Against COPE 9.732** 8.065**
(2.730) (2.867)
Linear Preference 3.568***
(0.730)
AFSCME 2.363
(2.339)
UAW 9.433***
(2.394)
IBEW 0.846
(4.310)
UFCW 0.351 \
(3.137)
Constant 12.399"** 16.811"** 24.914"** 21.882"**
(2.281 ) (2.730) ( 3.399) ( 3.141 )
Observations 537 537 537 537
Model R2 0.430 0.457 0.454 0.474
Adjusted R2 0.426 0.449 0.450 0.463
* p < 0.05; ** p < 0.01; *** p < 0.001 (all two-tailed tests).
138 ROLAND ZULLO
perceptions of candidates are strong predictors of survey two perceptions,

offering evidence of stability during the campaign cycle. More germane to our
theoretical question, however, is the absence of an effect from the union polit-
ical outreach. According to model 3.1, respondents increased their thermometer
ratings of the COPE-opposed candidate by an average of 12 points, but the
political outreach variables add little explanatory power. However, it does
appear that the introduction of the pre-outreach candidate preference variables
improves the fit of the model. A test comparing Models 3.1 and 3.2 is statis-
tically significant (F = 6.56; p < 0.001).
Most notable in Model 3.2 is the pattern of change exhibited by the
pre-outreach candidate preference variables. The greatest movement toward the
COPE-opposed candidate, an average of 27 thermometer points, is from those
who expressed strong early support for the COPE-opposed candidate. The
ranking of the magnitude of change in the thermometer scores then follows a
pattern that is exactly inverse to the expressed support for the COPE-endorsed
candidate: those leaning toward the COPE-opposed candidate, those respon-
dents whose preference was unknown, those leaning for the COPE-endorsed
candidate, and strong COPE-endorsed supporters.
Once again, the linear preference variable is in the expected direction and
statistically significant (p < 0.001). A one unit increase in the linear prefer-
ence scale is associated with about a 3.5 point decline in perception of the
COPE opposed candidate. A test comparing Models 3.2 and 3.3 indicate
no significant loss of predictive power with the linear preference variable
( F = 0.80; p = n.s.). Thus, pre-outreach candidate preference is linearly
associated with the change in perception for both the COPE-endorsed and
COPE-opposed candidates. The finding of a relationship between the expressed
strength of support for a candidate and changes in the perception of that
candidate during the election cycle support the psychological perspective.
Model 3.4 includes the four union affiliation variables. As a group, these
indicators provide only marginal improvement in the model (F = 2.07;
p < 0.085). Members of the UAW, however, do demonstrate suppression in
their change in candidate perception. Estimates from Model 3.4 are that the
increase in perception of the COPE-opposed candidate was approximately seven
points less for UAW members compared to other union affiliations.
The effect of political information on candidate preference switching was tested

using ordered probit models, where the dependent variable indicates the strength
of preference for the COPE endorsed candidate just prior to the election. Table
Table 4. O r d e r e d Probit of U n i o n M e m b e r C o n g r e s s i o n a l C a n d i d a t e
Preference (Standard Errors in Parentheses).
Independent Variable Model 4.1 Model 4.2
No Outreach Omitted Omitted

First 2 Mail Pieces 0.009 -0.033
(0.132) (0.134)
3rd Mail Piece 0.180 -0.193
(0.129) (0.130)
Strong For COPE 1.145"** 1.187"**
(0.150) (0.153)
Lean For COPE 0.124 0.143
(0.165) (0.166)
Lean Against COPE 0.753*** 0.717"**
(0.188) (0.190)
Strong Against COPE 1.602"** -1.544"**
(0.162) (0.163)
AFSCME 0.045
(0.129)
UAW 0.269
(0.139)
IBEW -0.333
(0.232)
UFCW -0.488*
(0.194)
ixl 1.049 --1.090
(0.105) (0.123)
t~2 0.618 -0.652
(0.100) (0.118)
ix3 0.012 0.034
(0.098) (0.116)
Observations 582 582

- Log L 595.905 588.213
Model ×2 233.45 248.84
d.f. 6 10
* p < 0.05; ** p < 0.01; *** p < 0.001 (all two-tailed tests).
4 estimates the i m p a c t o f the union outreach on pre-election candidate prefer-

ence (survey two), controlling for pre-outreach candidate preference.
T h e e v i d e n c e suggests that the political outreach had no effect on m e m b e r
preferences. Coefficients for 2 mail pieces and 3 m a i l pieces are statistically
insignificant for all models. Rather, the findings show that candidate preferences
140 ROLAND ZULLO
are relatively robust during the political campaign cycle. Predicted probabilities
from Model 4.1 are that 86% of those who strongly favored the COPE-endorsed
candidate in August held the same view just before the election. Similarly, 74%
of those who strongly favored the COPE-opposed candidate maintained their
preference status.
Model 4.2 includes the union affiliation variables. A likelihood ratio test
indicates that the four union variables contribute significantly to the model fit
(X2 = 15.38; p < 0.01). Further evaluation of the coefficients demonstrates
remarkable variation across affiliated groups. The predicted probability that a
UAW member was either strong or leaning toward the COPE-endorsed
candidate just before the election was 81%, followed by AFSCME at 65%, the
IBEW at 61%, and the UFCW at 49%. Understanding this variation will
undoubtedly require the inclusion of demographic, organizational and economic
factors that vary across union groups. However, the fact that these results were
obtained after controlling for the pre-outreach candidate preferences implies that
the outreach tactics by regional and local affiliates of these internationals explain
some of the variation.
Discussion and Limitations
The aim of this study was to test for an effect of COPE political education
tactics on union member perceptions of, and preferences for, political candi-
dates. In 1996, the Wisconsin AFL-CIO mailed a series of informational
brochures on congressional candidates and campaign issues to the homes of
union members. The outreach went beyond an endorsement: the brochures
presented candidate positions on issues that were identified as important to union
members, and the information contained in the brochures was referenced to
candidate position papers and voting records. The WI AFL-CIO used this direct
informational tactic to educate union members on the issues and ultimately
persuade them to vote for COPE-endorsed candidates. This study tested the
effectiveness of that approach using experimental methods.
Contrary to rational voter theory of information and political behavior, the
results provide little evidence that the outreach changed union members' voting
pattems. A modest impact on union members' perception of the COPE-endorsed
candidate was detected in the positive association between the first two mail-
ings and union member perceptions of the COPE endorsed candidate. However,
the coefficient for this measure becomes statistically marginal with the intro-
duction of pre-campaign preference measures. Moreover, the third brochure,
which contained the clearest comparative data, produced a coefficient that was
negative and statistically insignificant. If union members were objectively
digesting AFL-CIO information for electoral guidance, then one would expect
a positive and sizeable effect from this final mailing. Further, the outreach had
no impact on members' perception of the COPE-opposed candidate. Even more
critical for labor, there was no association between the union sponsored outreach
and shifts in union member preference toward the COPE-endorsed candidates.
Returning to the formal tests, the weak support for hypothesis one and the
rejection of hypotheses two and three stand in contradiction to the predic-
tions based on the rational voter model.
Why was the effect of the WI AFL-CIO outreach so miniscule? It may be
that the AFL-CIO's effects are cumulative; shifting member preferences toward
labor candidates gradually over multiple election cycles, and this study was
unable to capture an effect due to a limited observation period. It may also be
that local union bodies engaging in member communications at the workplace
and in communities generate the most powerful forms of union persuasion.
While this study did not directly measure local union activities, the union affil-
iation variables did demonstrate predictive power. Affiliation with the UAW,
in particular, was associated with significant movement toward the COPE-
endorsed candidate and away from the COPE-opposed candidate. In
post-election interviews, UAW officials described their political outreach as an
ongoing, member-to-member process, which relies heavily on a strong steward
system for the transmission of political information. No other union in the
sample had such a comprehensive workplace-level strategy for political mobi-
lization. The implication from these preliminary clues is that the most effective
political outreach takes place at the union local or regional level, where grass-
roots political activities are an institutionalized dimension of union culture.
Additional results from this study support the psychological voter model. One
finding is that union members have opinions on candidates prior to exposure
to any union political outreach, and for the majority these opinions are robust
up to Election Day. This implies that the AFL-CIO may have to offer political
education well before election time, perhaps on a continual basis, and in a form
that distinguishes union sponsored information from competing political
messages. A second, more compelling result was that the change in candidate
perception leading up to the election tends to reinforce pre-campaign candidate
preferences, supporting hypothesis four along with the theory that voters selec-
tively evaluate information from general media sources to match political
predispositions. If true, then organized labor may have to merge political
mobilization activities with programs that build stronger member identification
with the labor movement to overcome the tendency for voter polarization. In
the very least, changing union member voting behavior seems difficult, and to
meet the challenge organized labor will probably have to direct resources toward
142 ROLAND ZULLO
member-to-member outreach techniques. Experimental tests such as this would

help labor evaluate local-level outreach techniques, hone their strategy, and
effectively allocate resources.
In drawing these conclusions, it is important to emphasize that this research
was limited to three congressional districts in Wisconsin, and that only the
informational activities of the W I A F L - C I O were tested. Future research on
the relationship between A F L - C I O outreach and u n i o n voter behavior should
be extended to other political districts, and to the activities of local and
regional organizations. Testing a broader variety of A F L - C I O outreach
tactics, such as candidate persuasion calls and get-out-the-vote drives, might
prove valuable.
NOTES
1. The fourth piece was scheduled to arrive at union households at about the time the
survey for this study was to be administered. Because I could not be reassured that the
fourth mailing would arrive before the survey, it was omitted from the analysis.
2. Internationals and locals often conduct political outreach that is independent from
the efforts by the state and regional AFL-CIO bodies. I was unable to censor survey
respondents from these other forms of union outreach, and therefore must rely on the
random assignment process to control for these effects.
3. The non-response breakdown is as follows: refusals, 374 (19% of sample); cannot
locate (usually due to the lack of a telephone number) 285 (14% of sample); other
non-response, 441 (22% of sample). In addition, 17 were unusable because they were
located outside of the three congressional districts, and 8 were omitted because of
returned mail.
4. Since respondents in the group that received three mall pieces also received the
first two, the actual number of respondents that are exposed to the first two pieces of
literature is 548.
5. Over 95% of the second surveys were completed before election results were
announced on November 5.
6. Numerous other measures were collected in the survey, including the demographic
variables age and sex. While age and sex are often systematically associated with
political perceptions and preferences, tests indicated that these variables were
distributed randomly across the three groups and that they did not interact with the
union outreach. Age and sex were therefore omitted from the analysis to present a
parsimonious report.
ACKNOWLEDGMENTS
The author gratefully acknowledges Heejoon Park, Lydia Li, Craig Olson and
Stuart Eimer for their comments on earlier drafts. This research was funded by
a grant from the Wisconsin AFL-CIO.
REFERENCES
Alvarez, R. M. (1998). Information & Elections. Ann Arbor: University of Michigan Press.
Ansolabehere, S., Behr, R., & Iyengar, S. (1991). Mass Media and Elections: An Overview.
American Politics Quarterly, 19(1), 109-139.
Ansolabehere, S., & lyengar, S. (1995). Going Negative: How Attack Ads Shrink and Polarize the
Electorate. New York: The Free Press.
Bochel, J. M., & Denver, D. T. (1971). Canvassing, Turnout and Party Support: An Experiment.
British Journal of Political Science, 2, 257-269.
Bowler, S., Broughton, D., Donovan, T., & Snipp, J. (1992). The Informed Electorate? Voter
Responsiveness to Campaigns in Britain and Germany. In: S. Bowler & D. M. Farrell (Eds),
Electoral Strategies and Political Marketing (pp. 204-222). NY: St. Martins Press.
Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American Voter. New
York: John Wiley & Sons, Inc.
Campbell, D. T., & Stanley, J. C. (1963). Experimental and Quasi-Experimental Designs for
Research. Boston: Houghton Mifflin.
Cohen, J., & Rogers, J, (1983). On Democracy. New York: Penguin.
Delaney, J. T., Masters, M. F., & Schowchau, S. (1988). Unionism and Voter Turnout. Journal of
Labor Research, IX, 221-236.
Delaney, J. T., Masters, M. F., & Schowchau, S. (1990). Union Membership Voting for COPE-
Endorsed Candidates. Industrial and Labor Relations Review, 43(5), 621-635.
Downs, A. (1957). An Economic Theory of Democracy. New York: Harper & Row.
Eimer, S. (1999). From 'Business Unionism' to 'Social Movement' Unionism: the Case of the
AFL-CIO Milwaukee Labor Council. Labor Studies Journal, 24(2), 63-81.
Eldersveld, S. J. (1956). Experimental Propaganda Techniques and Voting Behavior. The American
Political Science Review, L, 154-165.
Eldersveld, S. J., & Dodge, R. W. (1954). Personal Contact or Mail Propaganda? An Experiment
in Voting Turnout and Attitude Change. In: D. Katz, D. Cartwright, S. Eldersveld &
A. McClung Lee (Eds), Public Opinion and Propaganda. NY: Holt, Rinehart and Winston.
Form, W. (1995). Segmented Labor, Fractured Politics. New York: Plenum Press.
Greenstein, F. I. (1969). Children and Politics. New Haven CT: Yale University Press.
Hart and Associates, Inc. (1995). A Nationwide Survey Among Union Members and the General
Public on Politics and Legislation. Report to the AFL-CIO, May.
Hess, R., & Tomey, J. (1967). The Development of Political Attitudes in Children. Chicago: Aldine.
Himmelweit, H. T., Humphreys, P., Jaeger, M., & Katz, M. (1981). How Voters Decide: A
Longitudinal Study of Political Attitudes and Voting Extending Over Fifteen Years. London:
Academic Press.
Hojnacki, M., & Baum, L. (1992). New Style Judicial Campaigns and the Voters: Economic Issues
and Union Members in Ohio. The Western Political Quarterly, 45, 921-948.
ICPSR, (1993). American National Election Study, 1992: Pre- and Post-Election Survey. Inter-
University Consortium for Political and Social Research. Ann Arbor.
Jennings, M. K., & Neimi, R. G. (1968). The Transmission of Political Values from Parent to
Child. The American Political Science Review, LX/l(1)(March), 169-184.
Juravich, T., & Shergold, P. (1988). The Impact of Unions on the Voting Behavior of Their
Members. Industrial and Labor Relations Review, 4•(3), 374-385.
Master, B. (1997). A New Political Strategy For American Unions. Working USA, M.E. Sharpe,
Inc. September-October, 16-29.
144 ROLAND ZULLO
Masters, M. F., & Delaney, J. T. (1987). Union Political Activities: A Review of the Empirical
Literature. Industrial and Labor Relations Review, 40(3), 336-353.
Patton, D. B., & Marrone, J. J. (1984). The Impact of Labor Endorsements: Union Members and
the 1980 Presidential Vote. Labor Studies Journal, (Spring), 3-18.
Sousa, D. J. (1993). Organized Labor in the Electorate. Political Research Quarterly, 46, 741-758.
Sudman, S., & Bradburn, N. (1983). Asking Questions: A Practical Guide to Questionnaire Design.
DC: Jossey-Bass.
Wisconsin State AFL-CIO, (1996). Campaign for Working Families: Moving Forward Again. Issue
Statement.
THE FEARS OF RESOURCE
STANDARDIZATION AND THE
CREATION OF AN ADVERSARIAL
WORKPLACE CLIMATE:
THE STRUGGLE TO ORGANIZE
A FACULTY UNION AT
ILLINOIS STATE UNIVERSITY
Victor G. Devinatz
ABSTRACT
In this article, I investigate the effect that the communication of an anti-

union message delivered by faculty union opponents and the university
administration had on a recent faculty union organizing drive at Illinois
State University. Specifically, I argue that the Illinois State University
Faculty Association's (ISUFA) certification election loss was due to the
union opponents' effective communication of a message that having a
faculty union and collective bargaining representation would impose more
costs than benefits through the standardization of college and departmental
resources across the university and by creating an adversarial climate
between the administration and faculty members. Conversely, the union

ISBN: 0-7623-0853-2
145
146 VICTOR G. DEVINATZ
was ineffective in combating these two arguments by convincing faculty of

the overriding benefits of unionization. The ISUFA failed to build a suffi-
cient "community of interest" (or solidarity) among the faculty by focusing
on a single issue or a set of related issues that extended to university-wide
issues as a whole or to external issues that confronted the faculty as a
profession. The article concludes with lessons learned from the campaign.
INTRODUCTION
For many people, academic labor is not really considered to be labor. Even the
lowest rung of academic labor, graduate research and teaching assistants, are
often considered by many people to be a privileged class of individuals. For
example, witness the attitude of the Omaha World-HeraM in its analysis of the
Yale teaching assistants' (TA) grade strike at the end of the 1995 fall semester
in the Graduate Employees Student Organization's drive to achieve collective
bargaining with the Yale University administration. In spite of facing long and
arduous graduate programs, an academic job market that would provide few of
these TAs with tenure-track positions upon the completion of their doctorates,
and average compensation a bit less than $10,000 per year in 1995, Omaha's
newspaper opined:
Yale's 721 teaching assistants are already part of the Information Age's new aristocracy
... At Yale, they will begin building lifelong networks of other influential people, meet
brilliant scholars of the opposite sex, marry, have brilliant children and generally live the
life of Riley. Talk about the haves and the have-nots: Yale's teaching assistants are among
the most pampered people in America (cited in Watt, 1997, p. 229).
And if graduate TAs and research assistants as the lowest stratum of academic
labor are considered to be privileged, how are college and university tenured
and tenure-track professors, as the highest stratum of academic labor, viewed
by the populace as a whole?
Nevertheless, at the end of the twentieth century and the beginning of the
twenty-first century, college and university faculty are confronted with a
variety of problems including attacks on the tenure system (Aronowitz, 1997),
the dramatic increase in the hiring of part-time faculty in the academy since
1970 (Pratt, 1997; Thompson, 1997), the growth and implementation of
distance learning (Rhoades, 1998), low salary increases and the corporatiza-
tion of the university with its consequent threat for the traditional shared
governance system between faculty, administrators and university governing
boards (Aronowitz, 2000). Because of these conditions, faculty have
conducted unionization campaigns in recent years at a number of colleges and
The Fears of Resource Standardization and the Creation of an Adversarial Workplace Climate 147
universities, including research institutions such as Southern Illinois

University (SIU) and the University of Minnesota. Thus, at the start of the
twenty-first century, collective bargaining appears to be a practical solution
for dealing with major threats to academic labor, to further encourage a demo-
cratic culture of higher learning in colleges and universities, to expand
workplace justice for the professoriate and to preserve the pursuit of truth in
the academy which is under attack from a variety of external sources who
believe that institutions of higher learning should be operated according to
the business model.
The purpose of this paper is to discuss the attempt to create a faculty union
at Illinois State University (ISU) by examining the dynamics of the Illinois
State University Faculty Association's (ISUFA) organizing campaign from the
first exploratory meeting held at the end of July 1998 until the representation
election in March 2000. In analyzing a faculty union organizing drive at Queen's
University (Kingston, Ontario), Rastin (2000) demonstrates the importance of
the communication of the union's message to the faculty (in this case, an
emphasis on collegiality in generating support for the union) in establishing an
organizing climate that led to a successful certification vote. I use the same
concept of examining the role of communication in ISUFA's organizing drive
from a slightly different point of view; I investigate the effect that the
communication of an anti-union message delivered by the administration and
an anti-union faculty group had on creating the organizing climate which led
to an unsuccessful certification election vote.
Specifically, I argue that a major factor in the ISUFA's loss in the certi-
fication election was due to the union opponents' effective communication
of a two-pronged message that unionization would impose significantly more
costs than benefits for the university's faculty. First, according to the union
opponents, having a faculty union would lead to the standardization of college
and departmental resources across the university. Second, the union oppo-
nents argued that collective bargaining representation would disrupt the
harmonious and collegial environment between the faculty and administra-
tion at the university, thus creating an adversarial workplace climate in which
decision making would occur. Conversely, the union was ineffective in
combating these two arguments by convincing faculty members of the
overriding benefits of unionization. It failed to build a sufficient "commu-
nity of interest" (or solidarity) among the faculty by focusing on a single
issue or a set of related issues that extended to university-wide issues as a
whole (salaries, faculty governance, etc.), or to external issues that confronted
the faculty as a profession (increase in non-tenure track positions, attacks on
the tenure system, etc.).
The primary anti-union message communicated by administrators and a

faculty anti-union group, the Faculty for Shared Governance (FSG), played upon
the fears of faculty members that unionization would potentially lead to the
equalization of resources across the university which would have negative
consequences for a number of individual departments. Therefore, the message
that the administration and the FSG sent to the faculty was the following: you
will be better off as individuals and departments in continuing to compete for
university and college resources than to take a chance on collective bargaining
which will lead to a uniformity of outcomes and conditions for all faculty
members. Thus, for many faculty members at ISU, unionization may have been
perceived as a mechanism which could hurt them on an individual (or depart-
mental) level rather than as a vehicle for improving the situation for faculty as
members of the university community or as members of the professoriate as a
whole.
The success of the "standardization of resources" argument in leading to
the defeat of the ISUFA in the certification election can best be explained by
the median voter model of union behavior (White, 1982; Demsetz, 1993;
Babcock & Engberg, 1999). If the outcome of a union certification election
is determined by a simple majority vote of the bargaining unit and the voters'
preferences are well ordered (that is, single-peaked) and a function of one
variable, the median voter's preference is the result that will defeat all other
outcomes in a series of pairwise elections (Kaufman & Martinez-Vazquez,
1988).
Based on the array of faculty salaries and departmental allocation of
resources, the median member most likely falls within the College of Arts &
Science (CAS).With respect to applying the major principle of the median voter
model, the fear of the union opponents was that faculty members in CAS would
align with the low salary/low resource departments of the College of Fine Arts
(CFA), forming a majority coalition in favor of union representation and thus
redistributing salaries and resources from the high salary/high resource College
of Business (COB) and College of Education (COE) departments to the CAS.
This concern over salary and resource leveling was probably a major reason
for the ISUFA's loss in the certification election and will be analyzed in greater
detail in the conclusion of this article.
THE W H E E L E R AND M c C L E N D O N T H E O R Y OF
EMPLOYEE SUPPORT FOR UNIONIZATION
A useful guide for understanding the unfolding and the outcome of ISUFA's
organizing campaign is the integrative model of union joining developed by
Wheeler and McClendon (1991). According to this model, three "paths" exist
to union organization. One path, the rational calculation path, involves each
employee deciding on whether to vote in favor of union representation based
on the individual's subjective judgment of the benefits and costs that will occur
if the union wins the representation election. If the employee believes that the
anticipated benefits will exceed the anticipated costs of having a union, then a
vote will be cast for the union. However, if the employee believes that the costs
will exceed the benefits, then the employee will vote against union representa-
tion (Holley, Jennings & Wolters, 2001).
A second path, the emotional path, in Wheeler and McClendon's (1991)
model posits that an individual will take action against an employer and move
towards supporting unionization if an employee experiences either a specific
threat or a particular frustration imposed by the employer with respect to the
individual' s current conditions of employment (such as the employer decreasing
pay or refusing to give employees pay raises). However, this support for union-
ization may be modified by either facilitating conditions or inhibiting conditions.
Facilitating conditions include solidarity, instrumentality, and saliency, that is,
the occurrence of certain events or leaders making unionization appear attrac-
tive. Inhibiting conditions include fear of employer sanctions against the
employee for supporting unions and norms opposed to supporting unionization.
For example, in a study of employer campaign tactics, Lawler (1990) reports
that 25.9% of employers commit unfair labor practices by discriminating against
union supporters during the unionization drive. The third (and final) path, the
political/ideological beliefs path, posits that an employee will support union-
ization based on the individual's ideological commitment to unionism.
The role of the rational calculation path in the union organizing drive will
be discussed in the section below. The roles of the emotional path and the polit-
ical/ideological beliefs path will be discussed at relevant points later in the
paper.
PERCEIVED BENEFITS AND PERCEIVED COSTS OF

FACULTY UNIONIZATION AT ISU
With respect to utilizing the rational calculation path throughout the union
organizing drive, the ISUFA presented the following benefits that could
be potentially obtained if faculty members chose unionization although, as
mentioned earlier, it did not focus the campaign around a single issue
(or a set of related issues) in order to attract support: faculty salaries will
increase; the faculty would gain more representation in the decision-making
process; the faculty would have an independent organization providing
representation; the faculty would have an independent forum for discussing

professional issues; more tenured/tenure-track faculty would be hired and the
faculty would obtain more support for research. The FSG, on the other hand,
outlined the following potential costs that would be imposed on the faculty
if unionization was achieved, while basing its campaign on the two specific
issues mentioned above: another level of bureaucracy would be added to the
decision-making process; all faculty members would be required to pay union
dues; there would be increased polarization between faculty and administra-
tors; both faculty members and individual departments would lose autonomy
in negotiation over university resources; faculty members will go on strike;
there will be fewer resources for the library, computers and research; union
decisions will be made in an undemocratic manner and there will be more
outside intervention from union staff, mediators and lawyers.
In utilizing the rational calculation path, the union's strategy was to attempt
to cobble together a coalition of faculty members by appealing to each faculty
member on one (or more) of a number of benefits outlined during the orga-
nizing campaign. The union felt that this strategy would lead to a successful
certification election outcome because it believed that a majority of faculty
members would identify strongly with at least one of the union's positions on
the issues that the union presented in its platform and during the organizing
drive. On the other hand, the FSG's strategy was to try to get faculty members
to vote against unionization by focusing its anti-union message on two
potential costs of faculty collective bargaining, as mentioned above, the
standardization of departmental/college resources across the university and the
creation of an adversarial workplace climate.
As will be elaborated on throughout this article, a majority of faculty members
did not believe that having the ISUFA represent them for collective bargaining
purposes would be in their self-interest. Rather, they felt that the FSG's vision
of what would happen with respect to resource allocation at the university would
be more likely to occur. Specifically, these faculty members feared that an
intemal redistribution of resources from high salary/high resource colleges and
departments to low salary/low resource colleges and departments within the
university was most likely to occur.
BRIEF HISTORY OF F A C U L T Y COLLECTIVE

B A R G A I N I N G IN THE UNITED STATES
Faculty collective bargaining in the United States has been in existence for a
little over three decades, dating back to the late 1960s (Cameron, 1982). In the
first decade of faculty unionization in institutions of higher education, from
1969-1979, the number of unionized faculties increased approximately tenfold

from 24 to 227 (Garabino, 1980). However, by 1980, faculty unionism in the
private sector experienced a major setback with the Supreme Court's Yeshiva
decision. In this decision, the Court restricted the right of faculty members at
private colleges and universities to unionize when it ruled that Yeshiva
University faculty were "managerial employees" and thus did not have statu-
tory rights to negotiate under the National Labor Relations Act (Douglas, 1990).
While the common perception of academic collective bargaining is that the
process is still mostly limited to public sector primary and secondary school
teachers, unionization among college and university faculty has reached major
proportions. In 1994, there were 242,221 faculty located on 1,057 campuses
that were covered by collective bargaining agreements (Rhoades, 1998) with
98% of the faculty covered by labor contracts being represented by three labor
organizations, the American Association of University Professors (AAUP), the
National Education Association (NEA) and the American Federation of Teachers
(AFT) (Annunziato, 1994). Furthermore, faculty unionization is fairly concen-
trated with respect to geography. Even though collective bargaining units for
faculty are present in 31 states, 50% of all faculty union members are found
in California and New York and 83% of faculty covered by collective bargaining
agreements are found in 10 states which include California, New York,
Pennsylvania, Connecticut, New Jersey, Michigan, Florida, Massachusetts,
Washington and Illinois (Annunziato, 1994).
Although faculty at research universities are largely non-union, with major
exceptions being the State University of New York system and Rutgers
University, approximately 44% of full-time faculty (if part-time faculty are
included with full-time faculty, the figure falls to 26%) possess union
representation on 29% of all campuses throughout the United States. If one
only considers the unionization of full-time faculty in public universities,
union density increases to 63% of full-time faculty members and 60% of insti-
tutions. Eliminating the public research universities from this tally, unions
represent 89% of all faculty members. Compared with the 1994 union
densities for the workforce as a whole (16%) and the private sector work-
force (12%), unionization rates among college and university faculty were
considerably higher (Rhoades, 1998).
In the early 1970s, the primary issues of concern leading to faculty union-
ization centered on economic issues, specifically dissatisfaction with current
salary levels and fringe benefits (Williams & Zirkel, 1989). However, within a
few years, faculty discontent spread to include concern over nonmonetary issues
such as" (1) fewer new positions restricting jobs for new entrants and mobility
for senior faculty members; (2) more difficulty in obtaining tenure;
(3) worsening working conditions because of increased teaching loads, outside

income limitations, a decrease in the availability of sabbaticals, and the encour-
agement of early retirement; and (4) threatened layoffs for research staff,
administrative professionals and contingent (part-time and temporary) faculty
(Garbarino, 1973). Because of these concerns, contractual clauses dealing with
personnel issues (specifically, appointment, dismissal, tenure, seniority, staff
reduction, and promotion) and governance issues, (that is, the faculty member's
perceived role in institutional decision making) were included in faculty collec-
tive bargaining agreements throughout the mid-1970s and 1980s (Williams &
Zirkel, 1989).
With the onset of the 1990s, college and university faculty have been
confronted with budget crises in higher education that have resulted in down-
sizing of institutions of higher education, reduction and/or elimination of
academic programs leading to, in certain situations, faculty layoffs (Graf et al.,
1994) as well as renewed attacks on faculty governance, tenure and academic
freedom (Rosenthal et al., 1994; Pratt, 1994). In addition, in the 1990s, faculty
are working longer hours - an average of 54 hours per week - compared with
an average workload of 45 hours per week in 1977 (Rosenthal et al., 1994).
DETERMINANTS OF SUPPORT FOR UNIONIZATION

AMONG C O L L E G E / U N I V E R S I T Y FACULTY
A multitude of studies exist outlining the determinants of faculty support for

unionization at colleges and universities. Economic factors are clearly the
primary reasons for the unionization of college and university faculty in the
United States (Graf et al., 1994; Hemmasi & Graf, 1993; Williams & Zirkel,
1989; Ponak et al., 1987; Jones, 1986). Studies have consistently found the
existence of an inverse relationship between salary levels and faculty support
for unionization (Ali & Karim, 1992; Karim & Ali, 1993; Ausieker, 1976; Begin
& Browne, 1973; Graf et al., 1994; Hemmasi & Graf, 1993; Ladd & Lipset,
1973).
College and university faculty members who feel that their salaries are
inadequate are more likely to possess positive attitudes toward unionization
(Barbezt, 1989; Gress, 1976; Williams & Zirkel, 1989). A related finding is
that faculty members who find their salaries to be deficient, even if their salary
levels were equivalent to academicians at similar schools, were more likely to
become members of the union (Alien & Keaveny, 1981; Bigoness, 1978;
Hemmasi & Graf, 1993; Ponak et al., 1992). In summary, the major determi-
nant of unionization in institutions of higher education is faculty dissatisfaction
with their compensation packages (Kochan, 1980).
With respect to university governance issues, several studies have found that
degree of participation in self-governance is a major factor contributing to
faculty members feeling positively about unionization (Ali & Karim, 1992;
Karim & Ali, 1993). Faculty members who felt that they had little input in
decision making and wanted to have a higher level of participation were more
likely to support unionization (Graf et al., 1994; Hemmasi & Graf, 1993).
Based on demographic characteristics, there is differential support for union-
ization among sub-groups of faculty members. Bigoness (1978) found the
existence of an inverse relationship between positive attitudes towards union-
ization and age, academic rank, and salary. With respect to field of study, social
scientists demonstrated the most support for unionization, followed by profes-
sors working in the humanities and the natural sciences. Faculty members
teaching in business, engineering and related applied fields were least likely to
be in favor of unionization (Feuille & Blandin, 1974; Ladd & Lipset, 1973).
Because of tenure and job security concerns, younger, untenured professors
were found to be more favorably predisposed towards unionization (Ali &
Karim, 1992; Karim & Ali, 1993).
With respect to working conditions at colleges and universities, job satis-
faction has been found to be consistently related to support for faculty
unionization (Zalesny, 1985). Bigoness (1978) discovered that job dissatisfac-
tion (concerning work, pay, supervision, and promotional opportunities) was
significantly related to support for faculty unionization. Other studies have found
an inverse relationship between specific and overall job satisfaction at work and
attitudes toward faculty unionization and pro-union voting behavior (Hemmasi
& Graf, 1993; Heneman & Sandver, 1983).
Although compensation issues are the major determinant of faculty union-
ization attempts, other single issues can ignite such organizing drives. For
example, the organizing campaigns at the University of Minnesota and SIU
were "hot shop" campaigns because of perceived crises by faculty members -
a proposed draconian revision of the tenure code in the case of Minnesota
(McClure, 1999) and extreme faculty dissatisfaction with university governance
at SIU (Magney, 1999). However, in contrast at ISU, a "cold shop" campaign
ensued because the union did not develop and utilize the emotional path as
outlined in Wheeler and McClendon's (1991) model during this drive. In other
words, there were no specific single issues that led to a majority of faculty
members feeling threatened and/or frustrated on an individual basis and that
subsequently galvanized the majority of faculty members to support unioniza-
tion. While there was considerable support for the ISUFA, affiliated with the
National Education Association (NEA)/Illinois Education Association (lEA), to
represent faculty at ISU for collective bargaining purposes, the ISUFA lost the
representation election on March 8, 2000 by a margin of 335 votes (54%) to

283 votes (46%).
NOTE ON THIS STUDY'S METHODOLOGY
Before I begin my analysis let me briefly comment on the methodology used

in this article. As a member of the ISUFA Steering Committee, and the lead
organizer in the College of Business (COB), this paper is based on participant
observation and is an ethnographic study of the organizing drive from its incep-
tion until the holding of the certification election. I engaged in discussion
concerning the unionization drive with the vast majority of faculty in the COB,
as well as individual faculty members in other departments, and I kept notes
throughout most of the organizing drive. In addition, I was involved in all major
aspects of the organizing drive and have a (virtually) complete collection of
the written material distributed by the ISUFA, the administration, and the FSG,
the faculty anti-union group that emerged in February 2000.
BRIEF HISTORICAL BACKGROUND TO THE

ISUFA'S ORGANIZING DRIVE
The ISUFA organizing drive came within a few years after votes of "no confi-
dence" of ISU president Tom Wallace in the summer of 1995 and ISU provost
John Urice in the fall of 1997. The vote against Wallace was prompted by his
uncompromising attitude that it was acceptable that he received large (and
secret) salary supplements from the ISU foundation; another factor which hurt
Wallace among faculty members was his perceived arrogant attitude that he
should be allowed to run the university as he saw fit. As in the case with
Wallace, Urice's vote of "no confidence" was precipitated by a belief among
the faculty that the provost operated in an arrogant and authoritarian manner
in carrying out his official duties.
After the brief presidency of David Strand (1995-1999), Victor Boschini
assumed the post of president of ISU in July 1999; former College of Fine Arts
(CFA) Dean Alvin Goldfarb was appointed provost in July 1998. As dean of
CFA, Goldfarb had developed a well-deserved reputation of being a strong
faculty advocate. Boschini, having served as vice president of student affairs
at ISU for a little over a year, did not have much of a track record but was
similarly viewed as being "faculty-friendly." Thus, the Boschini-Goldfarb team
was perceived quite favorably by a large segment of the faculty (including many
union supporters) in contrast to the Wallace-Urice administration of the early
to mid-1990s.
In addition to having problems with the university administration, the faculty

experienced problems with the newly reconstituted Board of Trustees. After the
dissolution of the Board of Regents, which governed Illinois State University,
Northern Illinois University and the University of Illinois-Springfield (formerly
Sangemon State University), each of these universities received its own Board
of Trustees that was responsible for governing the university. With the need to
rewrite the ISU constitution because of the implementation of the new governing
board, a conflict with the Board of Trustees developed in 1997. Faculty concern
over the future of shared governance peaked in 1998 "when the Board adopted
a new constitution in 1998, without the consent of the Academic Senate." The
standoff with the Academic Senate ended in January 1999, when the Board of
Trustees issued a Memorandum of Understanding that outlined the faculty's
"primary responsibility for academic issues, faculty affairs issues and educa-
tional issues related to student life." The return of faculty prerogatives contained
in the Memorandum led to a vote in which the Academic Senate overwhelm-
ingly approved the new constitution ("Has Shared Governance Broken Down
at ISU?" 8 February 2000).
THE FIRST STAGE OF THE CAMPAIGN,

JULY 1998-JANUARY 1999
The first meeting to explore the possibility of unionization at ISU occurred at

the end of July 1998 at the university's student center. Hazel Loucks, IEA's
Higher Education Director at the time, led the informational meeting that was
attended by six faculty members. A second exploratory (and informational)
meeting was held early in September 1998 that was attended by 25 faculty
members; several other informational meetings, led by the lEA, were held
throughout September that were sparsely attended. However, what emerged
from these meetings was the realization that there was a small (but committed)
core of faculty members willing to undertake the rigors of a faculty organizing
drive.
Towards the end of September, a number of the core activists decided that it
would be a good idea, out of respect, to approach Hank Campbell, the president
of the AFT local on campus, to determine if the AFT had any interest in con-
ducting a unionization drive among ISU faculty. The APT and the AAUP had
unsuccessfully competed against each other in an attempt to unionize the ISU
faculty in the late 1980s although a small inactive group of AFT members
remained on campus. A meeting was held with Campbell at the end of September
1998, attended by a dozen union activists, to discuss the AFT local's position on
staging a union organizing drive. Campbell's position was that he did not think
that unionization could succeed at ISU at this time and was even skeptical that
authorization cards could be collected from a sufficient number of faculty
members in order to qualify for the holding of a certification election.
With this information, the small group of approximately 15 union activists
decided to form a Steering/Organizing Committee and to affiliate with the
NEA/IEA. Although a major issue that provided the impetus for beginning the
organizing drive at this time was the conflict occurring with the Board of
Trustees, this core group of activists was composed of a majority of faculty
members who were ideologically committed to unionism although a few of
these activists supported unionism because of many years of dissatisfaction with
a succession of university administration regimes. Therefore, this core group
differed from the vast majority of ISU faculty members in taking the polit-
ical/ideological path (Wheeler & McClendon, 1991) in its support for collective
bargaining. These activists did not support unionization simply on the basis of
developing a balance sheet on the perceived benefits and costs of faculty union-
ization. Unfortunately, as will be discussed in more detail in the article's
conclusion, this political/ideological path was, on balance, a negative influence
in the unionization campaign because relatively few ISU faculty members,
mostly concentrated in the history, sociology and political science departments,
were ideologically committed to unionism.
From October 1998 through the beginning of December 1998, the committee
met once or twice a week in the evenings to hammer out a platform for the
ISUFA. Each meeting lasted for approximately two hours and potential planks
were discussed among the participants in a vigorous yet non-confrontational
manner.
In the final platform, there were 10 planks that the Steering Committee felt
were issues of concern to the ISU faculty as a whole and around which to
unite the faculty in constructing a democratic academic unionism, hopefully
culminating in a successful certification election vote. Included in the platform
were the following planks: a commitment to "meaningful shared governance"
at ISU in which the faculty would "exercise primary responsibility for the acad-
emic program and faculty status decisions" as well as "an appropriate role in
all university decisions"; support for the Academic Senate's "continuing
adherence to the principles of shared governance," as contained in AAUP
principles; "full participation" in all university decisions concerning the budget
and resource allocation; the preservation of disciplinary diversity; the promo-
tion of cooperation among faculty throughout the university; the implementation
of policies providing child care, maternity leave and "flexible tenure clocks" in
certain situations; "the introduction of new technologies such as distance
learning" when such techniques do not result in faculty exploitation and lead
to enhancement of the educational experience for students; a non-negotiable

maintenance of tenure; the guarantee of academic freedom as outlined by AAUP
principles, and the democratic participation of all union members in all
collective bargaining decisions. The platform was distributed to the faculty
members through their university mailboxes during finals week in December
1998 and, again, when classes resumed in January 1999 ("Illinois State
University Faculty Association Platform" 21 January 1999).
THE ASSESSMENT & THE AUTHORIZATION CARD

DRIVE, JANUARY 1999-DECEMBER 1999
At the end of January 1999, a public meeting was held to officially launch the
ISUFA. Besides the attendance of approximately 70 faculty members, the
meeting was well attended by the media, which included the local television
stations, radio stations and newspapers. Ron Strickland, the ISUFA president
and a professor in the English department, made some brief remarks inaugu-
rating the organization before inviting the attendees to stay for the buffet lunch.
Beginning in the 1999 Spring Semester and continuing into the summer of
1999, the ISUFA engaged in an assessment of all tenured and tenure-track
faculty members employed by ISU. Departmental representatives conducted this
procedure and the purpose of this assessment was to determine the faculty's
initial reaction towards faculty unionization and collective bargaining. For
departments that had no representatives, union activists telephoned faculty
members from the IEA office in both April 1999 and June 1999. Approximately
580 faculty (of 680 faculty) were assessed through this procedure and the results
indicated that a little more than one-third were in favor of faculty unionization,
one-third were undecided and a little less than one-third were opposed to faculty
unionization. Many members of the Steering Committee were encouraged by
these results, especially a number of the veterans from the earlier unionization
attempt in the late 1980s, because of the support present in the COB. Although
in the previous campaign, there was virtually no support in this college, the
assessment indicated that approximately one-third of COB faculty supported
faculty collective bargaining.
After the tabulation of these results, it was decided that the authorization
card drive would be launched at the start of the 1999 Fall Semester. A deci-
sion was made at this time to only solicit cards from faculty members who
supported the union. The reasoning behind this decision was that the union
wanted to have a good idea with respect to the union's base of support before
filing for a certification election. Because of rules established by the Illinois
Education Labor Relations Board (IELRB), the union had a six-month period
to collect authorization cards. Since the first cards were collected at the end
of August 1999, the ISUFA had until the end of February 2000 to file for a
certification election with the labor board.
During the first few weeks of the authorization card drive, cards were
collected rather quickly and easily. In the four-week period from September
14, 1999 to October 12, 1999, over 150 authorization cards were signed and
by the end of this period, the ISUFA had exceeded the 30% minimum required
to file for an election. However, while it had taken only six weeks to collect
cards from 30% of the bargaining unit, it took another six weeks to collect an
additional 10% of cards from bargaining unit members. Thus, by Thanksgiving,
the ISUFA had collected cards from a little more than 40% of the eligible
bargaining unit. The breakdown by college is presented in Table 1.
It is not surprising that there were differential success rates between the
various colleges during the authorization card drive. The strongest support came
from the Colleges of Arts & Sciences (CAS), Fine Arts (CFA) and the Library.
The weakest support came from the Colleges of Applied Science & Technology
(CAST) and Nursing. It should be noted, however, that the total lack of support
in the College of Nursing may have been due to the fact that it became part
of ISU in July 1999 and had only seven tenured/tenure-track faculty members
at the time of the card drive.
Nevertheless within the CAS and CAST, there were major differences with
respect to support during the card drive. For example, in six departments within
CAS, which included History, Sociology & Anthropology, Foreign Languages,
Political Science, Philosophy and Psychology, 75% of the faculty members had
signed cards. However, within the physical sciences departments, which
included Biology, Chemistry and Physics, only 14.8% of these faculty members
signed cards. In the CAST, 66.7% of the faculty members in the Criminal
Justices Department signed cards while only 7.4% of the faculty members in
the departments of Agriculture and Health Sciences signed cards. These results
are generally consistent with past studies indicating that social science and
Table 1. Faculty Signature of Authorization Cards.
College Affiliation Percentage Signed

Applied Science & Technology 24.3
Arts & Sciences 51.4
Business 41.3
Education 40.0
Fine Arts 48.8
Library 48.3
Nursing 0.0
humanities professors are more likely to support faculty unionization than

professors in business or related applied fields (Feuille & Blandin, 1974; Ladd
& Lipset, 1973).
While the union was collecting cards during the 1999 Fall Semester, the
ISUFA attempted to obtain support from the ISU chapter of the AAUP. The
ISUFA believed that the AAUP's endorsement of the organizing drive might
help sway undecided faculty members to support unionization. Although the
organization's membership was somewhat passive and fairly inactive throughout
the 1990s, the chapter's activist leadership did a good job of keeping the faculty
informed of developments at the university that could negatively impact the
faculty. Even though approximately two-thirds of the AAUP membership had
signed authorization cards by the end of November 1999, there was still a
significant percentage of AAUP members who were opposed to both faculty
unionization and to the organization's endorsement of the organizing drive.
While a number of undecided AAUP members had grown more sympathetic
to the organizing drive from fall 1998 to fall 1999, certain members were
concemed with the views of the national AAUP office concerning the orga-
nizing drive. In order to resolve this issue, Pat Shaw of the national office
attended a meeting of the local chapter in November 1999. Shaw stated that
the national AAUP's position was that whenever an organizing drive occurred
among faculty on a campus, the AAUP would not interfere with organizing
efforts but would support the unionization drive.
A meeting of the ISUFA Steering Committee was held on the evening of
November 22, 1999 to discuss when to file with the IELRB for a certification
election. The choice of the dates boiled down to December 6, 1999 or early
January 2000 although it was determined that a final decision would not be
made until the meeting with the AAUP in early December. However, most of
the Steering Committee members were leaning towards early January so that
the election could be held by early March 2000. Although no faculty anti-union
group had emerged during the collection of authorization cards, as was the case
during the last organizing drive at ISU, the feeling was that filing in early
January 2000 would provide the ISUFA with a better opportunity to respond
to a potential anti-union campaign conducted by the administration over the
four-week break between the two semesters.
At the meeting of the ISUFA and the AAUP held in the late afternoon on
December l, 1999, the AAUP representatives present favored the endorsement
of the ISUFA organizing drive and they promised the ISUFA representatives
that they would poll the AAUP membership in January concerning an endorse-
ment. In addition, the AAUP representatives requested that the ISUFA endorse
the AAUP's principles on academic freedom, tenure and shared governance.
On the evening of December 1, 1999, the ISUFA Steering Committee met

to discuss the ISUFA/AAUP meeting held earlier in the day and to finalize the
date concerning when the ISUFA should file for a certification election with
the IELRB. After a short discussion of the meeting, the ISUFA Steering
Committee agreed to endorse the AAUP's principles. An animated discussion
occurred of when to file for the certification election. Peter Miller, the IEA staff
organizer, argued in favor of delaying this decision and called for the Steering
Committee to spend the next three months continuing to collect authorization
cards, or until the ISUFA had signed authorization cards from at least 50% of
the potential bargaining unit. Miller felt that filing with only a little over 40%
of cards was too risky in terms of the union winning the certification election.
Although the Steering Committee members were not philosophically opposed
to Miller's position, from a practical standpoint, they felt that this position was
untenable. Since the number of signed authorization cards had slowed to a
trickle within the past two weeks, committee members believed that even if an
intensive effort was made over the next three months, the 50% mark could not
be achieved. In addition, only a handful of Steering Committee members were
still actively soliciting cards and they concurred with the other committee
members concerning the possibility of collecting many more cards.
While there was not a clear consensus among the committee members of
whether to file in early to mid-December or in early January, the Steering
Committee members decided to file for an election around December 10, 1999
with the goal of having the election by the middle of February 2000. Six days
later, on December 7, the ISUFA Steering Committee met to discuss the estab-
lishment of committees for work on the union organizing drive once the ISUFA
filed for an election with the IELRB.
THE ELECTION CAMPAIGN,

JANUARY 2000-MARCH 2000
The campaign for ISUFA's representation election was officially kicked off on
January 20, 2000 with a buffet lunch held for all tenure-line faculty members.
One day later, on January 21, a four-page memorandum was prepared by Sharon
Stanford, Associate Vice President for Academic Administrative Services, which
was composed of "frequently asked questions" concerning collective bargaining
with answers and was distributed to all tenure-line faculty members. In
both tone and content, the memorandum appeared to be neutral, objective and
informative. However, the second question - " W h a t is negotiable under
collective bargaining?" - answered, "In short, almost everything" and provided
a list of "negotiated items" included as provisions of the collective bargaining
agreement negotiated between The Board of Trustees of SIU and the SIUCFA
IEA/NEA. In addition, the answer concluded, "The agreement also included a
section on Board of Trustee Rights including the following":
Section 1: As long as such actions and decisions are consistent with the other express
Articles of this Agreement, it is understood and agreed that the Board, on behalf of the
University, retains and reserves all of its powers and authorityto direct, manage,and control
all operations and activites (sic) of the Universityto the full extent of the law... (Stanford,
2000).
The inclusion of this section of the contract in the memorandum was clear; even
with a negotiated collective bargaining agreement, the Board of Trustees would
still retain the right to run the university as it wishes to, thus attempting to under-
cut one of the union's major issues that it was trying to rally faculty members
around - the threat to shared governance by the current Board of Trustees.
At the end of January and the beginning of February, an e-mail ballot
concerning whether the ISU AAUP chapter would officially endorse the ISUFA
organizing drive was held. During this period, a spirited debate occurred over
the proposed endorsement through the exchange of e-mails. Requiring a
two-thirds majority to obtain endorsement of the union organizing drive, the
proposal was decisively voted down by a margin of 12 votes for endorsement
to 14 votes against endorsement.
In an attempt to continue to educate the faculty concerning collective bar-
gaining, the ISUFA organized a weekly forum series, beginning on January 27
and running through February 24, which included topics such as what unions
can do for faculty, an open discussion of faculty collective bargaining at ISU,
collective bargaining in higher education, collective bargaining among
physicians, and the use of interest-based (also known as "win-win" or mutual
gains) bargaining by faculty unions and university administrations. The first
forum, in which Professor Stanley Aronowitz of the City University of New
York spoke, had the best attendance with approximately 70 people in the
audience. The open discussion of faculty collective bargaining at ISU, the
second forum, produced a lively discussion among union supporters and union
opponents and had 30 faculty members in the audience. The remaining forums
only had approximately l0 people in the audience.
Through early February, the major problem confronting the ISUFA was
trying to educate and mobilize a faculty that had little knowledge of the
mechanics of collective bargaining. Up to this point in time, there had been
little overt opposition to the organizing drive by the university administration
and no organized opposition from any anti-union faculty groups. The only
visible opposition to the ISUFA's campaign came from a small number of
individual faculty members, from different departments, who appeared to
operate in a disorganized and independent manner. However, this situation had

changed by the middle of February with the organization of an anti-union
faculty group named the Faculty for Shared Governance (FSG).
Although not explicitly identifying itself or its organizational structure until
relatively late in the campaign, the group described itself as "a loosely-knit
alliance of faculty members from all across campus" with "no formal leader-
ship structure" or "official "steering committee" ," but with "an editorial board
in crafting and editing our memoranda." Three faculty members from different
departments in the CAS "served as the public spokespersons for the group"
which included Neil Skaggs, a conservative economics professor, George
Gordon, a self-described "pro-labor liberal" out of the Political Science
Department and Charles Harris, a former chair of the English Department and
an AAUP member. An additional eight faculty members from a variety of
departments served as editorial board members. Finally, the FSG argued that
none of its members was an "administrator" but that all of its members were
"full-time, tenure-track faculty members."
As an organization, the FSG engaged in two major activities during its short
life. It distributed, what it referred to as "informational memoranda" to faculty
though the university's e-mail system as well as organizing "debates" between
ISUFA representatives and FSG representatives in half of the university's
colleges. However, FSG members did not appear to engage in much, if any,
one-on-one contact with ISU faculty members.
The ISUFA/FSG Debates on Collective Bargaining
Upon the organization of the FSG, there were a number of colleges throughout
the university that wanted to hear the positions of these two organizations with
respect to unionization. In a period of a little over a week from the middle to
late February, debates were held in the following three of the six colleges in
the university: CAST, Education, and CAS. The best attended of these debates
(based on the percentage of faculty in the college) was the one in the CAST
which also happened to be the first of the college debates held.
The structure of these debates followed a standardized format. After the
ISUFA and FSG panelists presented their positions, faculty members in the
audience were provided with the opportunity to ask the panelists questions.
From the audience's questions, it is clear that the issue of a perceived loss of
departmental flexibility and how the union would deal with departmental issues
was on the minds of those who attended these debates. Although a number of
the questions dealt with clarifications of what had been presented by the two
panels, a majority of the questions directed to the union team by audience
members focused on technical departmental issues. While the FSG continually

emphasized the negative effects to individual departments of the standardiza-
tion that would occur under collective bargaining, the ISUFA argued that it had
no interest in either micro-managing or regulating issues of concern to indi-
vidual departments. The union team's response was that the union supported
the handling of these issues at the departmental (or program) level and that
such issues would not be covered in the contract. In retrospect, the union should
have been better prepared for handling such queries as opposed to giving this
"blanket" response whenever such questions were raised.
The FSG and Other Union Opponents' E-Mail Informational

Memoranda Campaign
In the FSG memos, the group used the "standardization of resources" argument
as a primary strategy to play on the fears of the ISU faculty but also empha-
sized, at times, that unionization would foster an adversarial relationship among
faculty members and administrative staff, and thus lead to the damage, of an
otherwise harmonious academic community. For example, concerning the issue
of shared governance, the FSG argued that collective bargaining would inter-
fere with and undermine the cooperation, needed between faculty, the university
administration and the Board of Trustees to have a meaningful shared gover-
nance system. In memo number 7, written by FSG leader Charles Harris, Harris
pointed out why collective bargaining is incompatible with the practice of shared
governance:
But collective bargaining should not be confused with shared governance. If the structure
of shared governance is relatively horizontal, the structure of collective bargaining is binary.
Whereas the mode of shared governance is deliberative and cooperative, the mode of collec-
tive bargaining is legalistic and adversarial . . . Collective bargaining replaces the ethos of
collegial decision making with a legal contract. The very concept of bargaining implies
negotiation between separate parties. While these negotiations needn't be rancorous, the
adversarial structure of collective bargaining threatens the basic meaning of collegiality:
"the sharing of authority among colleagues" (Harris, 2000).
In addition, the group stated that faculty unionization could not protect shared
governance. In terms of the contract negotiated between the SIUCFA and SIU,
the FSG claimed that "shared governance at SIUC is above the contract" so
that "the continuance of shared governance at SIUC rests on exactly the same
willingness to cooperate upon which it depends here at ISU" ("Has Shared
Governance Broken Down at ISU?" 8 February 2000).
Although it appears that the FSG took the offensive in the union organizing
campaign at this time, ISUFA activists and supporters continued to meet on a
one-on-one basis with faculty members, to promote the union's platform and
to distribute union leaflets. In addition, the ISUFA responded to the FSG's
informational memoranda by distributing the union's response over the univer-
sity e-mail system, shortly after the FSG's positions appeared. For example,
addressing Harris' argument concerning the relationship between collective
bargaining and shared governance, ISUFA President Ron Strickland argued that
collective bargaining would not replace the structure of shared governance and
that these two models of university governance were not incompatible.
Strickland stated:
Collective bargaining agreements at institutions with faculty unions affiliated with the
National Education Association provide a "safety net" for shared governance, and a legal
foundation upon which faculty can enter into shared governanceon equal footing with the
administration and the Board of Trustees.
We can strengthen and protect our traditional role in shared governance with the ISUFA
(Strickland, 2000a).
Besides emphasizing how faculty unionization would undermine and threaten

shared governance, the FSG, through its communications, argued that
collective bargaining would actually lead to worse conditions for the majority,
if not all, of the faculty. Since the group claimed that "(t)he process of collec-
tive bargaining is one of give-and-take," and since "few faculty members are
willing to participate," the bargaining team will be composed of faculty
members "with axes to grind" who will be required to make "tradeoffs,"
implying that the final negotiated contract will only serve the interest of a
small percentage of the faculty, instead of representing the best interests of
all faculty members. Using the SIU contract as a model, the FSG claimed
that "the contract contains many articles that are likely to raise concerns
among ISU faculty members" which regulated "the minutiae of faculty life"
such as "the number of office hours, the content of syllabi, and procedures
for canceling an occasional class" ("Promise Us Everything, Give Us . . .
What?" 14 February 2000).
Once more, implementing the "standardization of resources" argument, the
FSG argued that collective bargaining would interfere with departmental
flexibility, that is, collective bargaining would not be sensitive to the varying
needs of individual university departments. The FSG argued that "the most
important academic issues" to faculty are departmental issues that would be
insufficiently handled under collective bargaining. According to the group, with
faculty unionization, standardization of conditions across departments would be
"inevitable," which "would actually diminish faculty voice on most of the issues
that concern us" ("Can You Hear My Voice?" 16 February 2000).
Since "(t)he major concerns of members of one department fail to resonate

with their colleagues in other departments," a department's faculty, as a minority
in the bargaining unit, would not be able to muster the votes needed to pursue
their issues through the collective bargaining process. However, because of the
negotiated contract, the F S G claimed that the department would no longer be
able to press its concerns "through "ordinary" administrative channels either."
This would be the case because the departmental chair would now serve as "a
manager in charge of ensuring that the collectively determined rules are
followed," rather than as an advocate pursuing the department's best interests
in front of both the college and university administrations ("Can You Hear My
Voice?" 16 February 2000).
Besides pointing out that the I S U F A would be unable to sufficiently repre-
sent the faculty's major concems at the departmental level, the F S G also claimed
that the ISUFA, would be ineffective in lobbying for the faculty's interests at
the state capitol in Springfield because of the competition for resources between
the different constituencies that the union was responsible for representing. The
F S G argued that since the l E A "represents 93,000 primary and secondary school
teachers" in Illinois, even if the I S U F A won the certification election, the
number of I E A members at the university level (SIU and ISU) would be
extremely small with respect to the number of l E A members who teach at
elementary, middle, and high schools. Based on the limitation of time and
resources and the competition between the l E A sufficiently representing the
interests of the primary/secondary schoolteachers and the interests of the univer-
sity faculty, the group claimed that the I E A ' s priority in lobbying at the state
level would be, out of necessity, the primary and secondary school teachers.
The F S G concluded:
Think about it logically. Whose interests are likely to be a priority with the IEA lobbyists
in Springfield? We think we know the answer to that one ("Can the Tail Really Wag the
Dog?" 21 February 2000).
In response to the F S G ' s concerns that the I S U F A would not be sufficiently

represented by the IEA in lobbying at the state capitol, the I S U F A argued that
all educators have similar, rather than opposing, interests. In an e-mail response
to "Can the Tail Really W a g the Dog?" union president Ron Strickland stated:
... (E)ducators hold many things in common, and working together we all achieve more.
The lEA has over 100,000 members, it is the largest political action committee in the state,
it contributes more to political campaigns in Illinois than any other organization. To extend
the FSG's metaphor of "dogs" and "tails," the IEA is a big dog. As university faculty, we
will benefit considerably by having access to IEA lobbying resources (Strickland 2000b).
While the I S U F A presented convincing evidence that ISU faculty members

were underpaid with respect to peer institutions and argued that higher salary
increases were needed, the F S G agreed with the union that unionization would
lead to higher faculty salaries. However, the F S G claimed that these salary
increases would come at a cost because of competition over resources.
Therefore, according to the FSG, the only reasonable way to fund higher salaries
would be through "internal reallocation within ISU" ("Would a Union Raise
Faculty Salaries?" 10 February 2000).
The faculty anti-union group argued that there were a number o f realistic
possibilities o f where salary increases could come from: "decreased adminis-
trative expenditures; reductions in funds for library or technological services,
departmental and college contractual and travel budgets, and other academic-
support areas; or a decreased number o f tenure-track faculty positions."
Because the university administration was in the process o f reorganizing in
order to make additional funds available for increased faculty salaries, the
F S G claimed that because o f this, "support services would almost certainly
suffer." Once more invoking the " p r o b l e m s " with the SIUC contract, the group
concluded:
Given steady enrollments, probable consequences include more temporary-line faculty

positions, larger average class sizes, and higher faculty teaching loads. Nothing in the SIUC
contract prohibits any of these effects from occurring ("Would a Union Raise Faculty
Salaries?" 10 February 2000).
The handful of e-mail m e m o s that were circulated in the COB by the admin-
istration and faculty members opposed to the unionization drive also utilized
the "standardization of resources" argument by pointing out that business
faculty would be worse off with collective bargaining because o f the c o l l e g e ' s
munificent funding level. Outlining his personal position in a m e m o a few
days before the certification election, F S G m e m b e r Ken Newgren o f the
Management and Quantitative Methods Department argued that faculty
members in the other ISU colleges were jealous of the resources of COB
faculty members and that under collective bargaining, the COB would
ultimately lose out because the faculty members in the other colleges would
not fight for its interests. He stated:
One thing I have learned from my time on the Academic Senate is that many faculty
members across campus are highly envious of the COB faculty in terms of salary and
teaching loads. Beginning salaries in some departments are under $40,000, and release time
often means 9 hours instead of a 12-hour load; I see no advantage for us. Since a union
negotiates all terms of employment, I see no incentive for it to improve the status of what
many of its members already refer to as the "rich college" (Newgren, 2000).
And in a memo distributed the day before the election by ten anti-union
COB faculty members, the m e m o ' s authors appealed to the major issues
raised by the FSG. They argued that an adversarial environment would be
created in the academy through the implementation of collective bargaining
and that unionization would decrease flexibility in the university. The authors
of this memo queried, "Why create an adversarial environment, especially
when the board and our top on-campus administrators have taken significant
recent steps to help assure faculty's strong role?" (Celuch et al., 2000).
In reference to decreased flexibility in the university, the authors claimed
that "(w)e also perceive a loss in valued freedom, as unionization would
(based on what the SIU-C contract shows) micro-manage much of our
existence." In one of the most bizarre claims appearing in the literature
circulated by anti-union forces, the authors provided an example of this
micro-management:
We fear that it could also reduce the protection of tenure by spelling out our responsibili-
ties in such detail that it would actually become easier for an antagonistic administration
(a future scenario that union backers seem to fear) to terminate a faculty member "for just
cause" (see Article XI, Section 2 of the SIU-C contract) (Celuch et al., 2000).
In a last minute appeal to the COB faculty (the e-mail message arrived at 4:45
p.m. on March 7, the day before the election), Dean Dixie Mills also referred
to the "standardization of resources" argument in opposing unionization and
keeping things status quo:
I believe that we are better served by a governance and resource allocation model that recog-
nizes the differences in staffing needs and other priorities across departments and colleges.
We need this to allow the College of Business to reach its full potential (Mills, 2000).
And in their only written public communication to the faculty in which they took
a position with respect to the upcoming certification election, President Victor
Boschini and Provost Alvin Goldfarb, sent a letter to all tenured/tenure-track
faculty which arrived by mail on March 6, two days before the election. In a
carefully worded letter that did not appear on university letterhead, the two lead-
ing university administrators counterposed the tradition of the shared governance
model with that of collective bargaining. Arguing that they were committed to
achieving competitive salaries for ISU faculty with peer institutions, Boschini and
Goldfarb pointed out that they had made significant progress within the last three
years and that they planned to continue to increase faculty salaries in the future. In
a similar vein to the FSG, Boschini and Goldfarb stated that collective bargaining
would lead to standardization across academic departments and a loss of
flexibility for individual academic departments to pursue their best interests:
We believe that collective bargaining arrangements tend to standardize and centralize

relationships. The flexibility and open communication that currently exists in individual
academic departments may be replaced by a "one-size-fits-all" contract. Each academic
department at Illinois State has unique characteristics and needs. Faculty, working with
department chairs, can champion these needs to college deans, administrators and governing
bodies. A contract that places everyone under the same umbrella might circumvent that
flexibility. We wonder, as many others have wondered, is this the right time for Illinois
State to change its faculty/administration governance relationship? (Boschini & Goldfarb,
2000).
ANALYZING THE CERTIFICATION ELECTION

RESULTS
As stated in the introduction to this paper, the I S U F A lost the certification

election held on March 8, 2000. Voter turnout was high, with 89% of the
eligible faculty casting a vote in the election. The I S U F A obtained 283 votes
(45.8%) and the "no union" option obtained 335 votes (54.2%). W h i l e the
outcome o f the election came as a surprise to many of the Steering C o m m i t t e e
m e m b e r s who b e l i e v e d that the union had w i d e s p r e a d support among the
faculty, I was not surprised by the results because, as a labor relations scholar,
I am aware o f the difficulties confronting unions in winning certification
elections.
I will admit that for a t w o - w e e k period in February, I shared m y Steering
C o m m i t t e e c o l l e a g u e s ' confidence and o p t i m i s m in the outcome o f the
election. However, a p p r o x i m a t e l y two weeks before the election, I had my
doubts that we w o u l d win the election. I felt that many of the Steering
C o m m i t t e e m e m b e r s were overconfident and were not going all out to bolster
and maintain contact with union supporters.
M y doubts were confirmed a couple of days before the election. Since we
had 334 faculty members (48.1% of the bargaining unit) indicate that they were
union supporters at the peak of our strength and 307 faculty members had
actually signed cards (Miller, 2000), we needed to have virtually all of these
faculty members vote in favor of the union in order to win the election because
of the expected high voter turnout. This meant that we could afford to have
only minimal slippage, an unlikely occurrence. In calling our union supporters
in a final "Get Out The Vote" on March 5-6, of the 174 faculty members
contacted, 159 were still supporters (91%), 10 had become neutral (6%) and
five (3%) were no longer in favor of the union. W h e n I learned of these numbers
on March 7, I knew that the union did not have a good chance to win the
election.
CONCLUSION: LESSONS ON ISUFA'S FAILED

ORGANIZING DRIVE
From the election results, it is apparent that the ISUFA was not able to success-
fully utilize the rational calculation, the emotional and the political/ideological
belief paths (Wheeler & McClendon, 1991) in its union organizing campaign.
The union was unable to build a sufficient community of interest for the ISU
faculty over the perceived benefits that unionization would bring concerning
university and national level issues as outlined by the ISUFA's platform. In
addition, the union was not able to exploit any single issue that resulted in a
majority of faculty members feeling threatened or frustrated by any of the
university administration's behaviors or actions.
Rather, because of the "standardization of resources" argument and the argu-
ment that an adversarial workplace climate would be created was continually
reinforced through communications by the administration and the FSG, it
appears that a majority of ISU faculty came to believe that collective bargaining
would not lead to an improvement in either their departmental or individual
situations but would lead to more adversarial relations between faculty members,
university administrators and the Board of Trustees. The oft-repeated threats of
collective bargaining bringing about the standardization of resources across
university departments while creating an adversarial workplace climate, unfor-
tunately resonated with too many faculty members who already were uncertain
or lacked sufficient knowledge concerning faculty unionization in the first place.
As stated in the introduction to this article, the success of the "standardiza-
tion of resources" argument in leading to the ISUFA's certification election loss
is best explained by the median voter model of union behavior. This concern
over salary and resource leveling from high salary/high resource departments
to low salary/low resource departments is supported by a post-election survey
conducted by Abacus Associates for the NEA in September 2000. The survey
indicated that 27% of the faculty earned less than $46,000 per year, 32% earned
between $46,000 and $59,999 per year and 41% earned over $60,000 per year.
Using salary level as a rough proxy for departmental membership, the study
found that union opponents were more likely to be faculty members with salaries
over $60,000 per year (or from high salary/high resource departments). The
survey also discovered that fine arts and liberal arts faculty members, that is
low salary/low resource departmental faculty members, were much more likely
to be union supporters. Finally, consistent with predictions of the median voter
model of union behavior, the study found that union supporters were more
likely to support pay equity across academic disciplines (or to put it in other
words, internal redistribution from high salary to low salary faculty) although
the study concluded that for the union to succeed in the future, it should not
focus its campaign around the issue of achieving pay equity across the
disciplines (Abacus Associates, 2000).
The median voter model also can be used to explain the union's certifica-
tion loss because of the faculty members' concerns that unionization would lead
to a standardization and loss of flexibility among the heterogeneous departments
and colleges at the university. At ISU (as well as at many other universities),
the working conditions of faculty members are actually rather heterogeneous
because of differential salaries, teaching loads, research expectations, graduate
student assistance, travel support, etc. among the various departments and
colleges within the university. Based on the election's outcome and the results
of a post-election survey, many ISU faculty members perceived that they benefit
from this heterogeneity (or flexibility) and believed that collective bargaining
will impose a standardization (or a forced homogeneity) with respect to
resources and working conditions at the university. The NEA post-election
survey indicated that 67% of union opponents believed that it was either
extremely likely or very likely that if the union won the election, faculty
members would lose autonomy in negotiating salary and/or course loads. With
respect to departments losing autonomy, 63% of union opponents believed that
such an outcome was either extremely likely or very likely to occur if the
ISUFA had won the certification vote (Abacus Associates, 2000).
Since the bargaining unit is composed of heterogeneous faculty members,
raising the threat of standardization of resources under collective bargaining
will appeal to the median voter and will result in the defeat of the union in the
certification election. Or to put it another way: an appeal to maintaining the
heterogeneity and flexibility among the faculty is a winning strategy if a hetero-
geneous faculty exists.
Thus, the FSG accomplished what the ISUFA failed to do; it successfully
focused its message around two major issues in order to generate and sustain
sufficient opposition to the union. On the other hand, the union was unable to
build commitment to the union centered on a major issue, or a couple of related
issues, that affected all faculty members. Rather, the union attempted to cobble
together a coalition of supporters by trying to appeal to as many faculty members
as possible in as broad a coalition as possible. The hope was that a majority
of faculty members would be attracted to the union through their support of at
least one plank in the platform. The ISUFA attempted to be, in essence, "every-
thing to everyone" which explains why the platform was maddeningly
amorphous and lacked sufficient detail in its ten planks.
This belief in the need for the maintenance of both individual faculty member
autonomy and departmental autonomy with respect to the negotiation of resources
keeps the faculty divided, further reinforcing the lack of trust and cooperation
among the faculty. And with a lack of trust and cooperation, why should faculty
members from one department, who are numerically a minority in the bargaining
unit, believe that the union would fight for their individual interests?
This argument is intertwined with, and reinforces, a broader ideological
orientation to their situation that is present in university faculty. This orienta-
tion, known as professionalism, according to Meisenhelder (1986, p. 382), "is
grounded in an individualistic consciousness." For the professoriate, profes-
sionalism requires that faculty respond to their work situation "through
competition within themselves, other faculty, and, particularly, with other
workers in order to secure the rewards of professional status." This
professionalism strengthens the professional culture and political beliefs of
faculty members, encouraging them to oppose unionization, which means that
the political/ideological beliefs path of the Wheeler and McClendon (1991)
model will generally be a negative influence in faculty unionization campaigns.
This was certainly the case with respect to the ISU organizing drive.
Unfortunately, the operation of professionalism in practice fails to address the
problems facing university faculty as a collective entity. Faculty members
continue to compete against each other for continually dwindling resources and
because they view themselves "as a self-governing community of independent
intellectuals," they do not see the need to organize as employees on the job.
Furthermore, refusing to view themselves as employees has other consequences,
most notably the reluctance of faculty to form alliances with other university
employees (non-tenure track faculty members, clerical/administrative workers,
maintenance workers, food service workers, etc.) or with other intellectual
workers in the teaching profession, such as primary and secondary school
teachers, who are also viewed as competitors for the state's educational resources.
In addition, during the campaign, the ISUFA did not sufficiently address the
other major concem raised by the FSG, specifically that the presence of a union
would create an adversarial climate between faculty members and the admin-
istration. As indicated by the election outcome and the NEA's post-election
survey, this issue also resonated with the faculty. According to the survey, 81%
of union opponents believed that if the ISUFA had been certified, it would lead
to an increase in polarization between faculty members and university admin-
istrators (Abacus Associates, 2000). Although the union argued that it would
cooperate and work with a faculty-friendly administration, the union did admit
that if the administration was antagonistic to the union, then it "would have to
play hardball" with the administration.
Research has demonstrated that a major concem of white-collar workers
during unionization drives is the fear that a unionized future will lead to a
conflict-ridden workplace (Cohen & Hurd, 1998). Although such workers are
less likely to vote for a union if they believe that unionization will contribute
to the occurrence of strikes (Maranto & Fiorito, 1987; Hurd & McElwain, 1988),
this fear extends beyond the proliferation of strikes to the idea that
workplace life on an everyday basis will be in "a state of perpetual conflict if
the organizing campaign succeeds" (Cohen & Hurd, 1998, p. 182). One
solution for dealing with this problem, according to Cohen and Hurd (1998,
pp. 194-195), is "to demonstrate the potential for the union to coexist as an
equal with management not only in negotiations and grievance handling but
also in solving a wide variety of workplace problems." These researchers state
that if management introduces and generates conflict, the union needs to respond
"without assuming responsibility for that conflict."
What lessons were learned from this organizing drive? I think the first, and
a primary, lesson is that many faculty members will not support a unionization
drive unless they feel that a union will be able to address the micro-issues, such
as the standardization/flexibility concerns, that affect them as faculty members
of individual departments. Unlike the FSG and the administration, which
successfully communicated their feelings of how collective bargaining could
negatively impact a wide range of micro-issues of concern to the faculty, the
ISUFA did a rather poor job of addressing these faculty members' concerns.
Addressing these micro-issues is of crucial importance for obtaining success in
any faculty organizing campaign because only a relatively small number of
faculty members have an ideological predisposition towards supporting
unionism or will be swayed by issues confronting the university or the profes-
sion as a whole. Although the union did a good job of addressing faculty
members' concerns with respect to the issues outlined in the platform, I think
that the ISUFA should have been better prepared to provide sufficient answers
to many of the faculty members' detailed and technical questions concerning
standardization and department level issues. A typical response that union
supporters provided to these types of questions was that "we will encourage
decision making to occur at the lowest possible level in the university and we
will respect departmental autonomy and flexibility."
In retrospect, the union should have researched other faculty collective
bargaining agreements to demonstrate how these issues have been handled by
other unionized faculties in order to alleviate faculty members' concerns about
standardization and micro-management. In the final analysis, this response was
not sufficient in convincing faculty members who lacked sufficient knowledge
or who already had considerable doubts about faculty unionization.
A second lesson to be learned from this campaign is that significant
additional support for the union will not be obtained upon the completion of
the authorization card drive. I think that there was an unspoken belief among
many ISUFA Steering Committee members that we would be able to pick up
additional support from faculty members in the weeks before the certification
election. Absent a major crisis or tactical error on the part of the administra-
tion, after the conclusion of the card drive, the union's major job is to
consolidate and maintain faculty members' support for the union in light of
opposition from the administration and any anti-union groups that emerge.
While a small amount of additional support might be obtained from faculty
members, the more serious possibility is the loss of a significant number of
supporters who signed authorization cards; this is what happened with respect
to the ISUFA organizing drive. As stated earlier in this article, although the
union obtained 307 signed authorization cards, only 283 bargaining unit
members cast a vote in favor of the union. Thus, 8% fewer faculty members
actually voted for the union than signed cards. This indicates that the union
should be prepared to experience at least some erosion of support once an
opposition campaign is launched.
Related to the second lesson, a third lesson, under normal circumstances, is
that the union should not file for a certification election unless the union has
collected a minimum of 50% of authorization cards from union supporters in
the potential bargaining unit. Even though the ISUFA filed with a little over
40% of cards signed by bargaining unit members, many Steering Committee
members believed that we could still win the election because we would retain
virtually all of our support as well as be able to pick up the support of a decent
number of "undecided" voters at the last minute. However, empirical research
demonstrates the importance of collecting a high percentage of authorization
cards from bargaining unit members in order for the union to have a good
chance to win the representation election. For example, one study found that
when 0nly 30% (the legal minimum) of bargaining unit members signed cards,
unions won only 11.1% of certification elections. However, in organizing drives
where 80 to 89% of the employees signed cards, the union achieved victory in
71.4% of the elections (Graham & Neilsen, 1991).
This rule should be strictly adhered to unless there is convincing support that
many faculty members are union supporters but are not signing cards because
they feel intimidated, which is the situation that occurred at SIU. However, this
was not the case at ISU. In my experience in collecting cards during this
campaign, only a small number of union supporters refused to sign cards. These
were primarily non-tenured faculty who refused to sign because they feared
administration retribution if for some reason this information became public.
Although, as mentioned earlier in the article, there is empirical evidence that a
significant percentage of employers retaliate against union activists and union
supporters during organizing campaigns (Lawler, 1990), there was no evidence

(or even any allegations) that the university administration, college deans or
department chairs retaliated in any way against either union activists or
supporters throughout the duration of the unionization drive at ISU.
A fourth lesson is that the organizational structure of the union should be
established before filing authorization cards for the certification election. A
major problem with the campaign of the ISUFA was that it did not have active
(or even any) departmental representatives in many departments throughout the
university. Since the success of union organizing campaigns depends on
effective one-on-one communication (Bronfenbrenner, 1997, 1998) within
departments, these efforts will be the most successful if the people doing the
convincing of fellow faculty members are actual members of the departments
rather than faculty members from other departments. In my experience of
collecting authorization cards from faculty members in the Department of Art,
having me promote collective bargaining was better than nothing but not as
effective as if I was a member of that department who could have contact with
these faculty members on a regular, one-on-one basis.
This is, in fact, what occurred in a number of departments. Since we did not
have a departmental representative that was doing effective work, we sent in a
faculty member from another department in order to have some coverage in
that department although the coverage was in many cases less than adequate.
Even though we knew that this was a problem, we thought that we could recruit
additional departmental representatives, or at least activate the ones that we
had, upon filing for the certification election, which we believed would generate
an increased level of excitement and enthusiasm for the union and its campaign.
A fifth lesson is that, if it all possible, a unionization campaign should fully
utilize the emotional path (Wheeler & McClendon, 1991) towards support for
unionization by conducting the drive around some kind of "crisis" issue which
a significant percentage, if not the majority, of the faculty feel directly affects
them. This appears to be an important factor for union certification election
success because as I noticed at ISU, outside of many Steering Committee
members and a number of faculty members in the CAS, very few faculty
members at ISU arrived at their support for the union through the political/
ideological beliefs path (Wheeler & McCiendon, 1991). That is, only a small
number of faculty members were ideologically committed to unionization under
virtually any circumstance. Thus, in order to have a successful unionization
campaign, there must be a single issue, or a couple of related issues, for uniting
the faculty in their drive towards unionization.
This certainly was not the case with respect to the unionization campaign at
ISU; as mentioned before, there was no single issue which unified the faculty
in their support for faculty unionization. From my experience as the major

organizer in the COB, and consistent with the ISUFA's organizing strategy of
appealing to as many faculty members as possible, there was not one but a
variety of issues that led to faculty members in the college supporting collec-
tive bargaining. For example, for some faculty members in the college, salary
issues were predominant while for other faculty, the threat of a loss of shared
governance or the increased (and increasing) power of the administration and
the loss of individual autonomy were perceived as the major problems
confronting ISU faculty. In addition, the feeling that the university was moving
towards a "customer" orientation and watering down the educational experi-
ence led to support for unionization. Finally, perceived external threats to the
profession from the Illinois Board of Higher Education (IBHE) or the Illinois
State Legislature was sometimes enough to garner faculty support for union-
ization from a handful of professors.
It appears unlikely that there was a single issue that would have been able
to unite the faculty in support of the union in March 2000. If the certification
election had been held five years earlier, the emotional path (Wheeler &
McClendon, 1991) may have been successfully used by exploiting faculty dissat-
isfaction with the Wallace administration. This may have been the issue that
could have united and catalyzed faculty support for the union. The issue of
inadequate salaries was potentially a rallying point for the union but the feeling
that the faculty-friendly Boschini administration was working on improving
salaries prevented this issue from becoming the unifier for faculty support for
unionization. The only remaining issue that potentially could have unified all
ISU faculty members was the external threats posed by the ISU Board of
Trustees, the IBHE and the state legislature to the individual faculty member's
autonomy. The ISUFA could have spent more time discussing and focusing on
this issue although it is unlikely that such an orientation would have changed
the outcome of the election. When the union did raise this issue throughout the
campaign, it failed to resonate with a large number of faculty members.
One final (but nevertheless crucial) lesson learned from the ISUFA union-
ization campaign: organizing college and university faculty does not end with
the holding of the certification election. Many members of the Steering
Committee viewed the certification election as the single defining event for
organizing ISU faculty; they believed that organizing faculty ended once the
certification election had taken place. If the ISUFA had actually won the
election, many Steering Committee members did not realize that the union
would have to continue to organize the faculty for the upcoming struggles. For
example, few committee members realized that we would have to continue to
organize (and to recruit) faculty in order to build a viable local union. Besides
176 V I C T O R G. D E V I N A T Z
recruiting members to build a strong union local, continued faculty organizing

would also be necessary to obtain a strong first contract in order to establish a
base for obtaining future union gains from the university administration.
Since the ISUFA did not win the certification election, organizing of the
ISU faculty still must occur if the university's faculty members are to even-
tually obtain representation. Summers (1991) points out a wide range of
activities that a union can participate in even after it has lost a certification
election; one of his suggestions is to have the union act as the employees'
voice on all employment-related issues that might arise in the workplace. The
focus, of course, is to build towards another unionization campaign, hopefully
culminating in a successful certification election outcome so that the ISUFA
can be the collective bargaining agent that represents the ISU faculty.
ACKNOWLEDGMENTS
I would like to thank the editors, David Lewin and Bruce E. Kaufman, and an
anonymous referee for putting me through a stimulating reviewing process.
Because of their efforts, the quality of this article has been dramatically
improved.
REFERENCES
Many of the cited references in this article are based on material distributed by either the ISUFA,
the FSG, the university administration or other union opponents during the ISU organizing drive.
Copies of these cited references are in the author's possession. For any questions with respect to
this material, please contact the author at his university address.
Abacus Associates (2000). Illinois State University Faculty Opinions On Certifying ISUFA.
Commissioned by The National Education Association (September).
Ali, S. M., & Karim, A. R. (1992). An Empirical Examination of Determinants of Faculty Attitudes
Toward Collective Bargaining. Journal of Collective Negotiations in the Public Sector, 21(1),
79-91.
Allen, R. E., & Keaveny, T. J. (1981). Correlations of University Faculty Interest in Unionization:
A Replication and Extension. Journal of Applied Psychology, 66(5), 582-588.
Annunziato, F. R. (1994). 1994 Directory of Faculty Contracts and Bargaining Agents in Institutions
of Higher Education, The National Center for the Study of Collective Bargaining in Higher
Education and the Professions - Baruch College, The City University of New York, New
York.
Aronowitz, S. (2000). The Knowledge Factory: Dismantling The Corporate University And Creating
True Higher Learning. Boston: Beacon Press.
Aronowitz, S. (1997). Academic Unionism and the Future of Higher Education. In: C. Nelson (Ed.),
Will Teach For Food: Academic Labor in Crisis (pp. 181-214). Minneapolis: University of
Minnesota Press.
Aussieker, B. (1976). Students and Bargaining at Public and Private Colleges. Monthly Labor
Review, 99(4), 31-33.
Babcock, L., & Engberg, J. (1999). Bargaining Unit Composition and the Returns to Education
and Tenure. Industrial and Labor Relations Review, 52(2), 163-178.
Barbezat, D. A. (1989). The Effect of Collective Bargaining on Salaries in Higher Education.
Industrial and Labor Relations Review, 42(3), 443455.
Begin, J., & Browne, S. (1974). The Emergence of Faculty Bargaining in New Jersey. Community
and Junior College Journal, 44(4), 18-19.
Bigoness, W. J. (1978). Correlates of Faculty Attitudes Toward Collective Bargaining. Journal of
Applied Psychology, 63(2), 228-233.
Boschini, V. J., & Goldfarb, A. (2000). A Letter to Colleagues. (March 3).
Bronfenbrenner, K., & Juravich, T. (1998). It Takes More Than House Calls: Organizing to
Win with a Comprehensive Union-Building Strategy. In: K. Bronfenbrenner, S. Friedman,
R. W. Hurd & R. L. Seeber (Eds), Organizing to Win: New Research on Union Strategies
(pp. 19-36). Ithaca, New York: CorneU University Press.
Bronfenbrenner, K. (1997). The Role of Union Strategies in NLRB Certification Elections. Industrial
and Labor Relations Review, 50(2), 195-212.
Cameron, K. (1982). The Relationship Between Faculty Unionism and Organizational Effectiveness.
Academy of Management Journal, 25(1), 6-24.
Can the Tail Really Wag the Dog? (2000). Faculty for Shared Governance Informational
Memorandum Number 6. (February 21).
Can You Hear My Voice? (2000). Faculty for Shared Governance Informational Memorandum
Number 4. (February 16).
Celuch, K., Goodwin, S., Leinicke, L., Love, D., Newgren, K., Osaosky, J., Rexroad, M., Showers,
L., Trefzger, J., & Williams, M. (2000). Letter to COB Colleagues. (March 7).
Cohen, L., & Hurd, R. W. (1998). Fear, Conflict, and Union Organizing. In: K. Bronfenbrenner, S.
Friedman, R. W. Hard & R. L. Seeber (Eds), Organizing to Win: New Research on Union
Strategies (pp. 181-196). Ithaca, New York: Cornell University Press.
Demsetz, R. S. (1993). Voting Behavior in Union Representation Elections: The Influence of Skill
Homogeneity and Skill Group Size. Industrial and Labor Relations Review, 47(1), 99-113.
Douglas, J. M. (1990). The Impact of NLRB v. Yeshiva University on Faculty Unionism at Public
Colleges and Universities. Journal of Collective Negotiations in the Public Sector, 19(1), 1-28.
Feuille, P., & Blandin, J. (1974). Faculty Job Satisfaction and Bargaining Sentiments: A Case Study.
Academy of Management Journal, 17(4), 678-692.
Garabino, J. (1980). Faculty Unionization: The Pre-Yeshiva Years, 1966-1979. Industrial Relations,
19(2), 221-230.
Garabino, J. (1973). The Emergence of Collective Bargaining. In: E. D. Duryea, R. S. F i s k &
Associates (Eds), Faculty Unions and Collective Bargaining (pp. 1-19). San Francisco: Jossey-
Bass.
Graf, L. A., Hemmasi, M., Newgren, K. E., & Nielsen, W. R. (1994). Profiles of Those Who Support
Collective Bargaining in Institutions of Higher Learning and Why: An Empirical Examination.
Journal of Collective Negotiations in the Public Sector, 23(2), 151-162.
Graham, H. E., & Neilsen, K. N. (1991). Union Representation Elections: A View From the Heart
of It All. Labor Law Journal, 42(7), 438-441.
Gress, J. (1976). Predicting Faculty Attitudes Toward Collective Bargaining. Research in Higher
Education, 4(3), 247-256.
Harris, C. (2000). What Are the Options? Faculty for Shared Governance Informational
178 VICTOR G. D E V I N A T Z
Has Shared Governance Broken Down at ISU? (2000). Faculty for Shared Governance Informational
Hemmasi, M., & Graf, L. A. (1993). Determinants of Faculty Voting Behavior in Union
Representation Elections: A Multivariate Model. Journal of Management, 19(1), 13-32.
Heneman, H. G., & Sandver, M. H. (1983). Predicting the Outcome of Union Certification Elections:
A Review of the Literature. Industrial and Labor Relations Review, 36(4), 537-559.
Holley, W. H. Jr., Jennings, K. M., & Wolters, R. S. (2001). The Labor Relations Process. (7th
ed.). Fort Worth, TX: Harcourt College Publishers.
Hurd, R. W., & McElwain, A. (1988). Organizing Clerical Workers: Determinants of Success.
Industrial and Labor Relations Review, 4l(3), 360-373.
Illinois State University Faculty Association Platform (1999). (January 21).
Jones, L. (1986). The Impact of Faculty Unions on Higher Education: A Reconsideration. Journal
of Collective Negotiations in the Public Sector, 15(1), 79-87.
Karim, A. R., & Ali, S. M. (1993). Demographic Differences and Faculty Attitude Toward Collective
Bargaining. Journal of Collective Negotiations in the Public Sector, 22(1), 87-97.
Kaufman, B. E., & Martinez-Vazquez, J. (1988). Voting for Wage Concessions: The Case of the
1982 GM-UAW Negotiations. Industrial and Labor Relations Review, 41(2), 183-194.
Kochan, T. A. (1980). Collective Bargaining and Industrial Relations. Homewood, IL: Irwin.
Ladd, E. C., & Lipset, S. M. (1973). Professors, Unions, and Higher Education. Berkeley, CA:
Carnegie Commission on Higher Education.
Lawler, J. J. (1990). Unionization and Deunionization: Strategy, Tactics, and Outcomes. Columbia,
SC: University of South Carolina Press.
Magney, J. R. (1999). Faculty Union Organizing on the Research Campus. Thought & Action,
15(1), 111-126.
Maranto, C. L., & Fiorito, J. (1987). The Effect of Union Characteristics on the Outcome of NLRB
Certification Elections. Industrial and Labor Relations Review, 40(2), 225-240.
McClure, B. (1999). Tenure at the University of Minnesota: A Postmortem. Thought & Action,
•5(2), 97-104.
Meisenhelder, T. (1986). The Class Position of College and University Faculty. The Social Science
Journal, 23(4), 375-389.
Miller, P. (2000). Illinois State University Faculty Association Post-Election Meeting Minutes (29
March).
Mills, D. (2000). Letter to Colleagues in the College of Business. (March 7).
Newgren, K. (2000). Letter to COB Colleagues (March 7).
Ponak, A., Harel, G., Thompson, M. & Kedem, D. (1987). Faculty Collective Bargaining: A Cross-
Cultural Survey. Journal of Collective Negotiations in the Public Sector, 18(3), 219-232.
Ponak, A., Thompson M., & Zerbe, W. (1992). Collective Bargaining Goals of University Faculty.
Research in Higher Education, 33(4), 415--431.
Pratt, L. R. (1997). Disposable Faculty: Part-time Exploitation as Management Strategy. In:
C. Nelson (Ed.), Will Teach For Food: Academic Labor in Crisis (pp. 264-277).
Minneapolis: University of Minnesota Press.
Pratt, L. R. (1994). A New Face for the Profession. Academe, (September-October), 38--41.
Promise Us Everything, Give U s . . . What? (2000). Faculty for Shared Governance Informational
Rastin, S. (2000). Organizing Tactics in a Faculty Unionization Drive at a Canadian University.
Labor Studies Journal, 25(2), 99-116.
Rhoades, G. (1998). Managed Professionals: Unionized Faculty and Restructuring Academic Labor.
Albany, NY: State University of New York Press.
Rosenthal, J. T., Cogan, M. L., Marshal, R., Meiland, J. W., Wion, P. K., & Molotsky, I. F. (1994).
Report: The Work of Faculty: Expectations, Priorities, and Rewards. Academe,
(January-February), 35-48.
Stanford, S. (2000). Frequently Asked Questions Concerning the Upcoming Union Vote (January
21).
Strickland, R. (2000a). The Best Option (February 23).
Strickland, R. (2000b). The Big Dog and its Tail (February 23).
Summers, C. (1991). Unions Without Majorities: The Potentials of the NLRA. Proceedings of the
43rd Meeting of the Industrial Relations Research Association, Madison, WI.: IRRA:
154-162.
Thompson, K. (1997). Alchemy in the Academy: Moving Part-time Faculty from Piecework to
Parity. In: C. Nelson (Ed.), Will Teach For Food: Academic Labor in Crisis (pp. 278-290).
Minneapolis: University of Minnesota Press.
Watt, S. (1997). On Apprentices and Company Towns. In: C. Nelson (Ed.), Will Teach For Food:
Academic Labor in Crisis (pp. 229-253). Minneapolis: University of Minnesota Press.
Wheeler, H. N., & McClendon, J. A. (1991). The Individual Decision to Unionize. In: G. Strauss,
D. G. Gallagher & J. Fiorito (Eds), The State of the Unions (pp. 47-83). Madison, WI:
Industrial Relations Research Association.
White, M. D. (1982). The Intra-Unit Wage Structure and Unions: A Median Voter Model. Industrial
and Labor Relations Review, 35(4), 565-577.
Williams, G. B., & Zirkel, P. A. (1989). Shift in Collective Bargaining Issues in Higher Education:
A Review of the Literature. Journal of Collective Negotiations in the Public Sector, 18(1),
73-86.
Would a Union Raise Faculty Salaries? (2000). Faculty for Shared Governance Informational
Memorandum Number 2 (February 10).
Zalesny, M. D. (1985). Comparison of Economic and Noneconomic Factors in Predicting Faculty
Vote Preference in a Union Representation Election. Journal of Applied Psychology, 70(2),
243-256.
THE AGING WORKFORCE AND THE
NEXT TURNING POINT
Daniel J. B. Mitchell
INTRODUCTION
The most important thing we do know from the 20th century that continues into
the 21st is that there was a baby boom in the former that is now aging in the
latter. Until the elderly boomers disappear - presumably in the 2060s absent a
medical miracle - they will have a major influence on American society. The
last act of the boomers will be to cause a turning point in the employment rela-
tionship and related social policy. In the 20th century, two such turning points
were triggered by economic fluctuations. But the next one will be demographic.
TURNING POINT h THE DEPRESSION/

WORLD WAR II PERIOD
Before the Great Depression, the American economy was laisser-faire in

character with a thin veneer of welfare capitalism at some large firms. The onset
of the Depression tended to undermine welfare capitalism, although - as my
colleague Sanford Jacoby (1997) points out - welfare capitalism was not
obliterated. Business as an institution was blamed for the economic debacle. A
move toward government intervention and the rise of unions produced mutual
reinforcement. Public policy supported unionization; unions - especially the
more radical CIO variety - supported a wider government role in the economy.

ISBN: 0-7623-0853-2
181
182 DANIEL J. B. MITCHELL
Demographics also played a role in creating new institutions in the 1930s.

Elderly assistance had been the responsibility of local authorities and private
charities before the New Deal. Poorhouses and (very) limited cash relief were
the basic instruments of old-age support. For men, working until health or death
forced an end to such activity was the norm. But the Depression made work
unavailable and families, localities, and charities much less able to care for
elderly persons.
The result was formation of what was termed at the time an "army of the aged"
that sought government pensions (Neuberger & Loe, 1936). Mainstream history
points to Social Security as the response to these elderly concerns. Yet even when
enacted, Social Security did not pay benefits until the 1940s, and many elderly
were ineligible. In fact, the army of the aged backed more radical schemes such
as the federal Townsend Plan, inadvertently providing Roosevelt's Social
Security proposal with a critical margin of support. 1 California, home of the
Townsendite agitation, saw the rise of the Ham and Eggers and other wacky
social movements aimed at harnessing electoral strength of elderly voters behind
state-level pension plans 2 (Mitchell, 2000).
W o r d War II required another dose of government economic control.3 And
the government's need for uninterrupted production of war materiel led to a
reinforcement of union growth. Employers came under pressure to recognize
unions to avoid strikes over recognition. Meanwhile, wartime wage controls
spread norms of contractual labor-management relationships.
THE NEW POSTWAR EQUILIBRIUM
Postwar America produced a new employment relationship centered on

collective bargaining in major industries. Government would provide social
insurance for non-workers, e.g. Social Security pensions (and later Medicare)
for retirees. It would aid other non-workers through unemployment insur-
ance, disability benefits, and "welfare" (later supplemented with Medicaid).
But at the workplace, welfare capitalism revived with government subsidy
through the tax code. Employers would provide health insurance for
employees and their dependents and pensions to be paid on top of the basic
Social Security entitlement. Workers could win or enhance such benefits
through collective bargaining. Or employers might simply choose to provide
or improve them. For those workers unlucky enough to find themselves at
workplaces where neither bargaining nor employer beneficence prevailed,
government would set a minimum wage and designate other limited labor
standards.
The Aging Workforce and the Next Turning Point 183
TURNING POINT I1: THE 1980S
The 1980s began with a severe slump - the worst since the Great Depression -
that pushed unemployment above 10%. A shift to the right occurred politically
with the advent of the Reagan administration. The union sector experienced
substantial erosion. Union-representation fell by 3.1 million workers between
1980 and 1985. Concession bargaining blossomed. Wages were frozen or cut at
many union workplaces. In others, de-unionization was pursued.
Unions had been the strongest manifestation of the postwar employment
relationship. But other aspects of that relationship began to crumble, too.
Contingent work began to grow. Job-based health care coverage began to slip.
A shift occurred from defined-benefit pension plans (favored by unions) to
defined-contribution plans. There was much rhetoric about workers managing
their own careers, saving for their own retirements, and generally being respon-
sible for themselves. Security was said to be a thing of the past. Job tenure of
older male workers began to erode. By the early 1990s, a mild recession
produced layoffs among managers and professionals, groups that had previously
been relatively immune from the business cycle.
THE S ~ F T I N G BALANCE OF ~ W E R
Although the prevailing view was that the labor market had become more
"competitive" by the 1990s, there was a puzzling element in pay determina-
tion. In the second half of the 1990s, a notable drop in unemployment occurred
as part of the Internet/stock market boom. Unemployment fell to levels not seen
since the Vietnam War. Employers around the country complained of labor
shortages and not just for high-tech employees. Construction workers, truck
drivers, and restaurant personnel were all said to be in short supply) In truly
competitive markets (gold futures, pork bellies, T-bills), shortages bring about
sharp price rises. Surely, in a competitive labor market, a wage explosion should
have occurred in the late 1990s. But it didn't.
Since that Great Depression experience, economists have been fretting over
why wages don't fall continuously in recessions. Usually, the economist's
answer has been some variant of the idea that wage cuts are bad for morale
and productivity (Bewley, 1999). Even in Hard Times, employee resistance,
whether organized or not, is able to prop up the wage. But this explanation is
not symmetrical since wage increases in boom times are not bad for morale
and employees surely do not resist them. So if wages don't respond like pork
belly prices to shortages, the answer must lie in the area of employer
resistance. Labor shortages during the Vietnam boom, roughly 1965-1969, did
push up wage inflation, with unions leading the way. 5 During the Internet/stock
market boom (1995-2000), wage inflation was relatively quiescent despite labor
shortages and certainly not more responsive than in the 1960s. 6 Employer power
was retarding upward wage responsiveness.
THE COORDINATED LABOR MARKET
Unlike some corporatist European countries, the U.S. has relatively decentralized
labor market institutions. Yet as Arthur Ross (1948) pointed out in the 1940s, the
American labor market has certain propensities for coordination. In a formal
sense, postwar pattern bargaining produced wage imitation across certain union
settlements. But in a less formal way, norms of wage setting and personnel
practice spread much wider, spilling over from the union to the non-union sector.
Providing health insurance and adjusting wages for inflation - even during
recessions - were examples of such norms. The norms were reinforced through
government action, e.g. tax-favored treatment of employer-provided insurance,
timely publication of the CPI so that wage setters could respond.
To put it controversially, the Depression/World War II experience fostered a
loose "employee cartel," activated through unions and favorable government
action and coordinated through imitation. But like OPEC and other cartels,
the employee cartel could overreach. And it did in the 1970s when the union/
non-union wage premium was pushed up in the face of adverse market conditions,
setting the stage for the second turning point.
American employers, too, can behave as an informal cartel. Through trade
associations, personnel publications, management consultants, and the popular
press, they discover and imitate what other employers are doing. At times
employers, too, can harness government action. Business-friendly think-tanks
can foster ideas about deregulating the labor market. Columnists and TV talking
heads can push the idea that old postwar employment norms are dead. They can
popularize the idea that employees should manage their own careers and not
expect security or benefits from employers. As these ideas circulate, employers
are cued to act in a loosely coordinated manner, implementing the new norms.
No conspiracy is implied; just the institution of imitation and "benchmarking."
It is not surprising in retrospect that the 1980s should have seen the ascen-
dancy of employer power and the eclipse of its employee counterpart. The cost
to employers of providing job security varies with the risk being insured. By the
1980s, uncertainty over economic conditions such as inflation was greater than
before. 7 Profit variability was also rising, in part because of the marked shift to
debt finance. 8 American abandonment of fixed exchange rates in the early 1970s
added to the uncertainty of the 1980s. Floating exchange rates had proved volatile
and - as the dollar peaked relative to foreign currencies in the mid-1980s -

American competitiveness was undermined. New sources of foreign competition
were arising in Asia. Product market deregulation in transportation, finance, and
communications made the outlook there more uncertain from the employer
perspective. The postwar deal became increasingly costly.
Initially, union wage premiums began to crumble at distressed firms during
the early phase of concession bargaining in the 1980s. And the message con-
veyed to other not-so-distressed firms was that unions were not as invincible as
once thought. More and more finns demanded concessions. With the Reagan
administration in office, there was less political constraint on such behavior.
Ultimately, if wages and benefits of union workers were shown to be vulnera-
ble, why should not the same be true of non-union? Why not look for savings in
labor costs throughout the workforce? Why bear the burden of the cost of
providing job security, even for managers and professionals?
To summarize, the 20th century saw a kind of jockeying between de facto
employer and employee cartels. Starting in the 1930s, the employee side
exercised power through unions and favorable government regulation. As Table
1 suggests, once strengthened by the Depression/World War II experience,
employees managed to increase their share of the corporate pie through the late
1970s. 9 The employer side tried to weaken unions and counter government
regulation in response. It sought to attenuate commitments entailed in postwar
employment norms. Table 1 suggests that employers checked the power of
employees during the 1980s and rolled it back in the 1990s. But this employer
success will soon collide with demographics.
DEMOGRAPHICS OF THE 21ST CENTURY
Throughout their history, baby boomers have been perceived as "trouble." As

children in the 1950s, they were the Johnnies who couldn't read. Then they were
teenage delinquents. In the 1960s, they were protesting college students, urban
rioters in minority neighborhoods, and rebellious strikers on factory floors. They
were said to push up structural unemployment in the 1970s, as they entered the
job market in large numbers. And in the 1990s, they were self-indulgent
consumers who were not saving for retirement.
Whatever the justification for these past images, as the boomers move through
the labor force and into retirement during the next few decades, they will pose
two problems. First, as aging workers, they will want job security, a commit-
ment from which employers have sought to escape. Second, as retirees, they
will want adequate income and health care, demands that require diverting
consumption from younger active workers.
Table 1. Labor's Percentage Share of Corporate National Income.
1929 74.5% 1979 82.3%

1949 76.6 1989 81.8
1959 78.1 2000 79.6
1969 80.3
Source: Bureau of EconomicAnalysis, U.S. Departmentof Commerce

If employers persist in pushing their current advantage, they will soon meet
resistance from its aging boomer workforce. Although some resistance might
take the form of union revival, it is much more likely to take political and legal
shape. Left/right political leanings of boomers concerning other socio-
economic issues may not be determinative. Neither liberals nor conservatives
want to be laid off. Neither wish to have their health care benefits denied or
rationed. Neither want to experience adverse conditions at work that threaten
mortgages or junior's college tuition or support of elderly parents.
Glimmers of this resistance can already be seen in legislative battles over
aspects of health care and litigation over attempts by employers to shift from
defined benefit pensions to cash balance plans. As the Dunlop Commission
pointed out in the mid-1990s, battles that used to be fought at the bargaining
table are now being joined in courtrooms. Look for matters related to age dis-
crimination and job-based pensions and health care to be the future issues of
choice for litigants. Rather than the pattern bargaining of the post-World War
II era, we may well have patterns in lawsuits - successes triggering imitation
suits - and legislation spilling from one state to another.
And what happens when the boomers retire? Who will provide resources for
their consumption, health needs, or long-term care? Virtually none of the
proposals currently in play to "privatize" Social Security do anything to enlarge
the economic pie in, say, 2030. They simply restructure the way in which
consumption tickets will be provided to retired boomers. If the individual accounts
which such proposals entail pay off for the boomers, i.e., provide them with more
income than they would have obtained from the present system, there will be less
of the pie available for active workers. That real wage squeeze - however it occurs
- could well trigger employee dissatisfaction aimed at employers. If the accounts
don't pay off, the boomers will feel cheated. The new army of the aged of the
21st century will then push for legal and legislative remedies regarding employer
pensions and whatever Social Security has become. As in the 1930s, the army of
the aged will become a third force along with employees and employers, demand-
ing its share of the economic pie. Unless the "new economy" provides so much
prosperity that everyone will be satisfied - an increasingly dubious proposition -
the 2030s may well be as exciting as the 1930s. I°
NOTES
1. The Townsend Plan promised elderly persons over 60 $200 a month if they
promised not to work (leaving jobs for the young) and to spend every penny during the
month (to stimulate the economy). $200/month ($400 for a couple) was an enormous
sum at the time.
2. The Ham and Eggs proposition on the California ballot in 1938 and 1939 offered
those over age 50 "Thirty Dollars Every Thursday" to be paid in a new California
currency. As in the Townsend proposal, recipients were forbidden to work. Ham and
Eggs received 45% of the vote in 1938 and might well have passed had the antics of
its con-men promoters not been brought to light. Pensionite movements in California,
an elderly state at the time, roiled state politics in the 1930s and 1940s. Related move-
ments developed in other parts of the country such as Colorado and Ohio, also elderly
states.
3. At the war's peak, over 40% of GDP was going to the military.
4. I searched the phrase "labor shortage" in Nexis/Lexis under "NEWS" and
"REGION" annually for 1995-2000. In 1995, the number of news media citations in
the Midwest, Northeast, Southeast, and West were, respectively, 211, 52, 124, and 106.
In 2000, the figures were 610, 566, 525, and 545.
5. Average hourly earnings inflation rose from around 4% per annum at the begin-
ning of the Vietnam boom to about 7% at the end. First-year wage increases under
major union settlements rose from about 4% to 8%. There were 268 major strikes in
1965 and 412 in 1969. Unemployment fell from 4.5% to 3.5%.
6. Union wages lagged non-union and strikes were rare in this period. According to
the Employment Cost Index, union wage inflation accelerated from 2.6% in 1995 to
3.4% in 2000. Non-union wage inflation accelerated from 3.6% to 4.0%. Unemployment
fell from 5.4% to 4.0%. There were 31 major strikes in 1995 and 39 in 2000.
7. The standard deviation of GDP price index inflation over the prior ten years had
fallen below 1% by the early 1960s. By the mid-1980s, it was about 21/2%. The
standard deviation of the percent share of before-tax profits in corporate national income
over the prior ten years rose from 2% or less in the 1960s to about 4% in the
mid-1980s.
8. Net interest as a percent of capital's share of corporate national income was
essentially zero until the mid-1960s. By the 1980s, it varied within the 20-30% range.
9. Corporate national income can be divided into the share going to labor (wages,
benefits, payroll taxes), the share going to profits (as defined in national income
accounting terms), and the share going to net interest. Labor's share of income is very
sensitive to the business cycle because profits are business-cycle sensitive. So labor's
share is boosted by recessions. The dates shown in Table 1 represent rough business
cycle peaks as dated by the National Bureau of Economic Research. At this writing,
it remains to be seen whether 2000 was a peak, i.e. whether 2001 will mark a
recession.
10. Note that the foreign equivalents of the baby boom - in Japan, Europe, and even
China - have been accumulating claims on the U.S. thanks to large and ongoing
American deficits in net exports. These claims will have to be "cashed in" as foreign
boomers retire and consume more than they produce. Thus, active American workers
will be supporting both foreign and domestic boomers.
188 DANIEL J. B. M I T C H E L L
REFERENCES
Bewley, T. F. (1999). Why Wages Don't Fall During a Recession. Cambridge, MA.: Harvard
University Press.
Jacoby, S. M. (1997). Modem Manors: Welfare Capitalism Since the New Deal. Princeton, N.J.:
Princeton University Press.
Mitchell, D. J. B. (2000). Pensions, Politics, and the Elderly: Historic Social Movements and Their
Lessons for Our Aging Society. Armonk, N.Y.: M.E. Sharpe.
Neuberger, R. L., & Low, K. (197311936]). An Army of the Aged: A History and Analysis of the
Townsend Old Age Pension Plan. New York: Da Capo Press.
Ross, A. M. (1948). Trade Union Wage Policy. Berkeley: University of California Press.
FROM WORKPLACE TO
CORPORATE GOVERNANCE:
LEVERAGING WORKER ASSETS
IN THE 21ST CENTURY
Stephen R. Sleigh
INTRODUCTION
The last quarter century was a difficult time for organized labor in the United
States. In the private sector, both absolute and relative measures of union
strength were headed in the same direction: down. In 1973 unions represented
nearly 15 million workers in the private sector, or 25% of the workforce; by
1999 that figure had dropped to 10 million workers and under 10% of the
workforce.JThis decline has forced organized labor to take measure of their
past and present position in the labor market and to make critical strategic
decisions about the future. One such strategic approach is the harnessing and
leveraging of organized labor's capital resources to further the goal of
increasing its power in the labor market and balance the overwhelming power
management have accumulated. This brief paper examines the recent devel-
opment of capital strategies by one labor organization, the International
The viewspresentedin this paperare thoseof the authorand not necessarilythoseof the International
Associationof Machinists and AerospaceWorkers.

ISBN: 0-7623-0853-2
189
i90 STEPHEN R. SLEIGH
Association of Machinists and Aerospace Workers (IAM), and assesses how

capital strategies may influence industrial relations in the years ahead.
The 20th century experience for most unions in the United States is
characterized as narrowly focnsing on bread and butter issues such as wages,
hours, and working conditions. Both the business and industrial models of
unionism adopted ~this focus, a~heit with di~erertt organizing strategies. As,a field
of research, industrial relations studied the,ramifications of this workplace focus
and help shape an understanding of how systerns~of governance work at the work-
place. The genre reached its peak by mid,century with .the publication o f John
Dunlop's Industrial Relations Systems. 2 I)unlop's system identified the salient
components of modern industrial society as the interaction between workers and
.their organizations, managers and their organizations, and government agencies
concerned with the workplace) Until the turmoil in the labor market in the 1980s
and 1990s, that description of a fairly stable system adequately characterized the
attributes of labor-management relations in the :United States.
The precipitous decline of organized labor and the cutting back of govern-
ment's role 4n regulating the collective side of the labor market that occurred
during the 1980s and 1990s occurred simultaneously with the ascendancy of the
shareholder value 'model of corporate governance that drove much of the
downsizing and union avoidance during that same period. In my view, the current
state o f industrial relations has resulted in the most imbalanced relationship
.between employees and employers since the 1920s. A result of this imbalance is
Clearly seen in the detailed survey work done by ~ h a r d Freeman and Joel
Rogers published 4n What Workers Want. 4 Employees are dissatisfied with the
level and means Of representation at the workplace. One response to the imbal-
ance between labor and management, and a means to increase workers voice in
an increasing number of areas is the expansion of organized labor's role beyond
the level of workplace governance to that of corporate governance. What may
be different in the years ahead is how labor uses capital strategies to move up a
notch in terms o f importance as partners in governance of the corporation.
Organized labor, if it is to regain its vibrancy and central role as the collective
voice for employees, will play a role in who makes investment decisions in
corporations, what types of investments they make, and how returns from invest-
ments are distributed. 5
The first section reviews past union attempts to expand beyond the traditional
sphere of wages, hours, and working conditions. This is a cursory overview at
best. No doubt many creative attempts to move into co-management of industry
can be cited particularly in the garment industry. 6 The point is that organized
labor in the United States did not develop a systematic approach to expanding
into corporate governance until fairly recently. After setting the context of
From W o r l ~ e to Corporate Governance 19t
today's labor market, and the heavy reliance on consumer debt to. mainlain living
standards, the following sections look at two inila'atives latmched by the I_AM it~
the late 1990s that use capital strategies: creation of a mrtion-specific, mutual ftmd
and efforts to bring executive compensation into alignment with overall
organizational objectives. These two examples provide a snapshot of what is
rapidly becomi~ag a streaming series of activities by unions in the area of
corporate strategy. Taken together, the parts are beginning to add up to more than
the pieces as organized labor comes to the realization that one of its greatest
strengths is the combined purchasing power and retirement assets of its members.
The conclusion will circle back to the main theme of what can IR in the 21st
century learn from IR in the 20th century.
LOOKING BACKWARDS: SUCCESS AND FAILURE

OF NON.TRADITIONAL TACTICS
The IAM, like other AFL unions, experimented with mutual aid programs and
more sophisticated capital strategies from its earliest days. In the 1920s, for
example, the IAM created the Mt. Vemon Savings Bank. For thirteen years,
from 1920 to 1933, Mt. Vernon "brought a good return to the IAM, which had
purchased some of the original s t o c k . . . [F]or most of the thirteen years of its
existence, [the Bank] was a source of great pride to the [IAM]. ''7
Despite the apparent success of the Mt. Vernon Savings Bank the idea of
leveraging its assets, of expanding the bank's activities into areas that supported
the union's goals of expanding and servicing members, never really took off.
With the collapse of so many savings banks during the depression in the 1930s,
this effort fell apart.
Many years later the idea of harnessing the pension power that union's had
amassed during the fat years from World War II to the mid-1970s came to the
fore. In their book, The North Will Rise Again, Randy Barber and Jeremy Rifldn
provided a framework for harnessing employee pension funds. Written in 1978,
some of the predictions now seem hopelessly naive and misguided - that for
example, capitalism was on its last legs. Underneath the rhetoric was a
compelling idea: workers, and their unions, had a tremendous tool to use in
their pension funds if only they could learn how to manage these assets for
growth, which would ensure the fund's objective of providing secure pensions,
and investments in union-friendly businesses which would ensure the growth
of union-represented enterprises. 8
At the same time Rifkin and Barber were laying out their framework for a
socialized economy based on worker pensions, management's guru, Peter
Drucker, also recognized the power of pension funds. In his book, The Unseen
192 STEPHEN R. SLEIGH
Revolution: How Pension Fund Socialism Came to America, written in 1976,

Drucker takes the same set of facts and turns the analysis inside out arguing
that all that ownership will add up to a more committed employee loyal to
capital, not labor. 9
Despite the predictions from the left and right, pension fund activism did not
live up to its hilling throughout the 1980s or 1990s. With the election of John
Sweeney as head of the AFL-CIO in 1995, capital strategies received its first big
push from the top of the labor movement. Shortly after taking office, Sweeney
created the Office of Investments, and in 1999 launched the Center for Working
Capital. Under the direction of the AFL-CIO's Richard Trnmka both of these
initiatives gave capital strategies a front row seat in the ongoing effort to expand
workers' voice opportunities. As Trumka noted in 1999, "Worker funds account
for more than $3 trillion in capital investment, making them a major force in the
global markets. We're making sure that worker beneficiary voices are heard on
Wall Street. ''l° Through the efforts of the AFL-CIO and its affiliate unions, such
as the IAM, many of the ideas that Rifkin and Barber, and Drucker, thought
would come of age in the 1980s are now coming to fruition in the first decade
of the 21 st century.
PIECES OF A CAPITAL STRATEGY:

THE IAM E X P E R I E N C E
News about the economy over the past eight years (1992-2000) tends towards
a relentlessly optimistic view. From all reports it appears the business cycle
has officially been repealed. President Clinton summarized this view in his State
of the Union speech in 1999 by saying, "Because of the hard work and high
purpose of the American people, these are good times for America. We have
more than 14 million new jobs. The lowest unemployment in 24 years. The
lowest core inflation in 30 years. Incomes are rising, and we have the highest
home ownership in history." The business community second's the motion with
the recent statement from the Business Council, You could title our economic
report as The Best of Times . . . . [n]ot one of us has seen better times than
these.tl Working families see the world differently. Despite the eight-year long
economic expansion, working families are working longer and harder to
maintain their standard of living. One measure of the mismatch between the
view from the top and that from the middle is the surging consumer debt that
working families are taking on. Recognizing the economic reality our members
face the IAM undertook an effort to create a union friendly savings device the
IAM Shares mutual fund.
From Workplace to Corporate Governance 193
UNEQUAL TIDES: INCOME INEQUALITY DURING

BOOM TIMES
By now, the subject of income inequality has become commonplace among
economic analysts, policy makers, and union leaders. The Economic Policy
Institute, a labor-supported but fully independent economic think tank, has virtu-
ally made careers out of documenting the dual nature of the U.S. economy.
The latest figures do show that, finally, low and middle wage workers are bene-
fitting from the growing economy. Nonetheless, workers' wages have remained
stagnant for the past twenty years, with real wage increases occurring only in
the last two years since 1979. With unemployment down to 4.3% in April,
1999, the tight labor market is producing real gains in hourly earnings for the
first time in almost twenty years. The real median wage grew 3.1% from the
first quarter of 1998 to the first quarter in 1999. j2 Overall the increase in real,
inflation-adjusted, wages brought the median earner back to the level attained
in 1989. Male wage earners still lag behind the level of ten years ago, trailing
by 2.1% compared to 1989 for the median wage earner. The gains of the past
year were particularly strong at both the bottom and top 10% of the hourly
wage distribution. Male workers' wages increased 5.1% over the last year at
the bottom decile, and 3.9% at the top decile. The increase at the lower end
of the wage spectrum appears to be tied to the minimum wage, as their
percentage gains were greatest in the two periods over this business cycle when
the minimum wage was increased, j3
While these gains are certainly welcome to working families they pale in
comparison to the poor performance of the U.S. economy in lifting the living
standards of the vast group of people who work for wages and who make up
the middle class. From 1989 through 1997, median family income rose by
only $285 a year, or 0.6%. The typical family got their 0.6% raise through
hard work. The typical family worked 247 hours more per year in 1997
compared to 1996, while at the same time the productive capacity of the
economy improved about 8%. Bottom line, American families are working
harder to stay in the same place and are seeing little of the gains in the overall
economy. ~4
That is in fact what is going on in today's economy. One of my tasks at the
IAM is to provide assistance to our negotiators in the collective bargaining
process. We also do a quarterly summary of wage and contact settlement trends.
In this tight labor market, with skilled workers in high demand, we should be
negotiating wage packages that make up for the last twenty years of stagna-
tion. Instead, most of our agreements are coming in around 3% for initial wage
increases. That is better than concession bargaining for sure; and, it is true that
more settlements are coming in over 4% than under 2%, which again is a
measure of the strength of today's economy.
The truth, however, is that our members, and I believe this is true for other
wage earners, are more concerned about job security than about wage increases.
The threat of job loss in the manufacturing sector, where production can and
often is moved to lower labor cost areas, particularly industries that don't
compete in fight just-in-time delivery cycles where production can be moved
overseas, explains why job security is issue number one in all of the pre-
bargaining surveys we conduct. The effect of globalization, with its threat to
move production and resulting job loss, has been to weaken workers resolve
to capture their fair share of productivity increases. Unions, like the IAM, are
determined to put resolve back into a workforce that has been beaten down
pretty low.
KEEPING UP WITH THE JONESES
Workers may not have the same determination or organizational capacity to

take on employers for their fair share of the economic pie as they did twenty
years ago, but they have not changed their consuming habits. Indeed, the way
working families have coped with stagnant wages and aggressive employers is
to work more hours. For men, that means on average three hours per week
more in overtime than twenty years ago. For women, it has meant that the
opportunity to stay home and not to join the work force is increasingly a distant
option. Think about the forty-hour work week. The labor struggles at the end
of the last century were waged to bring down the work day to eight hours. One
hundred years later, just try to be a union official selling the idea of flex time,
work sharing, or reducing overtime opportunities. There is no surer path to
getting a one-way ticket back into the plant!
What we have is an economy that no longer produces regular gains for
typical wage earners, that no longer corresponds productivity gains and wage
gains, but which lives on the precept of ever greater consumption. How to
afford the middle-class lifestyle and expanding consumption? Work longer
hours, have more family members work, and run up ever increasing levels
of debt.
Juliet Schor has written a wonderful book about this titled, The Overspent
American, which is a sequel to her earlier book, The O v e r w o r k e d A m e r i c a n , S5
Schor contends that keeping up with the Joneses is passe that the Joneses
lifestyle is too quaint. We now set our sights on the lifestyles of those higher
up the organizational chart. "See-want-borrow-and-buy" is how she summarizes
today's consumers. When you earn less per hour than you did ten or twenty
years ago but you want to buy more there are two ways for workers to accom-
plish this: work longer hours and borrow more.
CONSUMER DEBT TODAY
In addition to doing many pre-bargaining surveys of specific groups of members,

we conduct an annual random sample survey of our membership which
numbers over 750,000 including retirees. One of the questions we track is
the satisfaction they have with the amount of money they are making (take
home pay) and the amount they are saving. On the pay front, 22% say they
are "Very Dissatisfied" while 35% say they are "Very Dissatisfied" with the
amount they are saving. At the other end of the spectrum, 9.8% say they are
"Very Satisfied" with take home pay while 4% say they are "Very Satisfied"
with savings/6 In fact, the latest data shows that the national savings rate
in the United States reached zero in 1998 and may have gone into the red
in the first quarter of 1999.
The results are consistent with trends across the country. Indeed recent
reports have indicated that the national savings rate is now approaching zero.
Home mortgage debt constitutes the largest portion of household debt today,
as it has since the boom in home ownership began in the late 1940s. Today,
home mortgage debt makes up just over two-thirds of household debt. The
robust economy, especially low interest rates have mitigated further increases
in home mortgage debt for those refinancing while also allowing many
households to borrow significantly more than they would have just a few
years ago.
The ratio of debt payments to family income increases dramatically as you
move down the income scale. For those making below the median income
of $25,000 debt constitutes nearly 20% of total family income. More trou-
bling, each year since 1989 that figure has gone u p . j7
Beyond home mortgage debt, two debt areas stand out for consumers: auto
loans and credit card installment payments. Since 1980 automobile credit has
soared from $350 billion to nearly $1.3 trillion. Credit card debt has also
soared, from $243 billion in 1990 to a projected $677 billion in 2000. j8 The
percent of family households with a general purpose credit card has slowly
increased from 55% in 1989 to 66% in 1995 with an estimated 75% by the
year 2000. A large percentage of credit card holders pay off their purchases
within a month and do not incur financing charges. However, those families
making less than $50,000 - over 31% - hardly ever pay off the balance,
according to the Federal Reserve System.
TWO STRATEGIC INITIATIVES
One response the IAM has come up with to address the savings problem is to
create the first union-focused mutual fund. About half of IAM members partic-
ipate in some form of tax advantaged savings plan administered by their
employer. On June 1, 1999, the IAM through State Street Global Advisors, a
nationally prominent mutual fund management company, launched an index
fund that is comprised of IAM-represented companies. The fund, called IAM
Shares, is structured to parallel the S&P 500. All investments entail some risk,
particularly those based on the equity market, but this new savings device will
provide a much needed way for IAM members to invest back into the compa-
nies they work for while diversifying risk away from an over reliance on their
specific firm's performance. IAM Shares is a publicly traded no-load mutual
fund and can be bought for a minimal initial investment and has extremely low
operating expenses (traded under the ticker symbol SIAMX). In the first nine
months since inception the IAM Shares fund tracked the S&P 500 very closely
and managed a total return of 14.2%. 19
Workers are working harder and longer hours in an attempt to maintain a
middle-class lifestyle. Easy access to credit cards and home mortgages have
made debt the easiest way to make up for shortcomings in wages. Of course,
in the long run, that does not seem like the fight formula for a good society.
As people work longer hours and fall deeper into debt it is no wonder that they
tend to see civil society and government as an impediment. Taxes take away
hard earned dollars that could be used to buy more things.
Arguments for stricter credit policy or for increasing burdens on working
people's ability to purchase a home are certainly misplaced. Indeed, job creation
in the manufacturing sector is heavily dependent on low interest rates. Juliet
Schor's somewhat utopian call for a society of "downshifters", who make the
decision to get off of the see-want-borrow-and-buy track, holds scant appeal in
the real world. I for one would not want to get up in front of a group of our
members and tell them they should stop buying new cars or tools, or that they
should think about vacationing at home this year.
The fundamental underlying problem in our economy today is the unequal
distribution of power that results in some doing extremely well while most see
stagnant or declining incomes. In his two recent books, Luxury Fever and The
Winner-Take-All Society, Robert Frank has detailed the shift that income
inequality has created at the top of society. The pay at the top of America's
largest corporations exemplify this problem. Years ago J.P. Morgan, no friend
of organized labor or working people, said that the head of an organization
should get paid no more than 20 times what the average worker gets at that
organization. In the last few years that ratio has become so bloated as to be
out of sight. The CEO of General Electric, for example, was paid $97 million
in salary, bonus, and stock options in 1998, or 2,425 times what the average
GE worker earned. The IAM has a unique approach to this problem as well.
In 1998, utilizing the employee's ownership stake in United Air Lines where
the members of the IAM and air line pilots own roughly 55% of the common
stock, the IAM proposed a new system of contingent compensation for senior
managers. Instead of a strict reliance on basing compensation on financial
performance, the IAM worked with UAL's management to adopt a new
compensation formula that includes employee satisfaction and customer service
as equal parts to financial performance. The goals of the company are aligned
more closely in this way: satisfied employees deliver better service resulting in
more customers which creates a financially healthier company.
The ownership structure of UAL allowed for this innovative measure to go
forward. In 1994 the IAM and the Airline Pilots Association (ALPA) negoti-
ated the largest employee ownership deal in history. With a six year
concessionary period during which labor costs were reduced an average of 10%,
union members purchased 55% of the common stock of UAL. The UAL trans-
action represented the zenith of employee stock ownership plans both in the
airline industry and throughout the United States. A detailed account of this
transaction, and its implications for labor relations in the airline industry can
be found elsewhere. Despite the success of UAL's innovative ownership struc-
ture, the past CEO of UAL noted that, "We have not been copied in any big
way. ''2° Ownership clearly is the ultimate in terms of corporate govemance, so
the UAL, and other airline ESOPs at Northwest Airline and TWA, deserve
closer scrutiny then what is provided here. These transactions, however, all
occurred during a period of time during which the U.S. airline industry was
faced with almost insurmountable financial difficulties. Ownership was not
sought as a strategic approach by labor but rather as a defensive tactic to save
financially troubled companies} 1
The executive compensation proposal was adopted as a strategic initiative to
show the power of ownership to employees. The UAL effort provides a glimmer
of hope that excessive executive compensation can be reined in. Still, the effects
of these wide gaps is slowly tearing the fabric of society. Spending your way
into debt is one way that working families have found to keep afloat. That
doesn't make long term economic sense. What we need is a strengthening of
the labor market institutions that govern the distribution of productivity gains.
Clearly today shareholders and executive management control those levers. In
the interest of the entire society we' should take a closer look at making the
economy work for the workforce in a fairer more efficient manner. Encouraging
collective action through representatives, of their choosing is one obvious

solution. Encouraging workers arid~their representatives to exercise: control over
capital, instead of being passive consumers, is another avenue worth pursuing
in the interest of all of society.22
In. addition to these twc~ specific initiatives the IAM is actively invol~ed in,
two other areas that utilize capitol strategies. One dicect way of harnessing
organized labor's capital resources is to become more active shareholders. With
pension: fund assets iaa joiatly trusteed pians of over $I0 billion fine IAM has
the ability to pursue good corporate governance issues that have positive effects
on the workforce white also improving the company's bottom line. One such
shareholder resolution, for General. Electric to adopt a code of condt~ct on labor
standards based on internationally recognized core rights, was voted on irr Apri~
2000. The effort attracted the support of over 120 million shares at the 2000
GE shareholder meeting. This was far. short of a majority, in fact representing
just shy of 8% of all shares cast, but the support of several key institutional
investors such. as the California Public Employees Retirement System (CalPers)
and others made the effort worthwhile. The IAM, with the support and encour-
agement of the AFL-CIO's Office of Investment, and wor~ng through the
Council of Institutionat Investors, sees shareholder activism as an important
avenue through which workers' concerns for a fairer society can be addressed.
Another area of capital strategy that is just being developed is the active
creation of tmion-friend|y ventt~re capital. Despite many serious efforts, for
example at the Be|oit Corporation, a division of Harnischfeger Industries,
successful union-initia~ed bu~outs have proven elusive. These buyouts are
different then the employee stock ownership plans (ESOPs) that were commott
in the 1980s ~ early 1990s. W i ~ equity ready to be deployed, the union
brings an entirely different perspective to the process of saving and reinvigo-
rating financially troubled companies. The jury is stil,| out as to how effective
the application of trojan-friendly capital will be on corporate performance and
the creation of good jobs that pay prevailing union wages and benefits.
THE I M P L I C A T I O N S F O R I N D U S T ~ A L RELATIONS
At the start of this paper I pointed to the decline of unions in the United States
using density as a measure of strength. Density is h a l however, a good measure
of institutional strength or of the ability of institutional innovation. As John
Dunlop has said, "The share of the nonagricultural workforce in labor organ~-
zations,.is scarcely an all-purpose measure of union strength or influenceBat the
workplace, in a community, or in the larger society.''23 The development of
capital strategies in many different forms through shareholder resolutions,
From Workplace to~Corporate Governance 199
creation, of union-friendly venture capital, the creation~ of union-specific savings

and investing vehicles, and controlli~ excessive executive compensation, as
examples has a very real potential to transform the traditional model of indus-
trial relations in the United States. Against the backdrop of changes in corporate
strategy, particularly the relentless focus on shareholder value that has swept
across from Wall Street into the executive suite of corporations, labor's capital
strategy stands in stark contrast. Rather than using the downsize and distribute
model championed by General Electric, and emulated' across major corpora-
tions, labor's interest is in long-term growth and stability. In the long-term the
retain and reinvest model may well prove more economically attractive to
employees and shareholders alike. 24
Allan Kennedy's recent book, The End o f Shareholder Value, provides a
wealth of examples showing the limitations of the present short-term focus on
quarterly returns for employees, communities and shareholders alike. 25 With
the capital strategies highlighted in this paper, unions are showing that the
traditional scope of industrial relations in the United States - wages, hours, and
working conditions - are byproducts of a larger set of variables that unions are
just now beginning to learn how to harness. Expanding from a narrow focus
on workplace governance into the broader arena of corporate governance is one
of the areas that will broadly redefine the field of industrial relations. Workers
and unions will have to assess their respective roles as adversaries to capital,
and to address new ways to deal with management. As labor comes to grip
with its power in the capital market the need to reconfigure long standing
relationships will take great prominence.
NOTES
1. Barry Hirsch and David Macpherson, Union Membership and Earnings Data Book,
BNA Publications, Washington, D.C., 2000, pp. 9-t0.
2. John T. Dunlop, Industrial Relations System: Revised Edition, Harvard University
Press, Boston, MA, 1993.
3. Bruce Kaufman, The Origins & Evolution of Industrial Relations in the United
States, Cornell University Press, Ithaca, NY, 1993, p. 99.
4. Richard Freeman and Joel Rogers, What Workers Want, Comell University Press,
Ithaca, NY, I999.
5. This definition of what corporate governance mean is from Mary O'Sullivan,
Corporate Governance in Germany, Public Policy Brief from the Jerome Levy Economics
Institute of Bard College, 1998, p. 7.
6. See for example, Labor Will Rule, Steven Fraser, The Free Press, New York, NY,
1991.
7. Mark Pertman, The Machinists, Haveard University Press, Cambridge, MA, 196 I,
p. 66.
8. Jeremy Rifkin and Randy Barber, The North Will Rise Again, Beacon Press, Boston,
MA, 1978.
9. Peter Drucker, The Unseen Revolution: How Pension Fund Socialism Came to
America, Harper and Row, New York, 1976.
10. Richard Trumka, AFL-CIO press statement, February 16, 2000.
11. Business Leaders See Best of Times As They Predict Growth, Efficiencies, Daily
Labor Report, BNA, May 7, 1999.
12. Jared Bernstein, Real median wages finally recover 1989 level, QWES, 1999, 1,
p. 1.
13. EPI Press Release, April 29, 1999.
14. Lawrence Mishel, The State of Working America 1998-1999, Cornell University
Press, Ithaca, NY, 1999, p. 2.
15. Juliet Schor, The Overspent American, Harper Collins, New York, NY, 1998; and,
The Overworked American, Basic Books, New York, NY, 1993.
16.Stephen Sleigh, AIAM Pulse of the Union, IAM Strategic Resources, 1998. Data
available upon request.
17. U.S. Statistical Abstract 1999, Government Printing Office, Washington, D.C.,
Table 802.
18. Ibid, Table 820 and 822.
19. State Street Global Advisors Bulletin, April 2000.
20. Frank Swoboda, Departing Chairman Says United Is Worth Copying, Washington
Post, p. e-I, July 7, 1999.
21. See for example, Stephen R. Sleigh, The Difficulty of Sticking Together in Tough
Times, in: Airline Labor Relations in the Global Era, ILR Press, Ithaca, NY, 1994. The
UAL transaction was conceived during a period of financial difficulty but certainly did
not face the same dire circumstances that Northwest or TWA did.
22. See, Beth Almeida, Executive Compensation: A Window on Corporate Strategy,
IAM Strategic Resources, mimeo of presentation done April 12, 2000, for an example
of the IAM's strategy on executive compensation.
23. Quoted in Marick Masters, Unions at the Crossroads, Quorum Books, Westport,
CT, 1997, p.3.
24. See, William Lazonick, Japanese Corporate Governance and Strategy, Jerome
Levy Economics Institute Public Policy Brief 48, 1998.
25. Allan Kennedy, The End of Shareholder Value, Perseus Publishing, Cambridge,
MA, 2000.
GLOBALIZATION: SOME
IMPLICATIONS FOR 21ST CENTURY
U.S. LABOR MARKETS
Kenneth McLennan
INTRODUCTION
In the 1960s exports and imports represented only about 10% of the U.S.
gross domestic product (GDP). Starting in the mid-1970s, trade expanded
rapidly and by the end of the 1990s exports and imports were equivalent to
about 24% of GDP. Economic interdependence has also accelerated as U.S.
direct investment abroad and foreign direct investment in the United States
increased more than five-fold over the past two decades. The globalization
of the U.S. economy affected the competitive position of many manufacturing
industries whose product markets had previously been sheltered from inter-
national competition. The critical issue for management and labor in these
industries was, to borrow from the title of Professor Robert Z. Lawrence's
book, "Can America Compete?"
This powerful late 20th century economic development, and how manage-
ment in many industries responded to globalization, had a profound impact on
U.S. labor markets and labor-management relations. The key to restoring the
competitive position of U.S. industry was the higher rates of productivity
growth. This changed demand and supply in U.S. capital and labor markets.
Capital and labor resources moved out of industries in which the United States
had no comparative advantage and into industries producing high value added

ISBN: 0-7623-0853-2
201
202 KENNETH McLENNAN
goods and services. To remain competitive in global markets, U.S. industry had
to invest more heavily in new technologies - technologies which ultimately
produced the so-called "new economy." The higher skill requirements of the
new economy are changing the occupational and industrial composition and the
nature of employment for most workers in the United States. These labor market
changes also had, and will continue to have, a profound effect on labor-manage-
ment relations as unions and management in the private sector are forced to
adjust to the need for greater work force flexibility.
Globalization and the related changes in technology and productivity
increased the demand for a skilled work force. Most new labor market entrants
responded by increasing their investment in education. At the same time, for
some new entrants the poor quality of education at the K-12 level placed them
at a serious disadvantage, which has been reflected in a more unequal wage
distribution.
Similarly, U.S. immigration policy continued to encourage an increase in the
proportion of poorly educated, low-skilled workers in the immigrant population
and at the same time restrict the number of highly skilled foreign professionals
who can enter the United States annually. If U.S. public policies are to respond
successfully to the future demand in U.S. labor markets, education reform at
K-12 and immigration reform should be high on the political agenda.
PRODUCTIVITY GROWTH AND OCCUPATIONAL

RESTRUCTURING
In the 1970s U.S. industry had relatively low rates of productivity improve-
ment, and most industrialized countries and the rapidly growing Asian
economies had much higher rates of manufacturing productivity increases than
U.S. manufacturers. As globalization emerged, many U.S. manufacturers were
no longer the leaders in global markets.
For businesses in many U.S. industries, such as autos, electronics, primary
metals, computers, and textiles, economic survival depended on technological
innovation, more efficient manufacturing, and much higher product quality.
While many factors influence manufacturing excellence and efficiency, global-
ization of competition was a major stimulus for the rapid adoption of new
technologies by the end of the 1990s.
During the 1980s and 1990s, U.S. industry demonstrated the ability to
compete in global markets through reliance on the market system rather than
government industrial policies. In the 1980s and the 1990s, the rate of capital
investment, especially in equipment, increased and was especially strong in the
1990s, investment in equipment varied by industry. For example, in 1980 almost
GIobalization: Some Implications for 21st Century U.S. Labor Markets 203
10% of the net stock of total non-residential equipment was in information-

processing equipment - computers and peripherals and telecommunications
equipment. By 1986 it had grown to about 16% and now has risen to about
17%. l By the end of the 1990s, annual gross investment in software exceeded
investment in computers and peripheral equipment and was almost 90% of gross
expenditures on telecommunications equipment including business machines.
These changes in the rate and pattern of capital investment are now paying
off in higher rates of productivity growth. The dispersion of innovative
ideas, practices, and processes throughout the economy brought this surge in
productivity with extraordinarily high rates in a number of manufacturing
industries.
Table 1 shows large variations in average productivity improvement in
manufacturing and in the non-farm business sector in each of the past three
decades. The manufacturing growth rate for the 1970s declined to an average
of 1.9% per annum. This poor performance was reversed in the 1980s with an
average growth rate of 3.0% which in turn escalated to an extraordinary rate
of 5.1% per annum toward the end of the 1990s.
Toward the latter part of the last century the relative importance of major
industries in the U.S. economy changed significantly. Over the past two decades,
manufacturing's share of GDP has declined, to be surpassed by services, the
most rapidly growing sector of the economy. Within manufacturing, high growth
in output was concentrated in a limited number of industries. Industries such
as electrical and electronic equipment, which includes telecommunications
equipment, and machinery, except electrical, which includes computers,
substantially increased their relative contribution to GDP. Traditional manu-
facturing industries such as primary metals and motor vehicles and
equipment, while continuing to contribute importantly to GDP, no longer
dominated U.S. manufacturing as they did in the 1950s and 1960s.
Table 1. U.S. Productivity Growth: Average Annual Rate of Growth in

Output per Hour.
Manufacturing Non-farm Business

Years (Percent) (Percent)
1973-1980 1.9 1.3

1981-1990 3.0 1.5
1991-1999 4.3 2.0
1990-1995 3.5 1.5
1996-1999 5.1 2.6
Source: U.S. Department of Labor, Bureau of Labor Statistics.

According to the 1998-2008 Occupational Outlook prepared by the Bureau

of Labor Statistics, business services employment will be the most rapidly
growing sector of the economy over the next decade. For example, for the
period 1998-2008, the average annual increase in employment in computer and
data processing services will be 187,200 jobs; for health services, including
offices of health practitioners, about 196,000. 2
The continuing growth in services is in sharp contrast to the expected loss
of manufacturing jobs. Strong productivity growth in manufacturing industries
is likely to result in a net annual loss of some 9,000 jobs through 2008 despite
overall growth in manufacturing output. Some manufacturing industries will
have relatively strong employment growth. In terms of absolute number of
manufacturing jobs created over the next decade, electronic components, aero-
space, and plastics are the leaders.
The industries with the largest job losses are concentrated in the primary
metals and fabricated metal products sector. Job losses also are projected for
some industrial machinery and equipment industries such as metalworking
machinery and engines and turbines. The motor vehicles and equipment industry
is expected to lose jobs at the rate of 5,000 jobs per year through 2008.
The growing importance of the information-technology sector in the U.S.
economy has dramatically changed the occupational outlook. According to the
U.S. Department of Commerce:
By 2006, almost half of the U.S. workforce will be employed by industries
that are either major producers or intensive users of information technology
products and services. Innovation has increased demand for high paid, "core
IT workers" (e.g. computer scientists, engineers), created new IT occupations,
changed skill requirements for some non-IT occupations, and raised minimum
skill requirements for many other jobs. Wage gaps between workers in IT
industries and all other workers will continue to widen. 3
The U.S. Department of Commerce report on the emerging digital economy
points out that employment growth in the information-technology-producing
industries (e.g. computer and semiconductor industries) outperformed employ-
ment growth in the economy for 1989-1997 and predicts these industries will
employ about six million workers by 2006. The greatest job growth will be in
information technology-using industries (e.g. business services, finance, insur-
ance and real estate, and wholesale and retail trade) which will have some 45
million workers by 2006.
In addition to the changing industrial composition of U.S. employment, there
will be a significant shift toward information-technology-related jobs ranging
from engineering and computer systems professionals to data processing
equipment repairers. On average, these occupations require a higher level of
Globalization: Some Implications for 21st Century U.S. Labor Markets 205
educational attainment than the average level for the work force and "workers
[will find] that they must be 'multi-skilled' and must commit to a lifetime of
learning and retraining in order to remain competitive in rapidly changing labor
markets. ''4
These powerful economic changes have changed the type of skills demanded
in the "new economy." The technological changes of the 1980s and 1990s have
biased labor demand toward much higher skills than in the 1960s and 1970s.
R E S P O N D I N G TO FUTURE L A B O R M A R K E T
DEVELOPMENTS
The challenges posed by globalization over the past two decades and how the
U.S. economy adapted to globalization can form the basis for how industry and
the work force responds to labor markets in the early part of this century. The
first lesson from the response to globalization is that reliance on free adjust-
ment to market forces rather than a series of market interventions through new
government programs is the most effective strategy for meeting the increasing
demand for a more skilled work force. Second, government policies can assist
in the adjustment to the technology bias toward greater skills by providing some
incentives for individuals to acquire skills and by changing current policies that
now tend to lower the average skill level of the work force. Finally, the U.S.
labor movement has generally failed to respond successfully to the changes in
the structure of U.S. industry and unions now represent a declining proportion
of private sector employees. This has had a dramatic effect on labor relations.
The implications for labor relations, changing demand for work force skills,
and the role of labor market public policies in the 21st century are discussed
briefly.
Flexibility in Labor Relations and Human Resource Management
In the public sector, labor unions continue to play a relatively important role
in determining public employees' wages and conditions of employment. While
public sector collective bargaining varies considerably among federal, state, and
local jurisdictions, since the early 1990s some 42% of public employees are
unionized. In this sector the multilateral bargaining model with unions, manage-
ment, politicians, and taxpayers is likely to continue in the 21st century as it
did in the past.
In contrast, collective bargaining in the private sector is now a much less
important institution than in the past as the rate of private sector unionization
has plummeted to a little less than 10% of the work force from a peak of about
206 KENNETI4 Mc LENNA2g
40% following World War II. h ~ private sector, glob ~:zation and techno-
logical imaovation have forced management in ma~tffacturing industries to strive
to. improve productivity and reduce the costs of producfiorL This has resu~t~t
in a decline in manufacturing employment and a rapid shift in the distribution
of employment by industry. Globalization changed the competitive environment
for many industries which in the 1970s faced little competition in U.S. markets..
The era of collective agreements with large annuat wage increases and generous
benefits was no longer feasible in the new competitive environment.
Resources in the rnanafacturi'ng sector have moved from the o l d - l , ~ indus-
tries in the primary metals, textile, and automotive sectors - sectors with
relatively high rates of unionization - to computers, electronics, telecommuni-
cations, and pharmaceuticals - sectors with little or no unionizatioa. Labor
resources also have moved to notwaanufacturing industries such as business
services. This is the major reason for the huge decline in the proportion of
private-sector workers represented by unions since the 1960s.
Unions, with few exceptions, have been unable to attract workers in the
expanding industries. Workers in these industries areola average younger and
more mobile than workers in the traditional "old-line" manufacturing indus-
tries. As a result, workers in expanding industries are less interested in the
typical wage-and-benefits packages traditionally negotiated by unions.
To increase private sector unionization in the 21st century, it will be neces-
sary for unions to take a more flexible approach to collective bargaining and
offer workers the contract provisions which are appropriate for a more mobile
work force. This means promoting defined contribution, and cash balance,
pension plans and recognizing that defined benefit plans are tess beneficial to
a more mobile work force. Unions also have to recognize that the work force
of the future is likely to be more interested in pay-for-performance and the
opportunity to participate in stock options.
The traditional union-collective-bargaining strategy of improving base pay
and negotiating for first-dollar health care plans and defined pension plans,
which tends to encourage workers to remain with their current employers and
continue to be union members, is not particularly attractive to workers in
emerging industries. In general, unions have failed to recognize that workers
in these industries are potential customers and that their needs vary by type of
industry.
Globalization has forced management to adopt more flexible human resource
strategies, not only for employee benefits but also for how work is organized in
the business and especially at the plant level. Most businesses have found that
to be competitive in global markets, corporate offices need to he non-bureau-
cratic and decision making decentralized to operational levels with increasing
worker :participation in a wide range of decisions affecting product development,

work processes, and ,quality of products and services. To restore collective
bargaining as an important labor market institution, unions need to adopt
bargaining strategies which recognize the need for variations in agreements to
enhance :plant level productivity and offer greater flexibility in the terms and
conditions of employment they negotiate .for the work force of the future.
Automatic Labor Market Adjustments
Globalization and the increased interdependence of the U.S. economy with

economies around the world have resulted in a higher standard of living for
most Americans. At the same time, workers displaced from industries due to
foreign competition and low-skilled U.S. workers forced to compete with low-
skilled workers in emerging economies have suffered economically. In
responding to globalization, U.S. industry has increased the pace of techno-
logical change which has required a more skilled work force than in the past.
Globalization and technological change biased toward greater skills has
increased the demand for skilled workers and reduced the demand for unskilled
workers which in turn has led to a more unequal distribution of wages. For
example, in the first half of the 1970s, the ratio of wages at the eightieth
percentile to wages at the twentieth percentile of the wage distribution was
about 2.1. By 1995 the ratio was about 2.6. 5
Over the past two decades, more and more developed economies are relying
on the market system to determine what and how products and services are
provided in their economies. The unique advantage of the market system is that
it adjusts automatically to changes in demand. The labor market also adjusts
automatically to the increasing demand for a more skilled work force than in
the past. Two types of adjustments are currently occurring and will play an
important role in meeting the demand for greater skills and help reverse the
wage inequality trend which has developed over the past two decades.
First, the labor market is now increasing the return to investment in human
capital. In the early 1970s, the ratio of wages of college graduates to high
school graduates was actually declining; in 1976 it was about 1.25. By 1995
the ratio had risen to a little over 1.50. The labor market has automatically
rewarded those who acquire greater skills.
The increasing return from investing in human capital development will
affect the supply of skilled workers. Currently there is a great deal of discus-
sion about the rising cost of a college education. No doubt institutions of
higher learning need to improve productivity to avoid raising tuition costs at
a higher rate than the rate of inflation, but for the future work force investing
in a college education is a good deal. The increasing return to a college educa-

tion will gradually increase the supply of skilled workers.
The second adjustment is the change in relative wage rates as employers
experiencing skill "shortages" offer greater-than-average wage increases to
workers in specific occupations in strong demand. For example, over the period
1991-1998 the average annual earnings increase for private sector production
and nonsupervisory employees in private industry was relatively stable, fluctu-
ating between a low of a little over 2% to a high of 4% in 1997. The average
annual increases for employees in computer and data processing services were
more volatile and generally significantly higher than the average for private
industry. In 1997, for example, the annual increase in earnings for this group
of workers was almost 8%. 6
The lesson learned from the extraordinary changes in the U.S. economy stim-
ulated by globalization is that flexibility in the operation of labor markets is
critical in meeting the changing demand for skills in the work force.
Public Policies and the Demand for Work Force Skills
Flexibility in labor markets is a necessary condition for meeting the demand

for a skilled work force in the 21st century but sometimes it is not a sufficient
condition. Public policies can play an important role in stimulating the supply
of skilled workers. Several types of policies are relevant to achieving a more
skilled work force.
Training and Education Policies

Over the last 40 years, the government has introduced a wide range of training
programs to meet a wide variety of goals and serve an array of constituency
groups. As these training efforts proliferated into a huge number of programs
their impact in assisting workers meet the demand for skills in the labor market
was generally modest at best.
After many years Congress successfully consolidated most of these numerous
programs with passage of The Workforce Investment Act of 1998. While this
act is now in the process of implementation and it is clearly too early to judge
its success, it appears to be a step in the right direction. The act offers
considerable flexibility for local communities in program design and includes
considerable participation of business representatives in determining the type
of training that will be offered.
The government can play a small role in attracting students into disciplines
which are critical to the training of high skilled professionals. The recent
acceleration in technological change has increased the demand for electrical
Globalization: Some Implicationsfor 21st Century U.S. Labor Markets 209
engineers, mathematicians, and computer scientists. U.S. colleges and univer-

sities have responded by increasing the number of graduates in most, but not
all, of these fields.
In the 1990s employees with at least a bachelor's degree responded to the
increasing demand for high-skilled workers. The number of scientists and
engineers increased 7.9% over the period 1993-1997. The increase in computer
and information sciences professionals was significantly higher at 12.8%. For
electrical engineers specializing in computer systems, the increase was 11.6%,
well above the average increase for all scientists and engineers.
The increase in professionals in mathematics, however, was well below
average. The rate of increase for this specialty was only 4.3%, and there were
fewer operations research engineers in 1997 than in 1993, by 4.7%. The annual
number of bachelor's degrees awarded in mathematics/computer sciences
peaked in 1986 and declined steadily through 1995. 7
Through a scholarship program targeted on the basic discipline underlying
the skill in short supply the government can accelerate the number of profes-
sionals entering an occupation critical to the work force of the early 21 st century.
Such programs should, however, be small in scope and only available for educa-
tion in a well-defined basic discipline. Changes in relative earnings should be
the major means for increasing the supply of skilled workers.
Ultimately, however, the key to increasing the supply of highly skilled
workers depends on the quality of education at the K-12 levels. Improved perfor-
mance in basic skills, especially in science and mathematics, is critical to the
future flow of graduates from higher education institutions.
States and local communities are responding to the need to improve educa-
tional performance at the K-12 level of education. Many school systems are
adopting achievement standards and reforming school governance with more
local control over the teaching curriculum. Teacher unions are accepting pay-
for-performance compensation systems and promoting quality measures for
improving education. Globalization produced foreign competition which forced
a number of U.S. industries to improve quality and restore their leadership in
global markets. Public school systems can learn from this experience. To
improve the quality of underperforming public schools, it is essential to intro-
duce some degree of competition within the public school system and between
public and private schools.
Reform of U.S. Immigration Polices

In the past, immigration was an important source of skilled labor for the United
States. On average, immigrants who entered the United States in the 1950s and
1960s had levels of education and skills similar to the average for native
workers. The skills composition of successive waves of immigrants has declined

over much of the post-World War II period. For example, relative to natives,
immigrants were about 21% more likely to be high school dropouts in 1970,
but by 1990 they were more than twice as likely to be high school dropouts, s
This relative decline in educational attainment among immigrants is an
important source of the increasing wage gap between immigrants and natives.
For immigrant males who arrived in 1965-1969, wages were 16.6% lower than
native males. By 1990 wages of this group had risen to about the same level
as native males.
But for successive waves of immigrants the wage gap has widened. For
1970-1974 arrivals, the differential in 1980 was 18.9% lower and by 1990 was
9.3% lower. The 1990 differential for 1985-1989 arrivals was 31.7% lower. In
contrast, immigrants who arrived before 1964 had significantly higher wages
in 1990 than native males because they arrived with a relatively higher level
of education. 9 Recent immigrants, on average, are clearly at a serious economic
disadvantage in an economy that now requires, and will continue to require, a
higher level of work force skills than in the past.
A major reason for the high proportion of relatively unskilled workers in the
current generation of immigrants is the increasing number now coming from
countries with relatively low levels of educational achievement. The proportion
of male immigrants from countries with well-educated workers, such as
Germany and the United Kingdom, declined dramatically from 1961-1970 to
1981-1990. In the 1960s, 5.7% of male immigrants came from Germany, but
by the 1980s only 1.3% were German. For the United Kingdom, the decline
was from 6.4% to 2.2%. For citizens of both countries the average number of
years of schooling is 14 years or more.
In contrast to immigration from countries with strong education and training
systems, the proportion of male immigrants from Mexico increased from 13.7%
in the 1960s to 22.6% in the 1980s. The average years of schooling in Mexico
is about one-half the number for Germany and the United Kingdom. Other
countries in Latin America and Africa, with low levels of educational achieve-
ment, also have become more important sources of immigrants.
The trend toward a higher proportion of low-skilled immigrants from
countries with much lower levels of academic achievement is reflected in
significantly lower average wages for immigrants compared to native workers.
Such wage differentials either did not exist, or were very small, for immigrants
entering in the 1960s. ~°
Reform of U.S. immigration policies should be high on the policy agenda
for the next Congress and the Administration. Reaching a consensus between
business and labor and other groups will be extremely difficult since any change
Globalization: Some Implicationsfor 21st Century U.S. Labor Markets 211
in policy to increase the average skill level of the immigrant population will
require modifying the current family preference system, the employment-based
visa program, and several temporary immigration programs including the H-1B
program for highly skilled foreign workers.
Reform of U.S. immigration policies to increase the skill level of the annual
flow of immigrants is perhaps the most important policy to reduce wage
inequality at the lower end of the wage distribution and at the same time increase
the net economic benefit of immigration to the country. The AFL-CIO, and
some business groups, oppose this type of immigration reform. Indeed, the AFL-
CIO has proposed providing illegal immigrants now in the country another
amnesty since unions believe they can increase membership by unionizing low-
wage workers in the service industry. Another amnesty for the five to seven
million illegal immigrants will simply encourage more illegal aliens to seek
entry to the United States, and it will depress further the wages of workers at
the bottom of the wage continuum. Such proposals may benefit union orga-
nizers, but they will do nothing for American workers, and they will fail to
maximize the net benefits from immigrant workers.
It will be difficult to reform immigration policies. Given the low rate of
unemployment and the apparent shortage of skilled workers, it should be
possible to increase the cap on the annual number of high-skilled workers who
can fill jobs under the H-IB program. The current cap on the annual number
of H-IB visas is 115,000 for 2000, 107,500 for 2001, and 65,000 for 2002.
The annual visa cap is typically set for only three years in advance. In recent
years the annual ceiling for visas is reached well before the end of the year.
For example, the Immigration and Naturalization Service (INS) announced in
March of 2001 that it was accepting no more applications for visas since it had
received sufficient applications to meet the annual limit of 115,000.
The program grants temporary resident status; visas are valid for three years
and can be renewed for an additional three years for a maximum of six years.
It is relatively easy for employers and workers with H-IB visas to receive the
three-year extension. To remain permanently in the United States, the H-1B
worker must be sponsored by an employer for one of the 140,000 annual
employment-based immigrant visas, which requires a separate and lengthy
approval process.
Several criteria define the level of skills necessary to qualify for a visa. For
example, applicants must have theoretical and practical experience in the appli-
cation of specialized knowledge, a bachelor's degree or higher or equivalent cer-
tification, and meet licensure requirements if available. Employers must attest
that H-1B workers are receiving the prevailing wage paid to native workers and
attest (in a Labor Condition Application) that working conditions for native
workers will not be affected adversely. Employers must inform bargaining agents
in union situations, or in nonunion facilities post notification prominently in the
workplace, that application for H-1B visas has been made. There also is an appeal
process and a fine and debarment system if employers violate the regulations.
If the current cap on the annual number of temporary visas made available
for hiring high-skilled workers were raised, or eliminated for a period of three
years, immigration policy would move in the direction of increasing the average
skill level of the immigration policy. Similarly, if the number of employment-
based visas available for permanent resident status were increased from its
current annual level of 140,000, perhaps to 300,000, the human capital level
of the immigrant population would increase. Ultimately, to raise the educational
level of the immigrant population, achieve a reduction in wage inequality, and
increase the net benefit from immigration to the U.S. economy, it will be neces-
sary to modify the preference for family members - children, spouses, parents,
and siblings of U.S. citizens or permanent residents who hold "green cards"
but are not citizens. Obviously, family preference should be afforded for spouses
and children. Should parents and siblings also have preference? The dominance
of family preference in the annual flow of 900,000 immigrants limits the propor-
tion of "independent" immigrants who enter the United States in response to
employer demand for employees.
The proposal to deny preference to some family relatives is not new; the U.S.
Presidential Commission on Immigration Reform has recommended that siblings
and adult children not be included in family preferenceJ ~ To migrate to the
United States, siblings and adult children of a permanent resident would be
required to enter with a temporary visa or a permanent employment-based visa.
If the primary goal of public policy is to maximize the contribution of
immigrants to economic growth, then the move toward skills-based criteria for
entry as a legal immigrant is essential. As stated in the recommendations of
the U.S. Commission on Immigration Reform, an immigration strategy more
weighted to skills will maximize immigrants' fiscal contribution, minimize any
fiscal cost of immigration, and protect economic opportunities for the small
proportion of U.S. workers who are unskilled.
If the basic program for legal permanent residents is reformed, changes in
the temporary immigration programs can permit flexibility in the supply of skills
required in labor markets in the next several decades.
CONCLUSION
The rapid changes in the structure of the U.S. economy resulting from global-
ization has shown that if business and the work force respond rapidly to
accelerating technical changes, the overall benefits in terms of economic growth

are substantial. The key to successful adjustment is flexibility in labor markets,
in collective bargaining arrangements, and in employee compensation and bene-
fits systems. Both management and labor should avoid rigidity in wages and
conditions of employment and adopt policies that accelerate adjustment to
change.
The technological changes spawned by increased competition from global-
ization are likely to continue well into the future as new phases of innovation
follow the information revolution. If management and labor fail to recognize
the continuous need for the development of a flexible and skilled work force,
the growth of GDP will be less than optimal and reduce the nation's ability to
maximize social benefits.
NOTES
1. Jeremy A. Leonard, Stimulating Manufacturing Productivity Growth Through
Investment in Computers, ER-449, Manufacturers Alliance/MAPI, June 1998.
2. See Monthly Labor Review, Bureau of Labor Statistics, U.S. Department of
Labor, November 1999, for series of five articles which include detailed employment
projections.
3. The Emerging Digital Economy I1, Executive Summary, U.S. Department of
Commerce, June 1999.
4. Ibid., p. 42.
5. Marvin H. Kosters, Wage Levels and Inequality, American Enterprise Institute,
1998.
6. Trends in Labor Costs: 1997-1997, Manufacturers Alliance/MAPI, February
1998, p. 8.
7. National Science Foundation, www.nsf.gov.
8. George J. Borjas, The Economics of Immigration, Journal of Economic Literature,
XXXII, December 1994, p. 1677.
9. Ibid., p. 1674.
10. George J. Borjas, Heaven's Door: Immigration Policy and the American Economy,
Princeton University Press, pp. 42-44, 1996.
11. U.S. Commission on Immigration Reform, Becoming an American: Immigration
and Immigration Policy, Washington, D.C., GPO, 1997.

7fdf3ac0-5b5c-4f0c-9959-d9a65d4de092

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

7fdf3ac0-5b5c-4f0c-9959-d9a65d4de092

Uploaded by

Copyright:

Available Formats

LIST OF CONTRIBUTORS

Joseph R. Blasi School of Management and Labor Relations,

Mollie Bowers Merrick School of Business,

Robert Bruno Institute of Labor and Industrial Relations,

Victor G. Devinatz Department Management and Quantitative

Anita Jose Department of Economics and Management,

Bruce E. Kaufman Department of Economics,

Maya K. Kroumova School of Management,

Douglas L. Kruse School of Management and Labor Relations,

David Lewin Anderson Graduate School of Man'agement,

Patrice M. Mareschal Graduate Department of Public Policy and

E. Patrick McDermott Center for Conflict Resolution,

Kenneth McLennan Manufacturers Alliance/MAPI, USA

Daniel J. B. Mitchell Anderson Graduate School of Management &

Ruth Obar Department of Economics and Management,

Stephen R. Sleigh International Association of Machinists and

James C. Sesil School of Management and Labor Relations,

Roland Zullo Institute of Labor and Industrial Relations,

Volume 11 of Advances in Industrial and Labor Relations (AILR) contains

E. Patrick McDermott, Anita Jose, Ruth Obar,

Advances in Industrial and Labor Relations, Volume 11, pages 1--40.

Mediation is a dispute resolution process in which a neutral third party,

accepted institutionalized process? The ultimate acceptability of a dispute

The Equal Employment Opportunity Commission (EEOC) was established as

A. The EEOC Strategies To Timely Process Charges

B. The E E O C Mediation Program

Beginning February 1991, the EEOC introduced an ADR pilot mediation

is not disclosed to EEOC investigators. Investigators are instructed to cease any

III. THE RESEARCH CATALYST

The EEOC decided in 1999 that it needed a comprehensive evaluation of its

IV. LITERATURE REVIEW

A review of the various studies that evaluated the effectiveness of ADR

Fairness/procedural justice requires that there is opportunity for assistance

Gordon, in his study of the issues of justice in a union-management grievance

The primary method of data collection is a survey of the participants, who

A. Profile of the Sample

Table 1. Comparison of the Characteristics Between Our Sample and the

EEOC Mediation Cases Research Sample

B. Procedural Elements and EEOC Mediation

a. Explanation, Scheduling, and Voice

Table 2. Participant Satisfaction with the Explanation, Scheduling, and

Charging Parties Respondents

Prior to my attendance at this mediation session today,

Table 3. Participants' Satisfaction with the Mediator.

Charging Parties Respondents

The mediator understood my needs. 4.30 4.31

As illustrated in this section on procedural elements, both the charging parties

Participant satisfaction with the distributive elements of mediation was measured

Table 4. Participant Satisfaction with the Distributive Elements o f

Charging Parties Respondents

Most of the options developed during the

D. The Influence o f Other Variables on Participant Satisfaction

We analyzed participant responses to see whether they varied based on medi-

Table 5. Responses of Participants Whose Claims Were Not Resolved in

Participant Total Mean Strongly Neither Agree/Strongly

Charging Parties 488 2.60 48.6% 22.1% 29.3%

Table 6. Participant Satisfaction Based on Mediator Type.

Statements Charging Parties Respondents

Explanation, Scheduling and Voice

Adequate Explanation 4.26 4.20 4.23 4.14

Mediator understood 4.32 4.23 4.31 4.30

Development of 3.99 3.83 3.99 4.02

Table 7. Participant Satisfaction Based on Representation.

Statements Charging Parties Respondents