Performance Appraisal
Performance Appraisal Systems
Problems with Performance Appraisals
A number of problems can be identified that pose a threat to the value of appraisal techniques. Most of these problems deal with the related issues of the validity and reliability of the instruments or techniques themselves. Validity is the extent to which an instrument actually measures what it intends to measure, whereas reliability is the extent to which the instrument consistently yields the same results each time it is used. Ideally, a good performance appraisal system will exhibit high levels of both validity and reliability. If not, serious questions must be raised concerning the utility (and possibly the legality) of the system.
It is possible to identify several common sources of error in performance appraisal systems. These include: (1) central tendency error, (2) strictness or leniency error, (3) halo effect, (4) recency error, and (5) personal biases.
Central Tendency Error. It has often been found that supervisors rate most of their employees within a narrow range. Regardless of how people actually perform, the rater fails to distinguish significant differences among group members and lumps everyone together in an "average" category. This is called central tendency error and is shown in Exhibit 8.3. In short, the central tendency error is the failure to recognize either very good or very poor performers.
Strictness or Leniency Error. A related rating problem exists when a supervisor is overly strict or overly lenient in evaluations (see Exhibit 8.3). In college classrooms, we hear of professors who are "tough graders" or, conversely, "easy A's". Similar situations exist in the workplace, where some supervisors see most subordinates as not measuring up to their high standards, whereas other supervisors see most subordinates as deserving of a high rating. As with central tendency error, strictness error and leniency error fail to distinguish adequately between good and bad performers and instead relegate almost everyone to the same or related categories.
Halo Effect. The halo effect exists where a supervisor assigns the same rating to each factor being evaluated for an individual. For example, an employee rated above average on quantity of performance may also be rated above average on quality of performance, interpersonal competence, attendance, and promotion readiness. In other words, the supervisor cannot effectively differentiate between relatively discrete categories and instead gives a global rating.
These types of bias are based on our perceptions of others. The halo effect occurs when managers have an overly positive view of a particular employee. This can impact the objectivity of reviews, with managers consistently giving an employee high ratings and failing to recognize areas for improvement.
Whether positive or negative, we also have a natural tendency to confirm our preconceived beliefs about people in the way we interpret or recall performance, which is known as confirmatory bias.
For example, a manager may have a preconception that her male report is more assertive. This could cause her to recall instances more easily in which her report asserted his position during a meeting. On the other hand, she may perceive her female report to be less assertive, predisposing her to forget when the report suggested an effective strategy or was successful in a tough negotiation.
The halo effect is often a consequence of people having a similarity bias for certain types of people. We naturally tend to favor and trust people who are similar to us. Whether it's people who also have a penchant for golf or people who remind us of a younger version of ourselves, favoritism that results from a similarity bias can give certain employees an unfair advantage over others. This can impact a team to the point that those employees may receive more coaching, better reviews and, as a result, more opportunities for advancement.
Recency Error. Oftentimes evaluators focus on an employee's most recent behavior in the evaluation process. This is known as the recency error. That is, in an annual evaluation, a supervisor may give undue emphasis to performance during the past months - or even weeks - and ignore performance levels prior to this. This practice, if known to employees, leads to a situation where employees may "float" for the initial months of the evaluation period and then overexert themselves in the last few months or weeks prior to evaluation. This practice leads to uneven performance and contributes to the attitude of "playing the game".
Personal Biases. Finally, it is not uncommon to find situations in which supervisors allow their own personal biases to influence their appraisals. Such biases include like or dislike for someone, as well as racial and sexual biases. Personal biases can interfere with the fairness and accuracy of an evaluation and are illegal in many situations.