Literature on Performance Appraisal and Recommendations for Practitioners
Issues for the accuracy and fairness in performance appraisal have been one of the core research interests in the area of human resource management. Here, I will review the literature on performance appraisal that focuses mainly on the internal causes of rater error and biases (test theory perspective and cognitive perspective), external causes of rater biases such as rating inflation (political and goal-oriented perspective), and due process perspective that mainly focuses on the fairness issues. Then, I will recommend the appropriate appraisal systems based on the literature review.
Test theory perspective
Many of the early research on performance appraisal are based on the test theory perspective. The main assumption of this perspective is that there is an objective performance and people can measure such performance accurately if appraisal instruments are appropriate. This perspective focuses on the issues of reliability and validity of performance constructs (e.g. Austin et al. 1993; Smith, 1967), performance standards (e.g, Bobko & Collela, 1991; Huber), rating format (e.g., Latham & Sarri) and so on.
With regard to the reliability and validity of performance constructs, Smith (1976) proposes the importance of "criterion-problem" suggesting that relevancy, reliability and practicality are required for performance criteria and that criterion contamination and criterion deficiency should be avoided. Campbell and his colleagues (e.g., Campbell, 1990) propose the theory of performance, which suggests that performance is basically multidimensional and behavioral. They propose the model in which job performance is determined by the multipricative function of (1) declarative knowledge, (2) procedural knowledge and skills, and (3) motivation. Also, they propose the latent variables of job performance such as core job proficiency, non-core proficiency, demonstrative effort, personal discipline, and so on. Their model is similar to what Borman & Motwidlo (1993) propose as task performance (TP) and contextual performance (CP).
In order to increase the reliability and validity of performance measurements, various approaches are recommended such as multiple raters, combination of objective and subjective criteria, and so on.
With regards to the performance standard, Bobko & Collela (199?) suggest that performance standard are external to the individual and for evaluative purpose and that it is different from individual goals as a person's internal aim. They propose the importance of the employee reaction and acceptance of performance standards. Huber and her colleagues suggest that performance standards should be clear and specific in order for raters to evaluate performance accurately using such standards.
There were also many studies about rating scales. The major concern was, "What kind of rating scales increase rating accuracy?" Behavioral Anchored Rating Scale (BARS), Behavioral Observation Scales, and other similar scales were proposed to increase the rating accuracy. However, as Landy and Farr (1980) suggest, rating scales appear to have little impact on increasing rating accuracy. Rather, researchers started to focus on the information processing aspect of human being that may affect rating error and biases.
In sum, test theory perspective focuses on the importance of performance definition, performance standards, and rating instruments such as rating format and multiple raters.
Cognitive theory perspective
One of the earliest models of the cognitive perspective is Wherry (1952)'s model, which consists of (1) ratee performance and random errors, (2) rater observation of the performance, and (3) rater recall of the performance. The basic premise of this approach is that human being is a information processor (Landy & Farr, 1980).
Many models have been proposed by cognitive researchers which consist of selective attention, encoding and storage, retrieval and integration, and judgement (e.g., DeNisi & Williams, 1988). The seminal work by Feldman (1981) is prominent in this area. He suggests that there are automatic and controlled processes in human cognition and in many cases, people rely on automatic process, which may cause various rating error or biases. Because individual's information processing, especially automatic process, rely heavily on the schema, prototype or categories, these prototypes play an important role in rating accuracy. That is, if important information is screened out in the selective attention process, and the behaviors in prototypes are recalled instead of actual ratees' behaviors, these cognitive processes will result in inaccurate observation, recall, and judgement.
Other researchers focus on the specific rating biases such as halo and leniency. For example, Cooper (1981) suggests that halo (i.e., the correlation among different performance dimension) has both true halo and illusory halo and that illusory halo can be caused by the cognitive distortion of the rater. Some researchers found that there was individual difference in rating leniency.
In sum, researchers in cognitive area suggest that human cognitive process play a major role in rating accuracy and biases. In order to increase rating accuracy, researchers recommend the frame of reference (FOR) training or attributional training, utilizing observation methods such as using performance diary, and so on.
Political and goal-oriented perspective
Researches such as Banks & Murphy (1985) suggest that accurate judgement and giving accurate rating is different. That is, people might be able to judge ratee performance but they might be motivated to giving biased ratings (e.g., leniency). This perspective focuses on the context in which performance appraisal is conducted. Also, other researchers such as Ferris and his colleagues focus on the political behaviors of the employees and suggest that political behaviors such as impression management affect the rating accuracy and biases.
According to the Murphy and Cleveland's (1995) goal-oriented perspective, the goal of performance appraisal is important. The goals of performance appraisal can be (1) personnel decision-making, (2) development of the employees, (3) effective communication, and (4) legal defensibility. Research shows that if performance appraisal is conducted as administrative purpose, the rating inflation will occur. Because supervisors are reluctant to give negative rating that leads to negative personnel decisions or they fear other negative outcomes such as retaliation or loss of reputation, they tend to give higher rating in performance appraisal that is for administrative purpose.
Research on influence tactics (e.g. Wayne & Liden, 1995; Ferris et al. 1999) shows that ratees' impression management such as supervisor focussed tactics (e.g., ingratiation) increases performance rating through liking and perceived similarities. Also, research shows that the use of influence tactics leads to the perceived fairness because of their "voice" effect. Therefore, these researches suggest that organizational context and political behaviors should be considered seriously in order to increase accuracy and fairness of performance appraisal.
Due process perspective
Other researchers focus more on the organizational justice perspective in performance appraisal in order to increase the appropriateness of the system. Greenberg and other researchers demonstrated that there are distributive justice (perceived fairness of the outcome) and procedural justice (perceived fairness of the procedure) and that procedural justice causes the satisfaction toward the organizations or systems while distributive justice cause outcome satisfaction. The conditions of procedurally fair system are consistency, free from biases, accuracy, communication, voice, ethics, and so on (Leventhal, 1980).
Based on the justice research, Folger et al. (1992) propose due process metaphor of the performance appraisal. According to them, the conditions of fair performance appraisal are (1) adequate notice, (2) fair hearing, and (3) judgement based on evidence.
Recommendations for fair and accurate appraisal systems
Based on the literature review above, some recommendations can be made for accurate and fair performance appraisal systems. First, the goal of performance appraisal should be specified (i.e., administrative, developmental, communicative, or legal purpose). Different approaches should be used for different purposes. If the system is likely to produce rating inflation, an appropriate goal setting strategies and rewards and punishments should be administered in order to eliminate the obstacles for the accurate ratings. For example, organizations should motivate raters to give accurate ratings by rewarding accreting ratings. Using multiple rates can be used to detect deviant raters. Also, political behaviors by employees should be avoided. That is, some systems or procedures that discourage the political behaviors should be developed.
Second, definition of job performance and setting clear and specific performance standards are very important not only for accurate ratings but also for increasing employee performance. Also, the standards should be noticed as early as possible. Adequate notice of performance standers increases the perceived fairness.
Third, rater training will increase rating accuracy. Because people tend to rely on cognitive categories or prototypes in information acquisition, storage and retrieval, and making judgement, frame-of-reference (FOR) training helps to share same performance categories among raters, which increase the rating accuracy.
Fourth, the use of multiple raters who are familiar with the ratee performance is recomended. Multiple rater may increase rating reliability through interrater agreement. Raters can be supervisors, subordinates, peers or customers. However, when using multi-source feedback or appraisals, the issue of accountabilities should be considered (e.g., Antonioni, 1994).
Fifth, the procedural justice and interactional justice should be included in the process. For example, giving voice for employees (e.g., participation to the appraisal process, fair hearing, and opportunities to demonstrate their performance) is effective way to increase perceived fairness. Honest feedback and quality of interpersonal communication are also the key to the fair procedure.
Future research directions
Future research is needed to understand performance appraisal accuracy and fairness. One of the hot issues may be the area of influence tactics (e.g., Ferries et al. 1999). Also, the research on multi-source feedback is on the way. The development of these areas should increase our understanding of the performance appraisal.