Improving performance evaluations


The performance evaluation process is a potential minefield of problems. For instance, evaluators can make leniency, halo, and similarity errors, or use the process for political purposes. They can unconsciously inflate evaluation (positive leniency), understate performance (negative leniency), or allow the assessment of one characteristic to unduly influence the assessment of other characteristics (the halo error). Some appraisers bias their evaluations by unconsciously favoring people who have qualities and traits similar to themselves (the similarity error). And, of course, some evaluators see the evaluation process as a political opportunity to overtly reward or punish employees they like or dislike. Although there are no protections that will guarantee accurate performance evaluations, the following suggestions can significantly help the process more objective and fair.

Emphasize Behavior Rather Than Traits

Many traits often considered to be related to good performance may have little or no relationship to performance. For example, traits such as loyalty initiative, courage, reliability and self-expression are intuitively appealing as characteristics in employees. But the relevant question is: Are individual who are evaluated as high on those traits higher performers than those who rate low? We can’t answer this question easily. We know that there are employees who rate high on these characteristics and are poor performers. We can find others who are excellent performers but do not score well on traits such as these. Our conclusion is that traits such as loyalty and initiative may be prized by managers there is no evidence to support that certain traits will be adequate synonyms for performances in a large cross section of jobs.

Another weakness of trait evaluation is the judgment itself. What is “loyalty�? When is an employee “reliable�? What one considers as “loyalty� the other may not. So traits suffer from weak inter-rater agreement.

Document Performance Behavior in a Journal

Journal helps evaluators to better organize information in their memory. The evidence indicates that keeping a journal of specific critical incidents for each employee helps evaluators to be more
accurate and less prone to rating errors. Journals, for instance, tend to reduce leniency and halo errors because they encourage the evaluator to focus on performance related behaviors rather than on traits.

Use Multiple Evaluators

As the number of evaluators increases, the probability of attaining more accurate information increases. If rater error tends to follow a normal curve, an increase in the number of appraisers will tend to find the majority congregating about the middle. You see this approach being used in athletic competitions in such sports as diving and gymnastics. A set of evaluators judges a performance, the highest and lowest scores are dropped and the final performance evaluation is made up from the cumulative scores of those remaining. The logic of multiple evaluators applies to organizations as well.

If an employee has had ten supervisors, nine having rated her excellent and one poor, we can discount the value of the one poor evaluation. Therefore, by moving employees about within the organization so as to gain a number of evaluations or by using multiple assessors (as provided in 360-degree appraisals), we increase the probability of achieving ore valid and reliable evaluations.

Evaluate selectively

Appraisers should evaluate only in areas in which they have some expertise. If raters make evaluations on only the dimension on which they are in a good position to rate, we increase the inter rater agreement and make the evaluation a more valid process. This approach also recognizes that different organizational levels often have different orientation toward those being rated and observe them in different settings. In general, therefore, we would recommend that appraisers should be as close as possible, in terms of organizational level, to the individual being evaluated. Conversely, the more levels that separate the evaluator and the person being evaluated, the less opportunity the evaluator has to observe the individual’s behavior and, not surprisingly, the greater the possibility for inaccuracies.

Train Evaluators

If you can’t find good evaluators, the alternative, is to make good evaluators. There is substantial evidence that training evaluators can make them more accurate raters.

Common errors such as Halo and leniency have been minimized or eliminated in Workshops where managers practice observing and rating behaviors. These workshops typically run from one to three days, but allocating many hours to training may not always be necessary. One case has been cited in which both halo and leniency errors were decreased immediately after exposing evaluators to explanatory training sessions lasting only five minutes. But the effects of training appear to diminish over time. This suggests the need for regular refresher sessions.