University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln
, &$+$,1$. /1+$,1 "2*152!*(" 1(-,0 , &$+$,1$. /1+$,1
Goal Seing and Performance Evaluation: An
Aributional Analysis
Dennis L. Dosse
University of Nebraska at Omaha
Carl I. Greenberg
University of Nebraska at Omaha
-**-41'(0 ,# ##(1(-, *4-/)0 1 '8.0#(&(1 *"-++-,02,*$#2+ , &$+$,1% ".2!
/1-%1'$ 20(,$00#+(,(01/ 1(-, , &$+$,1 ,#.$/ 1(-,0-++-,0 , &$+$,1
"($,"$0 ,#2 ,1(1 1(3$$1'-#0-++-,0 ,#1'$ 1/ 1$&(" , &$+$,1-*("5-++-,0
7(0/1("*$(0!/-2&'11-5-2%-/%/$$ ,#-.$, ""$00!51'$ , &$+$,1$. /1+$,1 1(&(1 *-++-,0,(3$/0(15-%$!/ 0) (,"-*,1' 0
!$$, ""$.1$#%-/(,"*20(-,(, , &$+$,1$. /1+$,1 "2*152!*(" 1(-,0!5 , 21'-/(6$# #+(,(01/ 1-/-%(&(1 *-++-,0,(3$/0(15-%
$!/ 0) (,"-*,
-00$8$,,(0 ,#/$$,!$/& /*- *$8(,& ,#$/%-/+ ,"$3 *2 1(-,,8/(!21(-, *, *50(0 Management
Department Faculty Publications
'8.0#(&(1 *"-++-,02,*$#2+ , &$+$,1% ".2!
( Academy of Management Journal
1981, Vol. 24, No. 4, 767-779.
Goal Setting and
Performance Evaluation:
An Attributional Analysis
DENNIS L. DOSSETT
CARL I. GREENBERG
University of Nebraska at Omaha
The present study considered the effect of goal setting
on supervisors' evaluations of employees' performance
and the causes attributed to that performance. Results
indicated that attributional distortions were greater in
the assigned than in the participative or self-set goal con-
ditions. Supervisors rated the high participatively-set
goal worker significantly higher in performance, ability,
effort, and goal commitment than they rated the low
participatively-set goal worker.
Performance appraisal rating errors have been given considerable atten-
tion in the literature. The thrust of most of this research has centered
around the psychometric properties of the rating scales (Landy & Farr,
1980) and on the reduction of rater errors through training (Latham, Wex-
ley, & Pursell, 1975). Of probably equal importance in committing rating
errors are the social and situational information cues available to the
evaluator. To date such cues have been largely ignored in the industrial
psychology and management literature (Landy & Farr, 1980).
Social psychologists, however, have been pursuing these informational
cues for some time (Kelley, 1967). Specifically, Weiner, Frieze, Kukla,
Reed, Rest, and Rosenbaum (1971) propose that attributions regarding
task success and failure affect the degree to which an observer attributes
task performance to internal or external causes. According to these
authors, internal attributions are the actor's effort and ability; external at-
tributions are task difficulty and luck. For example, Mitchell and Wood
(1980) found that when the consequences of poor performance were
major, supervisors made more internal causal attributions for that perfor-
mance. In addition, Garland and Price (1977) have shown that the success
of female managers was attributed by prejudiced male managers to luck
and an easy task, but unprejudiced male managers attributed the success
of their female counterparts to skill and hard work.
767
Published by Academy of Management. Used by permission.
768 Academy of Management Journal December
A highly salient set of information cues is provided to the evaluator not
only by the outcome of success or failure, but also by the standards by
which the employee's performance is judged and the method by which
those standards are established. The latter may be accomplished through
engineered work standards, a supervisor either assigning a work goal to
the employee or setting the goal participatively, or even by allowing for
self-set work goals. Thus, an interesting line of research would focus on
the attributions made during performance evaluation as a function of vari-
ous goal setting methods under conditions of task success or failure. That
is, if performance evaluations made by a supervisor are subject to attribu-
tional distortion, is the nature of such distortions dependent on goal at-
tainment? In addition, are the dynamics of different methods of goal set-
ting an appreciable factor in the final evaluation of performance?
To date, no research has systematically investigated the effects of pro-
cess variables on performance appraisal in the context of goal setting. It
seems reasonable that a supervisor's evaluation of an employee's perfor-
mance may be affected by information obtained during the goal setting
process. In considering this process, a supervisor has considerable infor-
mation that may be more or less salient when evaluating an employee's
performance.
One important informational cue is the question of who set the goal.
The degree of influence a supervisor has in setting a goal increases from
self-set, through participative, to assigned goal setting procedures. As a
supervisor's influence increases in setting the goal, the employee's task
performance becomes more "hedonically relevant" to the supervisor
(Jones & Davis, 1965). That is, the success or failure of the employee in
meeting the goal has positive or negative consequences for the supervisor.
Consequently, a supervisor may differentially distort an employee's per-
formance evaluation as a function of the type of goal setting process, espe-
cially if the employee failed to meet the goal. Such failure may be a threat
to the supervisor's self-esteem, the supervisor then making a defensive at-
tribution to avoid that threat (Jones & Davis, 1965). It therefore was hy-
pothesized that an employee's performance evaluation would be affected
by an interaction between his performance outcome and the manner in
which the goal was set. Specifically, it was expected that supervisors would
distort the causal attributions of performance and, hence, the perfor-
mance evaluation for a failing employee more in an assigned goal setting
condition than in either participative or self-set conditions. No significant
distortions were anticipated for successful employees.
A second piece of information that may affect a supervisor's evaluation
of an employee's performance is the difficulty of the goal to which an em-
ployee initially aspires relative to the agreed-upon goal when the goal is
participatively set. For example, a supervisor who is faced with an em-
ployee who initially suggests a very difficult goal is likely to draw different
inferences regarding the causal factors of the employee's performance
than when faced with an employee who initially suggests a relatively easy
1981 Dossett and Greenberg 769
goal. Furthermore, the initial goal suggested by an employee may ulti-
mately affect his performance rating given by his supervisor. Thus, in the
present study two types of participative goal setting conditions were devel-
oped. The employee either initially suggested a more difficult or less diffi-
cult goal relative to the final agreed-upon goal. It was predicted that an
employee suggesting a relatively hard goal would be perceived as being a
better worker with higher ability, goal commitment, and motivation to
perform the task than an employee suggesting a relatively easy goal. The
employee's initially suggested goal was expected to affect differentially the
supervisor's attributions for success or failure and thus the overall perfor-
mance evaluation, because the participative goal setting conditions were
hedonically relevant to the supervisor.
A final, and probably the most potent, determinant of a supervisor's
evaluation of an employee's performance is whether or not the employee
meets the performance goal. In the present study, half the supervisors saw
a worker succeed by exceeding the set goal. Successful goal attainment was
hypothesized to affect significantly supervisor's attributions for the causes
of the worker's performance and their subsequent overall performance
evaluation rating.
The present study was designed to ascertain the effects of different types
of goal setting procedures on a worker's performance evaluation and the
perceived causes of that performance. Subjects were shown one of several
videotapes of a supervisor and a worker setting a performance goal. After
receiving information about the worker's success or failure in meeting the
goal, the subjects assumed the role of the supervisor and evaluated the
worker's performance. Attribution research has demonstrated that role
players make the same attributions as those of actual participants (Bem,
1972). Thus, role players participated in one of three goal setting condi-
tions: participative, assigned, or self-set. The participative goal setting
condition was divided further into two conditions in which the worker
suggested either a relatively hard or easy goal as compared to the final goal
set. Each of these conditions was combined with the worker either suc-
ceeding or failing to meet the goal set. Sex of the subjects was treated as a
blocking variable.
METHOD
Subjects
The participants were 30 male and 50 female undergraduates attending
the University of Nebraska at Omaha. In exchange for their participation
in the experiment, subjects received extra credit in their respective courses.
There were 10 subjects in each of the 8 experimental conditions. However,
the proportion of males to females in each condition varied considerably.
770 Academy of Management Journal December
Procedure
Subjects were shown two videotaped sequences of a male worker. The
first tape (goal setting) involved the worker and his (male) supervisor dis-
cussing how the task should be done and setting a production goal for the
worker. Videotape, rather than a live supervisor-worker interaction, was
used to control goal difficulty and any unique "historical" factors that
could bias the results. Following this tape, the subjects completed a ques-
tionnaire assessing their impressions of the worker, the supervisor, and the
discussion betweep them.
The second videotape (work performance) was a five minute sequence
of the worker performing the task. The task involved collating five order
sheets in a specified random order and adding up numbers obtained from
the five sheets. This task was adapted from a similar collating task used by
Heller, Groff, and Solomon (1977). After subjects viewed this tape they
again completed a questionnaire, this time playing the role of the super-
visor. The subjects were given instructions that the questionnaire was a
performance evaluation for appraising the performance of the worker
they had just viewed. When the subjects had completed this questionnaire
they were debriefed and dismissed.
Goal Setting Manipulation
Four versions of the goal setting videotape were produced. All tapes
were identical except for the part pertaining to setting the goal. In one tape
the worker set his own goal, and in another the supervisor assigned the
goal to the worker. The two participation tapes differed only on the goal
initially suggested by the worker. In one tape (high-goal worker) the
worker initially suggested a goal of 70 completed orders per hour. In the
other tape (low-goal worker), the goal initially suggested was 50 correctly
completed orders per hour. In each of these tapes the supervisor re-
sponded by suggesting 50 or 70 completed orders per hour, respectively.
The final goal was set at 60 orders per hour in both conditions.
For example, in the high-goal worker condition the interaction pro-
ceeded as follows:
Supervisor: "How many orders per hour do you think you can cor-
rectly process? "
Worker: "I think I can do about 70 per hour."
Supervisor: "Well, I think that may be a little high. How do you feel
about a goal of 50?"
Worker: "Well, maybe so. Would 60 an hour be O.K.?"
Supervisor: "That sounds fine! That shouldn't be too difficult; most
of our people make their production goals pretty regu-
larly...."
The remaining two tapes constituted the assigned and self-set goal set-
ting conditions. In the assigned goal setting condition the supervisor told
