La relation entre le contexte de l'évaluation du rendement et l'indulgence de l'évaluateur

Résumé

La présente étude s'insère dans les multiples efforts de recherche déployés pour mieux circonscrire les paramètres de l'évaluation du rendement. À partir du modèle de Murphy et Cleveland (1995), les auteurs développent une méthodologie originale qui permet de tester empiriquement auprès de 106 fonctionnaires de la fonction publique québécoise la motivation de l'évaluateur à produire des évaluations indulgentes de leurs subordonnés. Les résultats révèlent que l'indulgence s'avère une réponse à un contexte défavorable d'évaluation : les variables contextuelles influencent significativement les appréciations faites par l'évaluateur.

Abstract

There has been a substantial amount of research on performance appraisal. Practically all of this work has focused on understanding and improving a rater's ability to provide accurate ratings. Thus a plethora of research exists regarding such issues as the effect of rating formats and training on the ability to perform accurate evaluations of subordinates. More recently, several researchers have suggested that increasing the quality of performance ratings can only occur through a better understanding of the cognitive processes that underlie performance judgments. This perspective argues that rating errors and inaccurate appraisals are a function of the rater's information processing capabilities.

Recent models of performance appraisal have focused on motivational factors rather then cognitive deficits as explanations for apparent rater errors. Despite recent calls for more work in this area, very few studies have investigated the motivation to inflate ratings in the performance appraisal context. The purpose of this paper is to examine the relationship between the performance appraisal context and rater motivation to inflate ratings. Hypotheses were developed from the Murphy and Cleveland (1995) model of rater leniency. The assumption underlying the study is that rating inaccuracy is predominantly intentional.

The participants in the study were 106 managers in the Quebec public sector. Rating inflation was defined as the discrepancy between public and private performance appraisal ratings for a target ratee. A standardized interview and two questionnaires were used to collect the data. To gather public rating for the target ratee, each participant was asked to get an anonymous copy of the target ratee's last performance appraisal from the human resource department. Each participant's private ratings of the target ratee were collected. Private performance ratings consisted of the rater's personal judgment of the employee's performance during the most recent performance appraisal period. These ratings were made during the interview on a copy of the appraisal form normally used by the ratee. After finishing the interview, the researcher gave participants a first questionnaire and one month later sent the second questionnaire. The questionnaires included measures of context variables.

As expected, raters' perceptions of the performance appraisal context are related to rating behaviour. Specifically, the results show that the quality of the interpersonal relationship between supervisor and subordinate influence rating inflation. The ratings of an employee in low-quality relationships are inflated. In contrast, supervisory ratings are more accurate for employees in high-quality relationships. A supervisons perceptions of subordinates' self rating of their performance is related to rating inflation. This accountability pressure might arise because supervisors wish to avoid conflict. The level of rating inflation varies across raters and in relation to the type of standard used to judge performance. The lack of clear performance standards is related to rating inflation. Discomfort in giving feedback was not significantly related to rater motivation to inflate ratings.

The results also indicate that the purpose of performance ratings effects rating inflation. Rating inflation will be more likely to occur when performance appraisal is not linked to human resource management decisions. Rater trust in the appraisal System is likely to affect rater motivation. Low trust in the System is related to rating inflation. As predicted, a rater's impression of management activities is related to rating inflation. Raters are likely to inflate ratings to maintain a positive image of the organization and to maintain an appropriate image vis-a-vis his or her subordinates. Consistent with the hypothesis, managers may be more likely to inflate ratings when there are political influences within the performance appraisal process. Overall, the findings suggest that the performance appraisal context does affect rater behaviour. This research helps to bridge the gap between practice and research in the area of performance appraisal.

Resumen

El présente estudio se intégra dentro de los multiples esfuerzos de investigaciòn desarrollados para mejor comprender los paràmetros de evaluaciòn del rendimiento. A partir del modelo de Murphy y Cleveland (1995), los autores desarrollaron una metodologia original que permite probar empìricamente de un grupo de 106 funcionarios de la funciòn publica de Québec, la motivaciòn del evaluador de producir evaluaciones indulgentes de sus subordinados. Los resultados revelaron que la indulgencia se muestra como una respuesta a un contexto desfavorable de evaluaciòn: las variables conceptuales influencian significativamente las apreciaciones hechas por el evaluador.