Performance appraisal continues to be a widely investigated organizational function. There has been considerable recent interest in examining types of performance appraisals other than the traditional supervisory appraisal. Specifically, peer or coworker appraisals are becoming more popular and are receiving more attention from organizational practitioners and scholars as a result of the continuous improvement philosophy being advanced in many organizations (London & Beatty, 1993) and the increasing use of 360-degree feedback systems (Gruner, 1997; London, Wohlers, & Gallagher, 1990).

A meta-analysis examining the extent of inter-rater agreement between peers and supervisors indicated a correlation between peer and supervisor ratings of .62 (Harris & Schaubroek, 1988). Thus, only 38% of the variance in ratings is accounted for by rater source, indicating that although there is some degree of overlap, peers and supervisors appear to have different frames of reference when evaluating other's performance. Thus, it is important to understand the cognitive processes of both peer and supervisor raters in an evaluative context.

Attributional research examining the actor-observer effect offers an explanation for the differences found between peers and supervisors (Jones & Nisbett, 1972; Martin & Klimoski, 1990; Monson & Snyder, 1977; Storms, 1973; Watson, 1982). The actor-observer effect is the tendency for individuals to attribute their own behavior to situational causes and others' behavior to dispositional or personal factors. Studies have demonstrated that when an individual evaluates his or her own performance, he or she is much more likely to consider situational influences on performance, while observers are more likely to infer that dispositional factors influenced behavior (Martin & Klimoski, 1990; Saulnier & Perlman, 1981).

In the performance appraisal literature, the actor-observer effect has been investigated primarily by studying the attributional differences that exist between the supervisor and the subordinate when each serves as a rater of the subordinate's performance (Martin & Klimoski, 1990; Mitchell & Wood, 1980; Reiss, Rosenfeld, Melburg, & Tedeschi, 1981). The perspective of the peer as a rater has received less attention and thus it is not clear whether peer raters will make attributions that are more consistent with those made by self-raters or supervisor raters.

Thus, the purpose of the present article is twofold. First, we wish to examine whether any differences exist between peers and supervisors in the causal attributions made for performance. Second, we examine whether attributional differences impact subsequent performance evaluations.

Campbell and Pritchard (1976), in their review of research on motivation, stated that there are facilitating and inhibiting conditions (i.e., system factors) not under the control of individuals that can influence their level of performance. For example, employees may receive services and help from others to do their job, or they may have to wait for previous stages of the workflow to be completed. These conditions or situational constraints have been found to influence affective criteria (i.e., frustration, satisfaction) and performance measures (O'Connor, Peters, & Segovis, 1983; O'Connor, Peters, Pooyan, Weekley, & Erenkrantz, 1984; Peters, Fisher, & O’Connor, 1982; Peters, O’Connor, Eulberg, & Watson, 1988; Peters, O’Connor, & Rudolph, 1980).

When employees work under inhibiting conditions, they are often frustrated, dissatisfied with the task, and perform at a lower level. In addition, ratings given to managers facing severe constraints have been found to be lower than those given to managers indicated fewer constraints on their performance (O’Connor et al., 1984). The results from these studies suggest that performance does suffer as a result of the presence of situational constraints. The implications for the present study are twofold. First, it is important to determine if raters will consider the influence of situational constraints on performance and attribute performance to constraints when they are present. Second, it is important to determine if performance ratings will be modified based on the assessment of causal attribution.

Whether raters can distinguish between person and system factors when making evaluations has been a matter of some debate between organizational researchers and scholars. Deming (1986) has voiced concerns regarding performance appraisal at the individual level because of the possibility that raters cannot distinguish between person and system factors. He notes that one of the primary problems with performance appraisal is that appraisals may erroneously attribute variation in performance to individuals rather than system-related problems.

In situations where system factors or constraints are operating, it is important to understand if raters are able to recognize these constraints and provide appropriate performance evaluations in the face of these constraints. Falsely attributing variation in performance to individual factors could lead management to focus on the employee as the cause of poor performance and lead to morale problems among employees.

A few studies have examined whether raters can separate person and system influences on productivity and whether system factors were taken into account when making ratings (Cardy, Sutton, Carson, & Dobbins, 1990; Carson, Cardy, & Dobbins, 1991). The results of these studies indicate that raters do not distinguish between person and system influences and that productivity levels outweigh both person and system factors when raters make appraisal ratings. However, neither of these studies examined the perspective of the peer or coworker in a rating context. The unique observational perspective of the peer may lead to different attributions and ratings than that of a supervisor or casual onlooker.

For example, in a typical organization, coworkers work on similar tasks in similar conditions and have the opportunity to examine the performance of one another in a variety of contexts. Supervisors, on the other hand, are not ordinarily working on the same tasks as subordinates and spend a considerable amount of time working on separate tasks in other areas of the working environment. Thus, there are observational perspective differences that could influence attributions for performance.

The influence of observational perspective on attributions was examined by Storms (1973) and results indicated that when pairs of subjects engaged in face-to-face conversation in front of pairs of observers, actors made situational attributions for their own behavior, while observers made personal attributions for the actors' behavior. However, when videotape was used to reverse the observational perspective, both the actors and observers changed their attributions.

In the present study, we posit that the role of the peer is more similar to that of the actor than the supervisor in a performance situation. Peers, because they typically work in similar situations, have a good understanding of all the conditions that influence performance including the ability and effort it requires, as well as situational factors that may enhance or impede performance. Therefore, it is likely that peers who serve as raters may make attributions that are more consistent with the actor than the observer. In other words, peers may be more likely to attribute the performance of their coworker to external rather than internal factors. Thus, the first hypothesis is proposed: Peer raters will be more likely than supervisors to attribute the cause of performance to situational factors when evaluating performance that is situationally constrained.

Research has indicated that supervisors and casual observers of performance do not take situational factors into account to a great extent when evaluating a performance that is constrained by situational factors (Cardy et al., 1990; Carson et al., 1991). Research comparing peers and supervisory ratings has suggested that peers assign higher ratings for performance than do supervisors (Schneier & Beatty, 1978). This rating difference may be due to the attributional differences that exist between peers and supervisors. Thus, the second hypothesis is proposed: Peer raters will assign higher ratings to ratees working under conditions of constraint than will supervisory raters.

One hundred and eleven undergraduate students enrolled in various psychology courses at a large Northeastern university served as raters in this study. Participants either received extra credit for participation in this study or fulfilled a course requirement for research participation. The mean age of the participants was 18.5 years.

Upon arrival, participants were greeted by an experimenter and told that they were going to be working in a simulated working environment. Participants were told that they would be working on a task where the objective was to schedule bands to a concert series based on the times the bands were free and the times that the concert halls were available.

Peer and supervisor manipulations. Participants who were assigned to the peer condition were told that they would be working independently on the same scheduling task for 20 minutes. Participants were also informed that they would be asked to rate their coworker’s performance at the completion of the task as well as complete a questionnaire regarding their experience in this research.

Participants who were assigned to the supervisor condition were told that they would be randomly assigned to serve as either a supervisor or subordinate in the simulated working environment. The experimenter flipped a coin to determine who would serve as subordinate. The participants who served as supervisors were told that they would be working on a separate managerial task (i.e., the development of a new list of bands, times and possible schedules) and would be asked to complete that task as well as rate the performance of his/her subordinate. The participants who were chosen to be subordinates were told that they would be working on a scheduling task that differed from task their supervisor would be completing and that their performance would be rated by their supervisor upon completion of the task.

These manipulations were chosen to mirror the typical differences that exist in many organizations. Specifically, most coworkers work on similar tasks while most supervisors complete tasks that differ from those their subordinates carry out.

Performance constraint manipulation. At the beginning of each experimental session, all participants were told that they would be working on a scheduling task for 20 minutes. In the unconstrained condition, participants were allowed 20 minutes to work on their assigned tasks and then completed the questionnaires. In the constrained condition, participants were interrupted after the experimenter received a knock on the door halfway through the session (after 10 minutes). The experimenter returned flustered, told the participants that he/she had to stop the session early because someone else needed to use the room, ended the task, and administered the questionnaires. Thus, this manipulation included factors beyond the individual’s control (i.e., inadequate time to complete the task), but that directly influenced performance.

Performance information. Upon completion of the task, the rater was provided with false information about the ratee’s performance. Specifically, the rater(s) in each condition were given information that the ratee’s performance was either positive (the person performed better than 90% of others who had completed the task) or average (the person performed better than 50% of others). The rater was asked to rate the individual based on the performance information provided. This performance information was provided to raters to give them some indication of the ratee’s performance level. Because this was a task that the participants were unfamiliar with, it is likely that they would have difficulty assigning performance ratings in the absence of employee comparison data.

Measures

Manipulation checks. To assess the constraint manipulation, raters were asked to indicate their agreement with the following item "I feel that there was enough time for the person to adequately complete the task." A 5-point Likert scale was used with 1 indicating strong disagreement and 5 indicating strong agreement. This item was completed following the completion of the attribution and rating scales.

Attribution ratings. The attributions for performance were examined using a 4-item measure. Each rater was asked to indicate the extent to which he believed the person put forth effort and ability on a 5 point Likert scale (1 = no effort or ability and 5 = very high effort or ability). To assess perceptions of situational constraint, raters were asked to indicate their strength of agreement if with the following item on a 5 point Likert scale (1 = strong disagreement and 5 = strong agreement): "Factors outside of the individual’s control influenced performance." Finally, raters were asked on a 5-point Likert scale whether the task was very easy (1) or very difficult (5).

Performance ratings. Five items were used for the rater’s evaluation of the ratee’s overall performance. Raters were asked to assess the ratee’s dependability, energy level, efficiency, quality of work, and overall performance. Rating were made using a 7-point Likert scale (1= very poor to 7 = excellent). Coefficient alpha for the performance measure was .94.

Manipulation checks. Raters indicated that the ratee believed that there was less time to adequately complete the task in constrained than unconstrained conditions (constrained M = 2.70, unconstrained M = 3.43), t = 3.199, p < .05. This suggests that the raters did perceive the impact of time on the performance levels of ratees.

Causal attributions. Using situational attributions as the dependent variable in a 2 (Rater) x 2 (Level of constraint) x 2 (performance information) analysis of variance (ANOVA) was used to assess whether peers made more situational attributions than supervisors in constrained situations. Descriptive statistics for attributions are found in Table 1. ANOVA results can be found in Table 2. A main effect was found for level of constraint (F (1,103) = 12.87, p <.01). Raters were more likely to attribute performance to situational constraints when viewing a situationally constrained performance (M = 3.95) compared to an unconstrained performance (M = 3.22).

In support of Hypothesis 1, an interaction of rater and level of constraint was found (F (1,103) = 7.22. p < .01). Figure 1 displays this interaction. Peer raters were more likely to attribute performance to situational constraints (M = 4.27) than supervisor raters (M = 3.63) in the constrained conditions. There were no other main effects or interactions for the attributional dependent variable of situational constraint.

Additional supplemental analyses were conducted to examine ability, effort and task difficulty causal attributions. A 2 (Rater) x 2 (Level of Constraint) x 2 (Performance Information) ANOVA was conducted for each of the ability, effort, and difficulty attributions. Table 3 displays these results. A main effect of level of constraint was found for ability attributions (F(1,103) = 5.69, p <.05). Raters were more likely to attribute the cause of performance to higher ability in the constrained conditions (M = 3.77) than in the unconstrained conditions (M = 3.33). A main effect of performance information also emerged (F(1,103) = 6.50, p <.05). Raters were more likely to attribute performance to higher ability when they received positive performance information about the ratee (M = 3.79) than when they received average performance information (M = 3.35). There were no other main effects or interactions for the ability dependent variable.

A main effect of performance information was found for effort attributions (F(1, 103) = 8.76, p <.01). ANOVA results for the dependent variable of effort can be found in Table 4. Raters were more likely to say that the ratee put forth more effort when they received positive information (M = 4.20) than when they received average information about the ratee’s performance (M = 3.73). There were no other main effects or interactions for the dependent variable of effort.

ANOVA results for task difficulty attributions are presented in Table 5. A main effect of rater was found for task difficulty attributions (F(1,103) = 3.92, p = .05). Peers were less likely to attribute performance to task difficulty (M = 2.52) than supervisors (M = 2.93).

Performance ratings. To assess whether peers assign higher ratings than supervisors when viewing a performance that is situationally constrained, a 2 (Rater) x 2 (Level of Constraint) x 2 (performance information) ANOVA was performed. Descriptive statistics are displayed in Table 6. ANOVA results are displayed in Table 7. A main effect of performance information was found for ratings (F(1,103) = 128.55, p < .01). Raters were more likely to provide higher performance ratings when they were told that the ratee performed better than 90% of others (M = 5.99) than when they were told the ratee performed better than 50% of others (M = 4.32).

Contrary to expectations, there was no interaction of rater and level of constraint. Peers were not more likely than supervisors to assign higher ratings to ratees whose performance was constrained. However, level of constraint did interact with performance information (F(1,103) = 4.31, p <.05). Figure 2 displays this interaction. An examination of the differences between ratings in the negative information condition for the constrained and unconstrained conditions indicated that ratings were higher when negative information was provided in the constrained condition, t = 2.36, p <.05. However, there were no significant differences in ratings between the constrained and unconstrained condition when positive information was provided, t = .40, p > .05.

The purpose of this laboratory experiment was to examine whether peer raters are more sensitive than supervisory raters to the impact of situational constraints on performance. Our data suggests that while peer raters are more likely to attribute performance to constraints when constraints are operating, they are not more likely to adjust performance ratings to account for the impact of constraints. Further, an examination of the internal attributions made for performance (i.e., effort and ability) demonstrates that peers and supervisors do not differ in their assessment of whether ability or effort play a role in performance. Both peers and supervisors attributed performance to higher ability in the constrained than unconstrained situation. Furthermore, raters attributed performance to high effort when they received positive performance information about the ratee’s performance.

Peers and supervisors did differ in their attributions about task difficulty. Supervisors were more likely than peers to say that the level of task difficulty affected performance. These results may be due to differential familiarity with the task between peers and supervisors. Supervisors did not actually engage in the task and may have relied on the performance information they received to judge the difficulty of the task.

Because the peer raters were more likely to recognize the impact of constraints on performance than the supervisory raters, we were surprised that our second hypothesis was not supported. We predicted that peer raters would assign higher ratings to ratees than supervisor raters in the constrained conditions because they would take into consideration the influence of constraints on resulting performance. Our results indicated that peers and raters provided similar ratings for ratees in both constrained and unconstrained situations. This finding may have been due to the lack of familiarity with the task and the salience of the performance information. It is possible that both peers and supervisors were relying extensively on the performance information they received (i.e., whether the ratee performed better than 50% or 90% of others) to assign performance ratings. This would account for the similar ratings found between peers and supervisors. Alternatively, peers may have been unwilling to inflate the ratings of their coworker’s performance (under constraints) because their performance would look worse by comparison.

One unhypothesized, yet interesting, finding suggests that raters give similar ratings in both constrained and unconstrained situations when the ratee performs relatively well compared to others, but when the ratee performs only average, ratings are higher in the constrained situation. This suggests that both peers and supervisors are adjusting their ratings of ratees who perform average in constrained conditions, but are not adjusting their ratings of ratees who perform well in constrained conditions. This implies that when constraints operate to significantly impede performance, ratees are afforded a level of consideration that slightly impaired performers are not afforded. In an organizational context, this means that a person who excels under conditions of constraint will be given lower ratings than they may truly deserve compared to a person who performs average under similar conditions of constraint.

Our research addresses Deming’s (1986) concern that raters cannot distinguish between person and system factors. We find that peers (when working under similar conditions to one another) are more able than supervisors to identify the influence of constraints on behavior, but are not more likely to adjust performance ratings upward to account for the impact of the constraints. Thus, his concern that performance appraisals may erroneously attribute variation in performance to individuals rather than system-related problems is supported by our performance rating data.

Certain limitations of the current research should be noted. In our simulated workplace, individuals were working together for only a brief time (i.e., 10 – 20 minutes) on a single task. This does not mirror the typical workplace and may not reflect the myriad of factors that influence attributions and ratings. Research has shown that peer ratings are fairly reliable only when the peers who make the ratings are similar to and well acquainted with the employees being rated (Landy & Guion, 1970; Mumford, 1983). In our study, raters were unfamiliar with the ratee both personally and professionally. Thus, future research should examine attributions in an organizational setting where familiarity between rater and ratee is maximized.

However, because there were no organizational ramifications of our performance ratings (i.e., salary and/or promotion were not contingent upon performance ratings), it is likely that ratings were a reflection of perceived performance rather than concerns about how organizational resources would be allocated. In an organizational context, performance ratings made by peers may be different when valued organizational resources are at stake. Future research should examine the attributions for performance and performance ratings in a complex organizational system to assess whether peer raters are able to discern the impact of constraints on the performance of their peers when numerous activities and responsibilities compete for their attention and cognitive resources.

Barrett, R. S. (1966). Performance rating. Chicago: Science Research Associates.

Campbell, J., & Pritchard, R. (1976). Motivation theory in industrial and organizational psychology. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 63-130). Skokie, IL: Rand McNally.

Cardy, R. L., Sutton, C. L., Carson, K. P., & Dobbins, G. H. (1990, August). Degree of responsibility: An empirical examination of person and system effects on performance ratings. Paper presented at the meeting of the Academy of Management, San Francisco, CA.

Carson, K. P., Cardy, R. L., & Dobbins, G. H. (1991). Performance appraisal as effective management or deadly management disease. Group and Organization Studies, 16, 143-159.

Cederblom, D., & Lounsbury, J. W. (1980). An investigation of user acceptance of peer evaluations. Personnel Psychology, 33, 567-579.

Deming, W. E. (1986). Out of the crisis. Cambridge, MA: Massachusetts Institute of Technology.

DeNisi, A. S., & Mitchell, J. L. (1978). An analysis of peer ratings as predictors and criterion measures and a proposed new application. Academy of Management Review, 3, 369-374.

Dougan, K. W. (1989). Ability and effort attributions: Do they affect how managers communicate performance feedback information? Academy of Management Journal, 32, 87-114.

Gruner, S. (1997). Feedback from everyone: Are 360-degree performance reviews a silly fad – or a smart management tool? Inc., 19 (2), 201-103.

Harris, M. M., & Schaubroeck, J. (1988). A meta-analysis of self-supervisor, self-peer, and peer-supervisor ratings. Personnel Psychology, 41, 43-62.

Hollander, E. P. (1957). The reliability of peer nominations under various conditions of administration. Journal of Applied Psychology, 41, 85-90.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Consequences of individual feedback on behavior in organizations. Journal of Applied Psychology, 64, 349-371.

Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: Divergent perceptions of causality. In E. E. Jones et al. (Eds.), Attribution: Perceiving the causes of behavior (pp. 79-94). Morristown, NJ: General Learning Press.

Landy, F. J. & Guion, R. M. (1970). Development of scales for the measurement of work motivation. Organizational Behavior and Human Performance, 5, 93-103.

London, M., & Beatty, R. (1993). 360-degree feedback as a competitive advantage. Human Resource Management, 32, 353-372.

London, M., Wohlers, A., & Gallagher, P. (1990). A feedback approach to management development. The Journal of Management Development, 9, 17-31.

Martin, S. L., & Klimoski, R. J. (1990). Use of verbal protocols to trace cognitions associated with self- and supervisor evaluations of performance. Organizational Behavior and Human Decision Processes, 46, 135-154.

Mitchell, T. R., & Wood, R. E. (1980). Supervisor's responses to subordinate poor performance: A test of an attributional model. Organizational Behavior and Human Performance, 25, 123-138.

Monson, T. C., & Snyder, M. (1977). Actors, observers, and the attribution process: Toward a reconceptualization. Journal of Experimental Social Psychology, 13, 89-111.

Mumford, M. D. (1983). Social comparison theory and the evaluation of peer evaluations: A review and some applied implications. Personnel Psychology, 36, 867-881.

O'Connor, E. J., Peters, L. H., & Segovis, J. C. (1983). Situational constraints, task-relevant abilities, and experienced frustration. Personnel Selection and Training Bulletin, 4, 178-188.

O'Connor, E. J., Peters, L. H., Pooyan, A., Weekley, J., Frank, B.,& Erenkrantz, B. (1984). Situational constraint effects on performance, affective reactions, and turnover: A field replication and extension. Journal of Applied Psychology, 69, 663-672.

Peters, L. H., Fisher, C. D., & O'Connor, E. J. (1982). The moderating effect of situational control of performance variance on the relationship between individual differences and performance. Personnel Psychology, 35, 609-621.

Peters, L. H., O'Connor, E. J., Eulberg, J. R., & Watson, T. W. (1988). An examination of situational constraints in Air Force work settings. Human Performance, 1, 133-144.

Peters, L. H., O'Connor, E. J., & Rudolph, C. J. (1980). The behavioral and affective consequences of performance-relevant situational variables. Organizational Behavior and Human Performance, 25, 79-96.

Saulnier, K., & Perlman, D. (1981). The actor-observer bias is alive and well in prison: A sequel to Wells. Personality and Social Psychology Bulletin, 7, 559-564.

Schneier, C. E., & Beatty, R. W. (1978). The influence of role prescriptions on the performance appraisal process. Academy of Management Journal, 21, 129-135.

Scholtes, P. R. (1987). An elaboration on Deming's teachings on performance appraisal. Madison, WI: Joiner Associates, Inc.

Snyder, M. L., & Jones, E. E. (1974). Attitude attribution when behavior is constrained. Journal of Experimental Social Psychology, 10, 585-600.

Storms, M. D. (1973). Videotape and the attribution process: Reversing actors' and observers' points of view. Journal of Personality and Social Psychology, 27, 165-175.

Watson, D. (1982). The actor and the observer: How are their perceptions of causality divergent? Psychological Bulletin, 92, 682-700.

West, S. G., Gunn, S. P., & Chernicky, P. (1975). Ubiquitous Watergate: An attributional analysis. Journal of Personality and Social Psychology, 32, 55-65.

Table 1.

Causal Attribution Cell Means and Standard Deviations

______________________________________________________________________________

Positive Performance Information Average Performance Information

____________________________ _____________________________

Unconstrained Constrained Unconstrained Constrained

Supervisor Rater

Constraints

Mean 3.308 3.733 3.583 3.533

SD .855 1.223 1.084 .990

Ability

Mean 3.692 4.200 2.917 3.800

SD .855 .862 .900 .775

Effort

Mean 4.308 4.400 3.583 3.667

SD .630 .828 1.165 .816

Difficulty

Mean 3.077 2.933 2.667 3.000

SD .954 1.280 1.231 .756

Peer Rater

Constraints

Mean 2.667 4.000 3.286 4.571

SD 1.231 1.095 1.490 .514

Ability

Mean 3.667 3.563 3.071 3.500

SD .888 1.209 .997 .941

Effort

Mean 4.083 4.000 4.000 3.643

SD .900 .816 .679 .842

Difficulty

Mean 2.417 2.562 2.214 2.857

SD 1.165 .964 .975 1.231

_______________________________________________________________________

Table 2

________________________________________________________________________

Source SS DF MS F p h²

________________________________________________________________________

Rater (A) .23 1 .23 .19 .662 .002

Performance Information(B) 2.75 1 2.75 2.30 .132 .022

Level of Constraint (C) 15.40 1 15.40 12.87 .001 .111

A x B 2.13 1 2.13 1.78 .185 .017

A x C 8.64 1 8.64 7.22 .008 .066

B x C .47 1 .47 .39 .532 .004

A x B x C .31 1 .31 .26 .609 .003

Within Cells 123.30 103 1.20

Total 152.34 110 1.38

______________________________________________________________________________