Professional Documents
Culture Documents
Stress KAP
Stress KAP
Steven E. Kaplan
Arizona State University
)*564)+6
The majority of public accounting firms now offer flexible work arrange-
ments to their professional employees. Presumably these arrangements
help accommodate employee needs to manage work and family de-
mands, while also improving job satisfaction and retention. The ability
of flexible work arrangements to achieve these goals has received little
attention. The current paper addresses this issue by reporting the results
of a survey of CPAs working under a flexible work arrangement and a
similar group of CPAs working under a standard arrangement but who
appear to be plausible candidates for a flexible work arrangement. The
survey elicited information about several key employment variables:
job-related stressors (e.g., role conflict, role ambiguity, and role over-
load), burnout tendencies (e.g., emotional exhaustion, reduced personal
accomplishment, and depersonalization) and behavioral job outcomes
(e.g., job satisfaction and turnover intentions).
Results show that CPAs on flexible work arrangements report
higher job satisfaction and lower turnover intentions than those on a
standard work arrangement. CPAs on flexible work arrangements gen-
erally have lower levels of burnout and stressors, though the reduced
personal accomplishment burnout dimension may be conditioned upon
whether the CPA has a mentor. Finally, for professionals switching to
a flexible work arrangement, respondents indicated a significant im-
provement in job satisfaction and turnover intentions as well as some
decline in burnout and stressors.
We gratefully acknowledge the assistance of the AICPA Women and Family Issues
Executive Committee, participating firms and professionals, and the University of Miami
McLamore Summer Award and the SBA Summer Research Support. Comments from
Janet Samuels, Jeff Cohen, and Louise Single on earlier drafts of this paper are also
appreciated.
2 Almer and Kaplan
INTRODUCTION
The structure of the American workforce is changing and today
fewer than 10 percent of employees are part of a “traditional” family in
which the male partner is the primary provider (Rose 1998). While
families may need two incomes, they may also require more flexible
work arrangements in order to manage family demands when both
parents are working. Attempting to be responsive to their employees’
family needs, while also trying to mitigate costly turnover and short-
ages of CPAs (Hiltebeitel et al. 2000), the vast majority of public ac-
counting firms now make flexible work arrangements available to their
employees (AICPA 1997). In general, excessive work-family stress has
been linked to high turnover rates in public accounting firms (Hooks
et al. 1997).
Despite the availability of flexible work arrangements, little evi-
dence is available to assess the extent to which such arrangements
are associated with key employee variables such as job stressors, burn-
out (e.g., a negative psychological response to interpersonal stressors),
job satisfaction, and turnover in the public accounting profession.
Research has not explored, for example, whether job-related stressors
might be affected by changing from a standard to a flexible work ar-
rangement. Further, empirical evidence is scant on the extent to which
flexible work arrangements are associated with burnout, job satisfac-
tion, and turnover intentions.
The purpose of the current paper is to provide evidence on the
extent to which employee’s use of flexible work arrangements are as-
sociated with key employment-related variables in the public account-
ing profession. This evidence contributes, in part, by extending the
work of Fogarty et al. (2000). Their study provides evidence on the
relationships among job stessors, burnout, and job outcomes among
members of the AICPA from five randomly selected states. However,
participants’ work arrangements were not reported or examined by
Fogarty et al. (2000). Our study examines whether flexible work ar-
rangements represent an antecedent to job stressors. Further, our
study also provides evidence on whether flexible work arrangements
are associated with burnout and job outcomes.
A questionnaire was designed and administered to public account-
ing professionals working under a flexible work arrangement as well
as to demographically similar professionals working under a standard
arrangement. The professionals working under the standard arrange-
ment were selected because of personal circumstances (i.e., recently
returned from maternity/paternity leave, with school-aged children,
etc.) that made them plausible potential candidates for a flexible work
arrangement. The questionnaire elicited information on the employ-
ees’ backgrounds, job stressors, burnout, job satisfaction, and turn-
over intentions. These variables are of interest to public accounting
firms attempting to assess the effects of flexible work arrangements
The Effects of Flexible Work Arrangements 3
and have been the focus of several previous studies of public account-
ing employees (Bamber et al. 1989; Collins and Killough 1992; Fogarty
et al. 2000; Rebele and Michaels 1990; Senatra 1980). Responses of
these two highly stressed groups of employees are compared under
the two work arrangements. A supplemental analysis compares flex-
ible respondents’ perceptions about flexible work arrangements to their
perceptions of their previous standard work arrangement. Responses
that take a “backward look” may be affected by hindsight, but they
provide useful information to firms about whether flex professionals
believe their situation is improved by having adopted a flexible work
arrangement.
The rest of the paper is organized as follows. First is a discussion
on the background of flexible work arrangements. Next, a research
question and hypotheses are presented. Third, the methodology and
results are provided. Last is a discussion of the results and limita-
tions of the research.
BACKGROUND
Flexible work arrangements began to gain popularity as an em-
ployee benefit in the early 1970s (Sladek 1995; Sullivan and Lussier
1995). Such arrangements can take a variety of forms from working
fewer hours or days, to working the same number of hours but with
greater flexibility over when the hours will be worked. Regardless of
the specific format, flexible work arrangements inject changes in the
professional’s work environment as the professional is no longer work-
ing during the same hours and/or as many hours as is the norm in
their office (Hooks and Higgs 2000). Advocates of flexible work ar-
rangements contend that the use of a flexible work arrangement by
public accounting employees should improve their job satisfaction and
strengthen their intentions to remain with the firm (Mattis 1990). How-
ever, research that investigates these claims is scarce.
Public accounting firms have recognized the need to offer flexible
work arrangements to their employees. For example, the results from
the AICPA’s Survey on Women’s Status and Work/Family Issues in
Public Accounting found that among firms with 20 or more AICPA
members, 76 percent offer flexible work times to their professionals
and offer part-time work arrangements (AICPA 1997; Doucet and Hooks
1999). The survey also reports that since an earlier AICPA study in
1993, the prevalence of flexible work arrangements has increased sig-
nificantly and over this same time period, the percentage of new moth-
ers returning to work on a part-time basis increased from 27 percent
to 38 percent. The increasing availability of flexible work arrangements
to public accounting professionals is consistent with the results from
a Coopers & Lybrand survey that found quality of life issues were
more important than monetary rewards among new entrants (Rose
1998).
4 Almer and Kaplan
1 Per conversations with Human Resources contacts at firms participating in this study,
because the firms only offer general guidelines for flexible work arrangements, dis-
cussions occur between the professional and the firm in order to establish the spe-
cifics of the arrangement.
The Effects of Flexible Work Arrangements 5
Burnout
The term “burnout” was first used by Freudenberger (1974), and
has been extensively researched in the occupational health and ap-
plied psychology literatures (Lee and Ashforth 1993). As discussed in
this literature, burnout refers to a negative psychological response to
interpersonal stressors and contains three separate dimensions: emo-
tional exhaustion, reduced personal accomplishment, and deperson-
alization (Cordes and Dougherty 1993). Emotional exhaustion is
“characterized by a lack of energy and a feeling that one’s emotional
resources are used up” (Cordes and Dougherty 1993, 623). Reduced
personal accomplishment involves low motivation and self-esteem.
Depersonalization refers to detachment and an emotional callousness
toward others (such as clients). Sanders (1998) reports that burnout
6 Almer and Kaplan
METHOD
A survey was distributed to public accountants employed among
five participating firms. Firm participation was enlisted through the
support of the Women and Family Issues Executive Committee (WFIEC)
of the AICPA. In offering their support, the Committee indicated that
little attention has been given to non-Big 5 firms and strongly recom-
mended that the study focus on this underresearched population. Thus,
none of the five participating firms are among the Big 5, though all
are among the 50 largest firms.3 Each firm agreeing to participate named
a liaison who was initially responsible for identifying potential partici-
pants. Two different groups of participants were sought from each par-
ticipating firm. One group, referred to as the flex group, was comprised
of employees working under a flexible work arrangement. The second
group was comprised of employees who were plausible potential
2 The work-related variables were values, expectations, satisfaction, commitment, and
intentions. The personal life variables were values, leisure preferences, spouse, and
family considerations.
3 The participating firms included one national firm, two large multi-office regional
firms, and two mid-sized multi-office regional firms.
8 Almer and Kaplan
a flexible work arrangement and did not switch to a flexible work ar-
rangement: forty-five flex group respondents stated on their question-
naires that they were hired under a flexible work arrangement; and
for 28 flex group respondents, it appears that they were hired under a
flexible work arrangement since their time on a flexible work arrange-
ment equaled the time employed by the firm. This subgroup is re-
ferred to as flexible permanent respondents. The remaining 100 flex
group respondents switched from a standard to a flexible work ar-
rangement. This subgroup is referred to as flex switch respondents.
For the potential flex group, 282 surveys were distributed and 140
useable surveys were returned, representing a 50 percent response
rate. Table 1 provides demographic information on respondents.
TABLE 1
Demographics by Work Arrangement
Potential
Flex Flex Significance
n = 173 n = 140 of Difference
Level in firm: Partner .5% 1%
Senior manager/
manager 27% 23%
Senior/supervisor 28% 35%
Staff 44% 41%
Administrative .5% 0% n.s.
Mentor: No 39% 36%
Yes 61% 64% n.s.
Years with current firm: 5.8 4.7 n.s.
Years as a CPA: 9.0 6.8 .004
Primary
Functional Area: Audit/Accounting 24% 43%
Tax 56% 26%
Consulting 14% 17%
Administrative 3% 7%
Multiple Areas 3% 7% .001
Gender: Female 94% 76%
Male 6% 24% .001
Spouse description:
Homemaker 3% 5%
Works at home 4% 1%
Works FT outside home 91% 85%
Works PT outside home 2% 9% .035
Number of children under
18 living at home: 1.5 1.1 n.s.
10 Almer and Kaplan
Stressors
Rizzo et al. (1970) developed measures for role conflict and role
ambiguity applicable to those in managerial positions or those posi-
tions where the individual enjoys some degree of autonomy. These
measures have been used extensively and shown to exhibit accept-
able psychometric properties (see Jackson and Schuler 1985). For the
current study, a shortened version of the Rizzo et al. (1970) question-
naire was used containing three items to measure role conflict and
five items to measure role ambiguity. Role overload was measured
using a three-item scale from Beehr et al. (1976). For each measure,
higher scores indicate more stress. Fogarty et al. (2000) took a similar
approach to measure these three stressors, though their study did
not use subsets of these measures.
5 The proportion of respondents indicating they had no mentor was somewhat higher
than that reported in other studies of public accounting professionals (e.g., Viator
2001). One possible explanation for the lower level of mentors is that this research
focuses on flex professionals and candidates for a flexible work arrangement, whereas
other research focuses on public accounting professionals in general.
6 The measures reported in the current study are a subset of the entire seven-page
survey sent to participants. In accordance with the intent of the AICPA WFIEC Re-
search Assistance Grant under which this study was conducted, a wider range of
issues was examined in the survey than is reported in the current paper.
TABLE 2
Questions Used to Measure Model Variables and Means
All Respondents
Mean Cronbach’s
Measure Questions (Scale) (Std. Dev.) Alpha
Role Stressors
Role Conflict (1) I have to work under vague directives or orders. 3.16 (1.78)
(Rizzo et al. 1970) (2) I work under incompatible policies and guidelines. 2.41 (1.52)
(3) I receive an assignment without the manpower
to complete it. 2.92 (1.76)
(1 = strongly disagree to 7 = strongly agree) 2.84 (1.36) .73
Role Ambiguity (1) I have just the right amount of work to do.a 4.12 (1.67)
(Rizzo et al. 1970) (2) I know that I have divided my time properly.a 3.37 (1.60)
The Effects of Flexible Work Arrangements
Mean Cronbach’s
Measure Questions (Scale) (Std. Dev.) Alpha
Job Outcomes
Job Satisfaction
(McNichols et al. 1978) (1) Please indicate which of the following statements
describes how well you like your job.
(1 = I hate it to 4 = I am indifferent to it to 7 = I love it) 5.54 (1.15)
(2) Please indicate which of the following statements
describes how well you like your job.
(1 = Never satisfied to 4 = Satisfied half the time to
7 = Always satisfied) 5.29 (1.06)
(3) Please indicate how you think you compare with other
people in your firm.
The Effects of Flexible Work Arrangements
Burnout
The three burnout dimensions were measured using subsets of
scales drawn from the Maslach Burnout Inventory (MBI) (Maslach and
Jackson 1981). Each of the three burnout dimensions was measured
using three items from the MBI. For each of the dimensions, higher
scores indicate greater feelings of burnout.
RESULTS
Descriptive Statistics
Table 2 presents descriptive statistics for the questions used to
measure each construct under study based upon current responses
for all professionals. Table 2 also presents Cronbach’s alpha, an index
of internal consistency of a measure, to assess reliability. With the
exception of role overload, the Cronbach’s alpha for each of the mea-
sures utilized in the study exceed 0.70, an acceptable standard for
the reliability of measures (Nunnally 1967).
Data Analysis
Multivariate analysis of covariance (MANCOVA) was used to test
the hypotheses and research question. MANCOVA is appropriate for
statistical analysis in this study because multiple items were used to
assess each dependent variable (Johnson and Wichern 1998). Fur-
ther, MANCOVA can accommodate the inclusion of background vari-
ables that might be associated with the dependent measure.7
Two steps were taken to determine which, if any, background vari-
ables should be included in the statistical analysis. First, the correla-
tion between each background variable and each dependent measure
was determined. Significant correlations (p < .05) were found between
having a mentor (e.g., yes or no), organizational level (e.g., partner,
senior manager/manager, senior/supervisor, and staff) gender, and
at least one of the dependent variables. Thus, mentor, organizational
7 We rejected structural equation modeling because two of the key independent vari-
ables in our study are manifest variables (e.g., flexible work arrangement and men-
tor status). The inclusion of manifest categorical variables requires the use of a much
larger sample size than we have. Specifically, when manifest categorical variables
are present in the model, a polychoric matrix must be determined (Loehlin 1998).
The Effects of Flexible Work Arrangements 15
Stressors
Research question 1 explores whether employees under a flexible
work arrangement will report lower role stress than employees under
a standard work arrangement. The study contained three constructs
related to role stress. These are role conflict, role ambiguity, and role
overload. For each construct a separate MANCOVA was conducted.
The results are presented in Table 3. As shown in Panels A, C, and D
of Table 3, work arrangement is significantly associated with role con-
flict, but not with role ambiguity or role overload. Table 3, Panel B
contains descriptive statistics regarding the effect of work arrange-
ments on role conflict. As shown, role conflict is significantly lower
under the flexible work arrangement for each of the three questions
used to measure role conflict.
As shown in Table 3, each of the control variables was signifi-
cantly associated with certain stressor variables. Mentoring relation-
ships significantly influenced role conflict and role ambiguity whereas
organizational level significantly impacted all three stressor variables.
Those with no mentor reported higher levels of role conflict and role
ambiguity. Regarding level, partners and staff reported lower levels of
role conflict than seniors and managers. Role ambiguity was inversely
related to level. No discernable pattern was detected for the effect of
level on role overload. Gender was significantly associated with role
overload—males had higher levels of role overload than females.
Burnout
Hypothesis 1 predicts that employees under a flexible work ar-
rangement will report less burnout than employees under a standard
work arrangement. The study contained three dimensions of burnout:
8 Type of flexible work arrangement was also examined to determine if it affected the
behavioral outcomes included in the study. HR contacts at participating firms indi-
cated that the flexible work arrangements varied from person to person, to match
the needs of the local office and each professional. Typically they were structured to
vary on the following dimensions: hours and days worked during busy season(s) and
nonbusy season(s), work location during normal business hours, and if the profes-
sional was “on-call” during normal business hours. Each of these variables was cap-
tured in our instrument and none were significant covariates in our analyses.
16
TABLE 3
Effect of Work Arrangement on Role Stress
Panel A: MANCOVA on Role Conflicta
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .97 3 2.99 .0312 Lower when flex.
Covariates
Mentor .94 3 5.99 .0006 Higher when no mentor.
Level .90 12 2.82 .0009 Partners and staff lower than
managers and seniors.
Gender .99 3 .78 .5017 No difference.
Panel B: Adjusted Cell Means for Effect of Work Arrangement on Role Conflicta
Flex Work Standard Work
Dependent Measure Arrangement Arrangement p-value
Question 1a 3.45 3.96 .0148
Question 2a 2.74 3.10 .0402
Question 3a 3.26 3.80 .0095
Panel C: MANCOVA on Role Ambiguitya
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .97 5 1.63 .1494 No difference.
Covariates
Mentor .92 5 5.02 .0002 Higher when no mentor.
Level .87 20 1.99 .0060 Inversely related.
Gender .97 5 1.76 .1210 No difference.
Job Outcomes
Hypotheses 3 and 4 predict that job outcomes among employees
under a flexible work arrangement will be more favorable than for
employees under a standard work arrangement. The study contained
two constructs related to job outcomes. These are job satisfaction and
turnover intentions. Similar to the above analysis, a separate
MANCOVA was conducted for each job outcome construct. The re-
sults from this analysis are presented in Table 5.
Panels A and C of Table 5 show that work arrangement is signifi-
cantly associated with job satisfaction and turnover intentions. Table
5, Panel B shows that the means for two of the three job satisfaction
questions were significantly higher under a flexible work arrangement
TABLE 4
Effect of Work Arrangement on Burnout
Panel A: MANCOVA on Emotional Exhaustiona
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .95 3 5.21 .0016 Lower when flex.
Covariates
Mentor .97 3 2.13 .0956 Lower when mentor.
Level .93 12 1.72 .0582 Higher at higher levels.
Gender .99 3 .17 .9152 No difference.
Panel B: Adjusted Cell Means for Effect of Work Arrangement on Emotional Exhaustiona
Flex Work Standard Work
Dependent Measure Arrangement Arrangement p-value
The Effects of Flexible Work Arrangements
Panel D: Adjusted Cell Means for Effect of Work Arrangement and Mentor on Reduced
Personal Accomplishmenta
Mentor No Mentor
Dependent Measure Flex Standard p-value Flex Standard p-value
Question 1a 3.07 2.71 .0414 2.89 3.24 .1125
Question 2a 3.17 3.02 .4776 3.23 3.85 .0187
Question 3a 3.13 2.71 .0192 2.85 3.33 .0309
Panel E: MANCOVA on Depersonalizationa
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .96 3 3.34 .0196 Lower when flex.
Covariates
Mentor .95 3 5.66 .0009 Lower when mentor.
Level .92 12 2.12 .0137 Higher at higher levels.
Gender .96 3 3.67 .0126 Males higher than females.
Panel F: Adjusted Cell Means for Effect of Work Arrangement on Depersonalizationa
Flex Work Standard Work
Dependent Measure Arrangement Arrangement p-value
Question 1a 2.56 2.75 .1699
Question 2a 2.76 3.21 .0022
Question 3a 2.41 2.86 .0080
Panel D: Adjusted Cell Means for Effect of Work Arrangement on Intentions to Staya
Flex Work Standard Work
Dependent Measure Arrangement Arrangement p-value
Question 1a 4.88 4.72 .3681
Question 2a 4.40 3.98 .0354
Supplemental Analysis
The above analyses utilize two demographically similar groups of
professionals who are experiencing similar life stressors (i.e., return-
ing from maternity/paternity leave, school-aged children, etc.), and
include a variety of covariates. However, it could be argued that de-
tected differences are not attributable to differences in work arrange-
ments, but to some other unspecified difference between the two
9 This analysis was also performed including job satisfaction as an additional covariate.
Work arrangement continued to be insignificant.
24
TABLE 6
Alternative Analysis—Effect of Work Arrangement on Burnout
Panel A: MANCOVA on Emotional Exhaustiona
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .97 3 2.86 .0371 Lower when flex.
Covariates
Mentor .98 3 1.42 .2356 No difference.
Level .94 12 1.31 .2076 No difference.
Gender .99 3 .11 .9485 No difference.
Role Conflictb .94 3 6.43 .0003 Higher when higher role conflict.
Role Ambiguityb .98 3 1.83 .1404 No difference.
Role Overloadb .82 3 22.06 .0001 Higher when higher role overload.
Panel B: MANCOVA on Reduced Personal Accomplishmenta
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .99 3 .80 .4907 No difference.
Work Arrangement If mentor, higher when flex. If no
× Mentor .96 3 3.69 .0124 mentor, lower when flex.
Covariates
Mentor .99 3 .74 .5267 No difference.
Level .93 12 1.83 .0393 Lower at higher levels.
Gender .99 3 .55 .6422 No difference.
Role Conflictb .98 3 1.15 .3257 No difference.
Role Ambiguityb .94 3 5.96 .0006 Higher when higher role ambiguity.
Role Overloadb .98 3 1.18 .3156 No difference.
TABLE 7
Alternative Analysis—Effect of Work Arrangement on Job Outcomes
Panel A: MANCOVA on Job Satisfactiona
Wilks’
Factor Lambda DF F p-value Pattern of Results
Work Arrangement .93 3 7.16 .0001 Higher when flex.
Covariates
Mentor .93 3 6.34 .0004 Higher when mentor.
Level .96 12 .80 .6432 No difference.
Gender .99 3 .44 .7202 No difference.
Role Conflictb .99 3 .93 .4260 No difference.
Role Ambiguityb .94 3 5.34 .0013 Higher when lower role ambiguity.
Role Overloadb .98 3 1.41 .2392 No difference.
Burnout—Emotional
Exhaustionb .91 3 9.44 .0001 Higher when lower emotional exhaustion.
Burnout—Reduced
Personal .95 3 5.11 .0018 Higher when lower reduced personal
Accomplishmentb accomplishment.
Burnout—
Depersonalizationb .93 3 7.51 .0001 Higher when lower depersonalization.
DISCUSSION
Before discussing the results of the study, several limitations should
be noted. First, subsets of the originally validated measures were used
to measure role stressors, burnout, and job satisfaction. Accordingly,
the reduced measures may have failed to capture the relevant dimen-
sions of the constructs, and the results reported here may not be com-
parable with other research that used the complete validated measures.
Second, the study relied upon the responses from a questionnaire dis-
tributed to professionals identified by each participating firm. Reli-
ance upon the firm to identify participants does not represent a random
process and limits the generalizability of the results. In particular,
professionals who may have adopted a flexible work arrangement, then
still chose to leave their current firm, were excluded. The response
rate was quite high, but there is always the possibility that those re-
sponding may systematically differ from those who did not respond.
Of final note, the focus of our study was on sampling professionals at
large non-Big 5 firms facing both work and family responsibilities that
are likely to create a high degree of stress. The effect of flexible work
arrangements may differ at larger or smaller CPA firms.
The study found one of the role stressors—role conflict—was sig-
nificantly lower among professionals under a flexible work arrange-
ment. This finding is important because it indicates that professionals
The Effects of Flexible Work Arrangements 29
TABLE 8
Supplemental Analysis on the Effects of Work Arrangement
Comparison of Flex Switch Noncurrent and Current Responses
REFERENCES
American Institute of Certified Public Accountants (AICPA). 1993. Sur-
vey on Women’s Status and Work/Family Issues in Public Accounting.
New York, NY: AICPA.
———. 1997. Survey on Women’s Status and Work/Family Issues in Public
Accounting. New York, NY: AICPA.
Almer, E. D., and S. E. Kaplan. 2000. Myths and realities of flexible
work arrangements. The CPA Journal LXX (4): 14–19.
Baltes, B. B., T. E. Briggs, J. W. Huff, J. A. Wright, and G. A. Neuman.
1999. Flexible and compressed workweek schedules: A meta-analy-
sis of their effects on work-related criteria. Journal of Applied Psy-
chology 84 (4): 496–513.
Bamber, M. E., D. Snowball, and R. M. Tubbs. 1989. Audit structure
and its relation to role conflict and role ambiguity: An empirical ex-
amination. The Accounting Review 64 (2): 285–299.
Beehr, T. A., J. T. Walsh, and T. D. Taber. 1976. Relationship of stress
to individually and organizationally valued states: Higher order needs
as a moderator. Journal of Applied Psychology 61 (1): 41–47.
Cohen, J. R., and L. E. Single. 2001. An examination of the perceived
impact of flexible work arrangements on professional opportunities
in public accounting. Journal of Business Ethics 32 (4): 317–318
Collins, K. M., and L. N. Killough. 1992. An empirical examination of
stress in public accounting. Accounting, Organizations and Society
17 (6): 535–547.
———. 1993. Stress and departures from the public accounting profes-
sion: A study of gender differences. Accounting Horizons 7 (1): 29–38.
Cordes, C. L., and T. W. Dougherty. 1993. A review and an integration
of research on job burnout. Academy of Management Review 18: 621–
656.
Dirsmith, M. W., and M. A. Covaleski. 1985. Informal communications,
nonformal communications and mentoring in public accounting firms.
Accounting, Organizations and Society 10 (2): 149–169.
Doucet, M. S., and K. L. Hooks. 1999. Toward an equal future. Journal
of Accountancy 187 (6): 71–76.
Fogarty, T. J., J. Singh, G. K. Rhoads, and R. K. Moore. 2000. Anteced-
ents and consequences of burnout in accounting: Beyond the role
stress model. Behavioral Research in Accounting 12: 31–67.
Freudenberger, H. 1974. Staff burn-out. Journal of Social Issues 30 (1):
159–165.
Hiltebeitel, K. M., B. A. Leauby, and J. M. Larkin. 2000. Job satisfac-
tion among entry-level accountants. The CPA Journal LXX (5): 76–78.
Hooks, K. L., P. Thomas, and W. Stout. 1997. Retention of women in
public accounting. Advances in Accounting 15: 17–48.
———, and J. L. Higgs. 2000. Changing environment and productivity
relationships in professional services firms. Working paper, Florida
Atlantic University.
Jackson, S., and R. Schuler. 1985. A meta-analysis and conceptual
critique of research on role ambiguity and role conflict in work set-
tings. Organizational Behavior and Human Performance 36 (1): 16–
78.
Johnson, R. A., and D. W. Wichern. 1998. Applied Multivariate Statisti-
cal Analysis. Englewood Cliffs, NJ: Prentice Hall.
The Effects of Flexible Work Arrangements 33