You are on page 1of 7

Journal of APJ)lltd Psyd)olQly Copyright 1990 by the Ame~ PsycholosiClll Asaoeialiocl, In<:.

1990. Vol. 75. "No. 4, 3'1-377 0021.9010/90/$00.7 s

Employee Reactions to Contextual and Session Components


of Performance Appraisal
William F. Giles and Kevin W Mossholder
Department of Management, Auburn University

A neglected area of performance appraisal research concerns the context within which the ap-
praisal process oc.curs. For a sample of exempt employees, measures were developed that assessed
system components of the appraisal context. The contribution of these variables (complexity, im-
plementation, and follow-up) to the prediction of 2 measures of employee reactions to performance
appraisal (review session satisfaction and appraisal system satisfaction) was compared with the
contribution of a more frequently studied set of variables-supervisory behaviors in the review
session. The relationship of a salary linkage variable to the 2 outcome criteria also was assessed.
The supervisory session variables were related to session satisfaction, and the system contextual
variables were primarily related to system satisfaction. Salary linlc.age was associated with system
satisfaction.

Performance appraisal research has concentrated on a num- weakly but positively associated with ratings of job competence
ber ofareas, for instance, the appraisal instrument, counseling but was not related to an overall performance rating. Bernardin
and development of appraisees, rater training programs, and and Beatty (1984, pp. 268-270) described a measure, labeled
cognitive processes (see Banks & Murphy, 1985; Napier & "'trust in the appraisal process," that assessed employees' per-
Latham, 1986). In terms of the adoption of performance ap- ceptions of the accuracy and fairness in the assessment of their
praisal innovations by organizations, the impact of this re- performance. In a field investigation, Bernardin and Beatty
search has been relatively limited in proportiQn to the amount found that lack of trust in the appraisal process was associated
of effort expended (Banks & Murphy, 1985). In comparison to with inflated appraisal ratings. Cleveland, Murphy, and Wil-
other performance appraisal topics, the context in which perfor- liams (1989) developed performance appraisal usage factors
mance appraisal is conducted has received only limited study. and studied the linkages between these factors and organiza-
This is particularly interesting because the environment in tional structure constructs.
which the performance appraisal process occurs has been desig- Of the three contextual componems specified in Landy and
nated as a source of considerable influence on the appraisal Farr's (J 980) model, rating purpose has received more sustained
process (Ilgen & Feldman, 1983; Landy & Farr, J980; Lawter, research attention (e.g, Dobbins, Cardy, & Truxillo, 1988; Farb
Mohrman, & Resnick, 1984; Mohrman & Lawler, 1981; Zam- & Werbel, 1986) than organization or position characteristics.
muto, London, & Rowland, 1982). Researchers may perceive rating purpose as more tilceJy to af-
As Carroll andSchneier(l982, p. 6) noted, a variety ofcontex- fect appraisal outcomes than the other two types ofcontextual
tual characteristics may influence performance appraisal.
variables, which are only distally linked to performance ap-
Landy and Farr (1980) designated as contextual factors those
praisal. Cleveland et al.'s (1989) results support the notion that it
that are not explicitly related to the nature of the rater, ratee, or
is difficult to link conceptually distant contextual variables
rating instrument but that may be considered part of the con-
directly with appraisal outcomes. Specifically, of 40 possible
text in which the rating occurs. Their process model of perfor-
com:lations between 10 organizational characteristics (e.g., cen-
mance rating contains three contextual components: rating
tralization, size) and four performance appraisal uses, they
purpose, organization characteristics, and position characteris-
tics. Mohrman and Lawler (1981) specified as contextual such found only 4 that reached significance at the .05 level.
factors as the job characteristics and functional areas of ap- A class of contextual variables that appears more likely than
praisees, the nature of interpersonal relationships, and the organization and position characteristics to affect appraisal
structure, climate, and culture of the organization. outcomes has been suggested by recent research. In a study
A few studies have examined the potential impact of such involving both manager and employee samples, Mount (1983)
contextual variables within the performance appraisal process. factor analyzed items pertaining to various aspects of the ap-
Svetlik, Prien, and Barrett (1964) found that job difficulty was praisal process (e.g., review session, instrument, system). In ad-
dition to factors generally dealing with characteristics of the
appraisal instrument and general satisfaction, he derived fac-
We thank Mick Mount and Mike Buckley for their helpful com- tors labeled Company Policy/Procedures and Communication
ments on an earlier draft of this article. of the Program for the employee sample, and Company Policy,
Correspondence concerning thisanicle should be addressed to Wil- Appraisal Procedures, and Communication of the Program for
liam F. Giles, Depanment of Management, Auburn University, Au· the managerial sample. Unfortunately, Mount did not examine
burn Universit); Alabama 36849. potential correlations between these factors and appraisal out-
371
372 WILLIAM F. GILES AND KEVIN W MOSSHOLDER

comes. In another stu~ Mount (1984) did investigate relation- utility, fairness, accuracy, improvement) have been used to as-
ships between appraisal system characteristics (in addition to sess individuals' reactions to various performance appraisal
other appraisal-related variables) and reactions to appraisal, variables, the one most frequently used appears to be satisfac-
but the use of single-item measures for the independent and tion (e.g. Dorfman, Stephan, & Loveland, J986; Mount, 1984;
dependent variables limits the usefulness of his results. Nemerotf & WexJey, J979; Prince & Lawler, 1986; Russell &
Mount's (1983, 1984) research is pertinent to the present Goode, 1988; Silverman & Wexley, 1984). One advantage of
study because it introduced a potentially important class of using satisfaction as a measure of individual reactions is that it
contextual variables for performance appraisal research. A con- appears to assess both fairness cognitions and simple affect
textual component encompassing system characteristics offers (Organ, 1988), thus affording a broader indicator of individ-
a conceptual midpoint between assessment of macro-level orga- uals' reaction to appraisal than more specific, cognitively ori-
nization characteristics (e.g., size and structure) and micro-level ented criteria (e.g, perceived utility).
features of the performance appraisal process (i.e., rater, ratce, The third purpose of this study was to assess the extent to
instrument). It is surprising that researchers interested in con- which system conte1'tual variables, as compared with supervi-
1extual variables have typically no1 considered characteristics sory session variables, were related to employee satisfaction
of the appraisal system per se (hereinafter referred to as system with appraisal. Supervisory session variables are those behav-
comextual variables) because system characteristics establish iors exhibited by the supervisor in the review session. These
the immediate framework within which appraisal sessions oc- variables were used for comparison purposes because previous
cur, evaluations are given, and appraisal information is pro- research assessing employee reactions to performance ap-
cessed. Therefore, from an applied perspective, system contex-
praisal components has focused considerable attention on char-
tual variables ought to have a greater effect on appraisal out-
acteristics of the appraisal discussion while neglecting charac-
comes than contextual variables having no inherent connection
teristics of the appraisal system (Mount, 1983}. These studies
with the appraisal process (e.g., organization size, decentraliza-
indicated that certain supervisory session variables-goal set-
tion, job difficulty).
ting, encouragement of employee participation, and supervi-
The present study had three interrelated purposes. Because
sory support or criticism-are consistently related to employee
of the scarcity of specific empirical studies pertaining to sys-
reactions (see Cederblom, 1982, for a review of research in these
tem contextual variables (Mount, 1983), the first purpose was
areas). Responses to our preliminary survey suggested that an-
to extend the development of measures that assess system con-
other session-related variable-salary linkage-should be con-
textual aspects of performance appraisal. Consequently, the do-
main within which pertinent items were written was framed in
sidered as well.
terms ofsystemic features of performance appraisal. When for- Because satisfaction criteria needed to be appropriate for the
mulating such items, we emphasized appraisal problems result- two sets of independent variables (supervisory session and sys-
ing from the situation in which an appraisal system is used tem contextual), satisfaction with the review session and satis-
(Carroll & Schneier.1982, pp. 18-20). For example, items were faction with the appraisal system were used as dependent vari-
developed to reflect situations in which there is a failure to ables. Each has been used in past research (e.g., Prince &
develop or apply policies reinforcing the basic importance of Lawler, 1986), but much Jess attention has been focused on
the appraisal system to the organization. Factors involved in the satisfaction with the appraisal system (Mount, 1984). This
design and implementation of appraisal systems (e.g., orienui- omission is potentially serious because the review session is
tion, monitoring, and evaluation) also were considered (Carroll only one component of the overall appraisal system. Moreover,
& Schneier, I 982, pp. 217-248; Mohrman & Lawler, 198 J). the perceptions that employees bave of the overall performance
To further aid in domain definition, we conducted a prelimi- appraisal system, regardless of the system's specific features,
nary survey with open-ended questions pertaining to the aP- will have much to do with its success (Mohrman & Lawler,
praisal process. Feedback from this survey helped us target the 1981).
specific system contextual characteristics perceived by employ- Because goal setting, participation.and support or criticism
ees as relevant. Subsequently, items were written that pertained have been clearly shown to be related to employee satisfaction
to specific system aspects of the appraisal process, for ex.ample, with the review session, we included them in this study to deter-
introduction, training, uses (e.g., promotion and training deci- mine whether system contextual and supervisory session vari-
sions), time demands, complexity, and management support. ables would exhibit different relationships with appraisal sys-
A second purpose of this study was to investigate the rela- tem satisfaction and session satisfaction. We hypothesized that
tionships between system contextual variables and employees' supervisory session variables would correlate more strongly
reactions to performance appraisal. Reaction measures were with employee session satisfaction than would system contex-
used as dependent variables because they typically indicate tual variables. This is logical because a supervisor's review ses-
overall appraisal system viability better than more narrow psy- sion behaviors ought to more directly affect subordinates' satis-
chometric indices, such as leniency, halo, and discriminability faction with review sessions. In contrast, we expected system
(Bernardin & Beatty, 1984, pp. 179-183). Moreover, regardless contextual variables to correlate more strongly with appr.tisal
of its psychometric soundness, an appraisal system wiJI ulti- system satisfaction than would supervisory session variables.
mately be unsuccessful if it is not accepted and supported by its The context in which perfunnance appraisal is conducted per-
users (Carroll & Schneier, 1982, p. 218). meates the whole appraisal process, whereas the review session
Although several types of dependent variables (e.g, perceived constitutes only a portion of the total system.
CONTEXTUAL AND SESSION COMPONENTS OF APPRAISAL 373

Method From a factor analysis (principal axes with varim.ax rotation) of22
relevant items in the first section of the questionnaire, four factors
Subjects were extracted: Fuoctionality, Complexity, Implementation, and Fol-
low-Up. Of the original sample of S I0 questionnaires, all those with
The original subject pool consisted of 510 exempt employees from a
complete data on the 22 ,items (n = 3 l 8) were used in the factor analysis.
national textile company. Data werccol.lected from multiple geographi-
The Functionality factor contained seven items (e.g., "The performance
cal locations. Ofthese employees, I02 were used as subjects in the main
appraisal system helps an employee and his/her manager to recognize
data analyses. In this subsample, 95% of the employees were men, their
a.nd discuss the employee'sstrong and weak areas;; alpha was.89. The
average tenure with the company was 8. I years, and their mean age was
Complexity factor contained four items (e.g., "The performance ap-
39.5 years. Eighty-three percent had obtained a bachelor's degree or
praisal system is too complex for the average salaried employee to
more advanced academic credentials. Subjects held lower and midlevel
management positions. The functional area most heavily represented understand"); alpha was .83. The Implementation factor contained
was production management (n == 53); other areas (e.g., finance/ac- four items (e.g., "The meetings and training sessions which were used to
counting, personnel, industrial engineering) were represented to a introduce the performance appraisal system were not very effective"
lesser extent (reverse scored}); alpha was.73. The Follow-Up factor contained three
items {e.g., "The performance appraisal system needs to be monitored
better to be sure that supervisors discuss appraisals with their subordi-
Procedure nates" (reverse scored)); alpha was .68.
We first developed a preliminary open-ended survey, which was dis- The system contextual factor of Functionality was not used in subS(>
tributed to a random sample of200 employees. Of these, 122 returned quent analyses because it appeared to be operating as a surrogate sys-
usablequestionnairesfor a return rateof61 %. Questions on this prelim- tem satisfaction measure. Support for this interpretation comes from a
inary form asked employees to describe what they perceived to be the number ofsources. First, although the concepts of utility and satisfac-
major positive aspects, limitations, and needed improvements of the tion are conceptually distinct. they may be so closely linked in daily
performance appraisal system. The company had implemented a cor- job experiences and subsequent perceptions that clear distinctions are
porate-wide performan<:e appraisal system approximately 2 years be. not made between them. The high correlation between Functionality
fore the data in this study were collected. and system satisfaction (r = .73) supports this notion. Second, unlike
For the final questionnaire, items pertaining to contextual issues the other system contextual factors {Complexity, lmplementation, and
were based on appraisal systems considerations (Carroll &. Schneier. Follow-Up), the Functionality factor did not focus on a specific system
1982; Mohrman & Lawler, 1981) and feedback from the preliminary facet; rather, it pertained to the utility of performance appraisal in
survey. Items concerning supervisory session behaviors were based on relation to a large number ofsupervisory functions (e.g., training, coun-
pertinent measures used in past research on appraisal sessions (Burke, seling, communication, development of employee potential). Third,
Weitzel, & Weir, 1978; Dorfman et al~ 1986; Greller, 1978; Nemeroff& utility has been used as a dependent variable, along with appraisal
Wellley, 1979; Prince &. Lawler, 1986) and feedback from the prelimi- satisfaction measures, in a number of studies (Greller, 1978; Prince &
nary survey. Lawler, 1986; Silverman & Wexley, 1984), indicating that it may more
The final questionnaire was sent to 805 exempt employees through logically be regarded as a genera] outcome variable than as a specific
the company mail; 510 surveys containing complete or partial data on system contextual variable. Fourth, Mount (1983) found a general Satis-
items measuring the study variables were received, a return rateof63%. faction factor in his factor analysis of appraisal items, suggesting that
From this sample, 102 employees met the following two criteria: ta) the presence of a general factor may be a recurring phenomenon.
They provided complete data on the study variables, and (b) they had From a factor analysis (principal axes with varimax rotation) of 17
participated in at least one performance appraisal review in the role of relevant items in the second section of the questionnaire, four factors
a
a subordinate and none in the role of supervisor. We used the latter were extracted: Participation. Goal Setting, Criticism, and Salary Link-
age. Of the original sample of 510 questionnaires, an those with com-
criterion because the focus of this study was on employee (not supervi·
so1' reactions to performance appraisal factors. We felt that it might be plete data on thel7 items(n ~ 373)wercuscd in the factor analysis. The
difficult for an employee who bad participated in appraisal sessions as Participation factor contained four items (e.g., "In the most recent per-
both an appraiser and an appraisee to resJ>Ond to the questionnaire formance appraisal session, the supervisor invited my participation");
solely from the ~ive of an appraisee, even if instructed to do so. alpha was .82. The Goal Setting factor contained four items (e.g., "After
my last performance appraisal session. there was no doubt in my mind
as to what goals I should accomplish"); alpha was .75. The Criticism
Measures footor contained five items (e.g., "The supervisor was quite critical of
Employees used a 6-J>Oint Likertscale ranging from strongly disagree my performance in the last appraisal session;; alpha was .69. The
to strongly agree to resJ>Ond to the items. The factors derived from Salary Linkage factor contained two items (e.g. "In the appraisal ses-
these items were scaled so that higher scores represented greater per- sion, my supervisor discussed how my salary adjustment is aifeaed by
ceived presence of a factor. my appraisal'1; alpha was .64.
The questionnaire was divided into two sections with separate in- Because no general factor emerged, all four factors were used in later
structions requiring resJ>Ondents to answer the questions from differ- analyses. Moreover, the Participation, Goal Setting, and Criticism fac-
ent perspectives. In the first section, which primarily contained items tors corresJ>Onded quite closely to those that Cederblom (1982) identi-
penaining to system aspects of performance appraisal, respondents fied as the three major process factors in the review session.
were simply asked to indicate their extent of agreement or disagree- The Complexity, Implementation, and Follow-Up factors contained
ment with each item. In the second section, which generally contained only items pertaining to system contextual aspects of performarn:e
items pertaining to supervisors' review session behaviors, resJ>Ondents appraisal and were therefore labeled system contextual factors. Like-
were instructed (a) to answer these items only if they had participated wise, the Participation, Goal Setting, and Criticism factors contained
as an appraisee in at least one appraisal session and (b) to respond to only items pertaining to supervisory bebavio~ in the appraisal session
items pertaining to the review session in terms of the last session in and were therefore categorized as supervisory session factors. The two-
which they had participated in the role of a subordinate. item Salary Linkage factor contained one supervisory session item
374 WILLIAM F. GILES AND KEVIN W. MOSSHOLDER

(previously listed) and one system contextual item (''My salary adjust- faction was .56. Dipboye and de Pontbriand (1981), Mount
ments have been closely linked to the appraisals l have received"). The (1984), and RusselJ and Goode (1988) obtained conelations
latter item was included in the second rather than the first section of between these variables of .74, .56, and .21 , respectively.
the questionnaire because it was one of a few contextual items that Simple c.orrelations between each of the supervisory session,
called for a response solely from an appraisee~ perspective. Respon- system contextual, and salary linkage variables and each of the
dents were instructed to answer items in the second section from an
appraisee's point of view.
two dependent variables were significant (p <.OJ); criticism
We developed five other measures for this s~ Session and system and complexity were negatively correlated with the dependent
satisfaction were used as criteria, and job satisfaction, satisfaction with variables. The most recent performance appraisal rating did
supervisor, and last appraisal rating served as controls. The Session not correlate with any of the other 11 variables. Previous re-
Satisfaction measure contained three items (e.g., ul felt quite satisfied searchers have obtained differing results regarding the influ-
with my last appraisal scssionj; alpha was .89. The system satisfaction ence of prior performance ratings on employee reactions to
measure contained three items(e.g., ~in general, I feel the company has various appraisal-related dimensions (see Dorfman et al., I 986;
an excellent periOnnance appraisal system"); alpha was .81. The job Landy & Farr, 1980; Russell & Goode, 1988). An element con-
satisfa1.tion measure contained three items (e.g., "I'am quite satisfied tributing to these differences may be the extent to which ratings
with my job"); alpha was. 78. The supervisor satisfaction measure con- are communicated to employees, either verbally or in terms of
tained four items (e.g., "My supervisor is a very competent person"); pay increases and promotions.
alpha was .79. Performance appraisal data from each employee's last The results of the hierarchical regression analyses are pre-
review were gathered from personnel files. The company used a perfor- sented in Table 2. The three control variables were entered as
mance appraisal form on which employees were rated (on a 9-point
scale) on up to 20 job dimensions. A global performance rating was the first set of variables and the supervisory session, system
derived for each employee by computing the mean of the supervisor's contextual, and salary linkage variables were added as the sec-
performance ratings on those job dimensions judged by the supervisor ond set. Within each of these two sets, individual variables were
as relevant to the employee~ job. entered according to stepwise criteria.
Scores for the other 11 measures (system contextual, supervisory For session satisfaction, the three control variables contrib-
session, salary linJcaae, and satisfaction variables) were obtained by uted an initial R 2 of .46, and the remaining seven variables
computing the mean of items constituting each measure. contributed an incremental R 2 of.31 (final R 2 =. 77). All super-
visory session variables and one system contextual variable (fol-
Design and Analysis low-up) were related to session satisfaction.
For system satisfaction, the three control variables contrib-
The system contextual, supervisory session, and salary linkage fac- uted an initial R 2 of .23, and the remaining seven variables
tors were used as independent variables in two multiple regression
contributed an additional R2 of. 34 toward the cumulative R2 of
analyses, one with session satisfaction serving as the dependent vari-
able and the other with system satisfaction in this role. Job satisfaction,
.'57. Two system contextual variables (complexity and imple-
satisfaction with supervisor, and the global performance rating served mentation) and the salary Jinkagc variable correlated with
as control variables in both analyses. The contributions of the seven system satisfaction, but none of the supervisory session vari-
independent variables to session satisfaction and system satisfaction ables did.
were determined only after the contributions made by the three control To test the hypothesis that, as a set, supervisory session vari-
variables were: assessed. This procedure minimized the possibility that ables were more related to session satisfaction th.an were the
the system contextual, supervisory session, and salary linkage vari- . system contextual variables, we entered each set ofindependent
ables were simply serving as surrogates of more global attitudes. variables in a regression equation (with session satisfaction as
Job satisfaction was used to control for variations in system or ses- the dependent variable) after the control variables and other set
sion satisfaction that might be due to an employee's overall job satisfac- of independent variables were entered. When the supervisory
tion, because it is possible for an individuals overall reaction to affect session set was entered last, its incremental R1 was .12 (p <
his or her reactions to specific satisfaction facets (Mount, 1984). The .001 ). When the set ofsystem contextual variables was entered
most recent appraisal ratings were used bc<::ause an employee's rating last, its incremental R 2 was .04 ( p < .0 I). It thus appears that the
may have an overriding effect on his or her reactions to the appraisal
session or system (Dorfman et al, 1986; Russell & Goode, 1988). Fi- supervisory session variables, as a set, were more related to
nal!~ supervisor satisfaction was used because there is evidence that
session satisfaction than were the system contextual variables.
session satisfaction is influenced by this variable (Russell & Goode, To determine if the set of system contextual variables was
1988). more related to system satisfaction than was the supervisory
session set, we performed asimilar analysis with system satisfac-
Results tion as the dependent variable. When the set ofsystem contex-
tual variables was entered last, its in<:remeotal R 2 was .14 ( p <
Means, standard deviations, and an intcrcorrelation matrix .00 I). When the supervisory session variable set was entered
of the variables are shown in Tuble 1. The average intercorrela· last, its incremental R2 was .01 (nSJ. These results indicate that
tion among the 10 independent variables was .27. There were a the system contextual variatSles as a set correlated more
number of relatively high intercorrelations between system con- strongly with system satisfaction than did the supervisory ses-
textual and supervisory session variables. This is not surprising sion variables.
because various performance appraisal process models have Because session satisfaction and system satisfaction are both
posited conne.ctions between the context of appraisal and the reaction outcomes to performance appraisal, an auxiliary analy-
review session (Carroll & Schneier, 1982, p. 218; Decotiis & sis was performed to assess the influence of possible overlap
Petit, 1978; Landy & Farr, 1980; Mohrman & Lawler, 1981 ). between these two criteria on the results reported in Table 2.
The correlation between session satisfaction and system satis- Specifically. for each regression equation involving a particular
CONTEXTUAL AND SESSION COMPONENTS OF APPRAISAL 375

Table 1
Means, Standard Deviations. and lntercorrelation Matrix ofStudy Variables
Variable M SD 2 3 4 5 6 7 8 9 10 11 12

1. Participation 4.37 .95 -.45• .60• .23• .52• -. 15 .46• ,39• .19 .5 1* .74• .45•
2. Criticism 2.32 .62 -.47* .04 -.53* .21• -.46• - .20• - .04 -.40• - .59• -.36•
3. Goal setting 4.26 .91 .35* .so• -.21• .43• .40• .12 57• .75* .46•
4. Salary linkage 3.10 1.20 .23• -.09 .21• .13 .03 .16 .31 • .43•
5. Implementation 3.77 .96 -.32• .56* .37• .08 .30* .56* .55*
6. Complexity 3.32 1.09 -. 17 -.05 .04 .00 - .26• -.43•
7. Follow-up 3.52 1.13 .23• .07 .Jo• .62• ,31•
8. Job satisfaction 4.90 .92 .03 .17 .41* .45*
9. Performance 6.30 .86 .16 .11 .04
10. Satisfaction with supervisor 4.80 .81 .61* .26*
I I . Session satisfaction 4.28 1.08 .56•
12. System satisfaction 3.74 1.03

Nole. Ne 102.
• p < .05 (nondirectional).

dependent variable, the other dependent variable was entered future research on contextual influences in the performance
into the equation at the first step along with the three control appraisal process. We identified three system-related contex-
variables. In this manner, any common variance between the tual measures, two of which (Complexity and Implementation)
two dependent variables would be partialcd out. Results from . correlated with employees' satisfaction with the appraisal sys-
the auxiliary analysis are presented in Table 3. A comparison of tem. The remaining measure (Follow-Up) was related to session
Tables 2 and 3 reveals only two changes from significance to satisfaction rather than to system satisfaction. The fact that
nonsignificance (or vice versa) for the beta weights and R 2 in- Follow-Up items referred to the session may account for I.his
crements. For system satisfaction, the beta weight for follow-up result even though the focus of this measure was on whether or
changed from - .13 (nSJ to - . l 9 ( p < .05), and the incremental not the session occurred (as opposed to what transpired in the
R 2 for supervisor satisfaction changed from .03 ( p < .OS) to session).
.01 (ns). As hypothesized, the set of system contextual variables was
more strongly related to employees' appraisal system satisfac-
Discussion tion than was the set of supervisory session variables. Con-
The results of this study suggest that additional attention
should be given to characteristics of the appraisal system in
Table 3
Auxiliary Hierarchical Regression Results for
Table 2 Session and System Satisfaction
Hierarchical Regression Results for Session
and System Satisfaction Session System
satisfaction satisfaction
Session System
satisfaction satisfaction Variable p• llR 2 p• t.R2

Variable fJ• ~2 {J' 4R1 Control


System satisfaction .15• .17...
Control Session satisfaction .3 1* .31 ...
Job satisfaction .10 .09-• .27... .20••• Job satisfaction .06 .02• .24- .Q6••
Performance - .02 .00 .00 .00 Performance -.02 .00 .01 .oo
Supervisor satisfaction .19** .37*** .04 .03* Supervisor satisfaction .is•• .37••• -.02 .01
Supervisory session Supervisory session
Participation .30... .17*** .09 .00 Participation .29*** .06- .00 .00
Goal setting .23•• .04*** -.0 1 .00 Goal settin11. .23.. .03••• -.08 .oo
Criticism -.17* .01* -.13 .01 Criticism -.1s• .01• -.07 .00
Contextual System contextual
Complexity -.09 .01 -.29*** .11••• Complexity -.05 .00 -.26"* .w-•
Implementation - .07 .00 .24• .05•• Implementation - .10 .00 .26** .03*
Follow-up .21•• .08*.. - . 13 .01 Follow-up .23••• .12- -.19* .02
Salary linkage .08 .00 .33*•• .10••• Salary linkage .03 .00 .30*0 .01·-
Note. N - 102. The three control variables were entered on the first Note. N = 102. The control variables were entered on the first step. For
step. For session satisfaction, R 1 = .77, adjusted R 2 = .76. For system session satisfaction, R 2 = •79, adjusted R2 = .77. For system satisfaction,
satisfaction, R 2 = .57, adjusted R 2 = .Sl. R 2 . . .58, adjusted R1 = .53.
•Beta weights are controlled for all other variables. • Beta weights are controlled for all other variables.
*p< .OS. **p<.01. -p<.001. • p <.OS. •• p < .01. ••• p < .001.
376 WILLIAM F. GILES ANO KEVIN W. MOSSHOLDER

versely, the set of supervisory session variables (as compared terms of the variables examined in this study, organizations
with the set of system contextual variables) correlated more might consider a number of possible interventions. In regard to
highly with employees' session satisfaction. The salary linkage system complexity, it appears desirable to involve future users of
variable related to system satisfaction. an appraisal system in its development. lt is likely that partici-
The finding that salary linkage was positively related to sys- pants would consider "user·friendliness" issues (Wiersma &
tem satisfaction is of particular interest. Traditionally, it has Latham, 1986), knowing that they will have to Jive with the
been recommended that salary discussions be separated from system in the future. Employee participation in system develop-
nonevaluative appraisal feedback because of the presumed neg- ment would be particularly vital in organizational units judged
ative repercussions (Meyer, Kay, & French, 1965; Sashkin, roost likely to perceive the system as overly complex (e.g., those
1981). Recent research (Prince & Lawler, 1986), however, has with educational levels below the average ofsystem users).
found no evidence that discussion of salary produces negative When implementing an appraisal system, organizations
consequences in relation to the review session. In fact, Prince should attend to raters' ability and willingness to give accurate
and Lawler (1986) found that discussion ofsalary had a positive ratings (Banks & Murphy, 1985; Longenecker et al, 1987). To
impact on employees' reactions to appraisal sessions under cer- improve appraisers' skills, training programs could cover a
tain conditions. The present study extends these findings by range of appraiser-related behavior (e.g., skills in the appraisal
demonstrating that linking salary to performance appraisal session) in addition to improving managers' capabilities to rate
leads to higher employee satisfaction with the performance ap- employees accurately (Sashkin, 1981 ). Managers must be ap-
p1aisal system. This suggests that system credibility is likely to prised of the crucial role performance appraisal serves in rela·
erode if subordinates perceive that organizational reinforce- tion to other organizational functions. Emphasizing top man-
mentS are not related to the appraisals they receive. agement's support ofthe appraisal program also ought to have a
The salary linkage variable may have correlated with system positive influence on participant motivation (Beer, Rub, Daw-
rather than session satisfaction because it primarily reflects the son, McCaa, & Kavanagh, 1978).
fairness of outcomes, whereas the participation, goal setting, Organizations need to monitor their appraisal systems to in-
and criticism variables reflect the fairness of procedures sure that employees have review sessions with their managers at
(Greenberg, 1986). Most employees would probably hold their specified times. Simply ltaving subordinates sign off on ap-
supervisors responsible for using fair procedures in the review praisal forms to indicate that an appraisal session bas occurred
session, but would probably perceive their salary adjustments is not adequate because employees might feel pressured to do
to be influenced by a number of factors (see Longenecker, Sims, so. Managers and subordinates might also be queried periodi-
& Gioia, 1987; Mohrman & Lawler, 1981) in addition to their cally about their appraisal session experiences (Carroll &
supervisors' input (e.g., the organization's compensation proce- Schneier, 1982, p. 246; Sashkin, 198 I) to detect when subordi-
dures, market pressures, opinions of higher level managers). nates are either not receiving appraisal sessions or are receiving
One potential limitation of this research is that all variables ones that are ineffectual.
(except performance) were assessed with the same question- It is important that employees understand the connections
naire, raising the possibility that common method variance between their performances, appraisals, and salary adjust-
may have affected the findings. Two points should be made ments. These linkages, referred to by Greenberg (1986) as dis-
regarding this issue. First, the results for the system contextual, tributive justice factors, are often poorly articulated in organiza-
supervisory session, and salary linkage variables were obtained tions. Moreover, the role of other influential factors (e.g, labor
only after job satisfaction and satisfaction with supervisor were market pressures) on the salary adjustment process should be
entered into the regression models. To the extent that any com· made explicit (Mohrman & Lawler, 1981). Organizations also
mon method variance was associated with either of these two ought to continuously monitor the relationship between perfor-
global measures, it would have been minimized in terms of its mance ratings and salary adjustments.
influence on variables entered later in the models. Research on system contextual variables needs to be per-
Second, a major purpose of the study was to determine the formed in other organizations to determine the generalizability
relative degree to which supervisory session and system contex· of the present findings. Moreover, as additional research is con-
tual variables were related to session and system satisfaction. ducted on contextual factors and knowledge of their possible
By definition (see Campbell & Fiske, 1959), method variance antecedents and oonsequences accumulates, path analytic tech-
should similarly inflate relationships among all variables as- niques can be used to investigate causal linkages. Another re-
sessed in the same questionnaire, so that the difference be-- search area could be the development of variables related to
tween relationships ought to be largely unaffecte.d. specific aspects of the relationship between performance ap-
These findings suggest that companies ought to use different praisal and salary (e.g., historx organizational policy).
types of interventions depending on whether they want to im- The results of this study mesh well with practicality issues
prove employees' satisfaction with performance appraisal ses- that have been raised in regard to performance appraisal re-
sions or with the appraisal system. For the former, attention search (Banks & Murphy, 1985; Napier & Latham, 1986;
should be directed toward eliciting certain types of managerial Wiersma & Latham, 1986). Organizations are more likely to be
behaviors in the appraisal session (i.e., more partfoipation and capable of influencing system contextual variables than, for
goal setting, less criticism). For the latter, interventions targeted example, raters' cognitive processes (see Banks & Murphy,
primarily at the system level (e.g, appraisal system design, pol- 1985). Removirtg excessive complexity from appraisal systems,
icy, and implementation) may offer more promise. properly introducing them into organizations, ensuring that su-
To provide an optimum context for performance appraisal in pervisors conduct appraisal sessions, and linking salary adjust-
CONTEXTUAL AND SESSION COMPONENTS OF APPRAISAL 377
ments with appraisal results are all actions within most organi- Landy, F. J., & Farr, J. L. (1980). Performance rating. Psychological
zations' capabilities. If such actions are neglected, it is likely Bulletin, 87, 72-107.
that the impact of the appraisal system will be substantially Lawler, E. E., Mohrman, A. M., & Resnick, S. M. (1984). Performance
reduced. appraisal revisited. Organizational Dynamics, 13(1), 20-35.
Longenecker, C. 0., Sims, H. P.,&. Gioia, D. A. (L987). Behind the
mask: The politics of employee appraisal. Academy ofManagement
References Execiitive, I, 183-193.
Meyer, H . H., Ka); E., & French, J. R. P. (1965). Split roles in perfor-
Banks, C. G., & Murphy, K. R. (1985). Toward narrowing the research- mance appraisal. Harvard Business Revie)I! 43, 123- 129.
practice gap in performance appraisal Personnel Psychology. 38, Mohrman, A. M., & Lawler, E. E. (198 L, August). Improving the contex-
335-345. tual fit of appraisal systems. Paper presented at the 89tb Annual
Beer, M., Ruh, R., Dawson, J. A., McCaa, 8. 8 ., &. Kavanagh, M. J. Convention of the American Psychological Association, Los An-
(1978). A performance manaaement system: Research, desian. in- geles.
troduction, and evaluation. Personnel Psychology, 31. 505- 535. Mount, M. K. (1983). Comparisons of managerial and employee satis-
Bernardin, H.J., &. Beatty, R. W. (I 9S4). Petformance appraisal: Assess· faction with a performance appraisal system. Personnel Psychology.
ing human behavior at work. Boston: Kent. 16. 99-110.
Burke, R. J., Weitzel, W., &. Weir, T. (L978). Characteristics of effective Mount, M. K. (L 984). Satisfaction with a performance appraisal sys-
employee performance review and development interviews: Replica· tem and appraisal discussion. Journal ofOccupaJional Behavior, 5.
tion and extension. Personnel Psychology, 3J. 903-919. 271-279.
Campbell, D. T., & Fiske, n W. (1959). Convergent and discriminant Napier, N. K., & Latham, G. P. (1986). Outcome expectancies of people
validation by the multitrait-multimcthod matrix. Psychological who conduct performance appraisals. Personnel Psychology. 39,
Bulletin, 56, 81- 105. 827- 837.
Carroll, S. J., & Schneier, C. E. (1982). Performance appraisal and review Nemeroff, W. F., & Wexley, K. N. (1979). An exploration of the relation-
systems. Glenview, JL: Scott, Foresman. ships between performance feedback interview characteristics and
Cederblom, D. (1982). The petfonnance appraisal interview: A review, interview outcomes as perceived by managers and subordinates.
implications, and suggestions. Academy of Management Review, 7. Journal ofOccupaJional Psychology, 52. 25-34.
219-227. Organ, D. W. (1988). A restatement of the satisfaction-performance
Cleveland, J. N. Murphy, K. R., & Williams, R. E. (1989). Multiple uses hypothesis. Journal of Management, 14, 541-551.
of performance appraisal: Prevalence and correlates. Journal ofAp- Prince, J. B., & Lawler, E. E. (1986). Does salary discussion hurt the
plied Psychology, 74. 130-135. developmental performance appraisal? Organizational Behavior
Decotiis, T. A., & Petit, A. (1978). The performance appraisal process: Qlld Human Decision Processes. 37, 357-375.
A model and some testable propositions. Academy of Managemenl Russdl, J.S.,& Goode, D. L. (1988). An analysis of managers' reactions
Review, 3, 635-646. to their own performance appraisal feedback . Journal of Applied
Dipboye, R. L. & de Pontbriand, R. (1981). Correlates of employee Psychology, 73, 63-67.
reactions to performance appraisals and appraisal systems. Journal Sashkin, M. (1981 ). Appraising appraisal: Ten lessons from research
of Applied Psychology. 66, 248-251. for practice. Organizational Dynamics. 9(3), 37- 50.
Dobbins, G. H., Cardy, R. L., & Truxillo, n M. (1988). The effects of Silverman, S. B. & Wexley, K. N. (1984). Reaction of employees to
purpose of appraisal and individual differences in stereotypes of performance appraisal interviews as a function of their participa-
women on sex differences in performance ratings: A laboratory and tion in rating scale development. Personnel Psychology, 37, 703-710.
field study. Journal ofApplied Psycholog>J 73, 551-558. Svetlik, 8., Prien, E~ & Barrett, G. (1964). Relationships between job
Dorfman, P. W, Stephan, W G., & Loveland, J. (1986). Performance difficulty, employee's attitude toward bis job, and supervisory rat-
appraisal behaviors: Supervisor perceptions and subordinate reac- ings of employee effectiveness. Journal of Applied Psychology, 48,
tions. Personnel Psychology, 39, 579-597. 320-324.
Fam, J., & Werbel, J. 0. (1986). Effects of purpose of the appraisal and Wiersma, U., & Latham, G. P. (1986). The practicality of behavioral
expectation of validation on self-appraisal leniency. Journal of Ap- observation scales, behavioral ellpectation scales, and trait scales.
plied Psychology, 71. 527-529. Personnel Psyclwlogy. 39, 619-628.
Greenberg. J. (1986). Determinants of perceived fairness of perfor· Zammuto, R. F~ London, M~ & Rowland, K. M. (1982). Organization
mance evaluations. Journal ofApplied Psychology. 7J, 340-342. and rater differences in performance appraisal. Personnel Psychol-
Greller, M. M . (1978). The nature ofsubordinate part~ipation in the ogy, 35. 643-658.
appraisal interview. Academy ofManagement Journal. 21. 646-658.
Jlgen, 0. R., & Feldman. J. M. (1983). Performance appraisal: A pro- Received October 3, 1988
cess focus. ln B. Staw & L. l. Cummings(Eds~. Research in organiza- Revision received November 21, 1989
rional beha1>ior (Vol. S, pp. 141-190). Greenwich, CT: JAl Press. Accepted December 22, 1989 •

You might also like