Most companies have a formal performance appraisal system in which employee job performance is rated on a regular basis, usually once a year. A good performance appraisal system can greatly benefit an organization. It helps direct employee behavior toward organizational goals by letting employees know what is expected of them, and it yields information for making employment decisions, such as those regarding pay raises, promotions, and discharges.
Developing and implementing an effective system is no easy task, however. For instance, one study found that a majority of companies—65 percent—are dissatisfied with their performance appraisal systems. Analysts have found that a fairly low degree of reliability and validity remains a major bug in most appraisal systems. Many such systems are met with considerable resistance by those whose performance is being appraised, thus hampering the possibilities for effectiveness. While accurate and informative appraisal systems can be a major asset to a business, they are too often an unrealized goal.
There are three major steps in the performance appraisal process: identification, measurement, and management. With identification, the behaviors necessary for successful performance are determined. Measurement involves choosing the appropriate instrument for appraisal and assessing performance. Management, which is the ultimate goal, is the reinforcing of good performance and the correction of poor performance. Each step is described below. Additionally, management by objectives, which involves evaluating performance without a traditional performance appraisal, is described.
The organization must determine for each job family the skills and behaviors that are necessary to achieve effective performance. The organization should identify dimensions, which are broad aspects of performance. For instance, "quality of work" is a dimension required in many jobs. To determine which dimensions are important to job performance, the organization should rely on an accurate and up-to-date job analysis. Job descriptions written from job analyses should offer a detailed and valid picture of which job behaviors are necessary for successful performance.
In the identification stage, the company must also choose who will rate employee performance. Supervisors, peers, and the employees themselves may provide performance ratings. In most instances, performance appraisals are the responsibility of the immediate supervisor of an employee. Supervisors rate performance because they are usually the ones most familiar with the employee's work. Additionally, appraisals serve as management tools for supervisors, giving them a means to direct and monitor employee behavior. Indeed, if supervisors are not allowed to make the appraisals, their authority and control over their subordinates could be diminished.
While supervisory ratings can be quite valuable, some companies have added peer appraisals to replace or supplement those given by the supervisor. Naturally, peers and supervisors each view an individual's performance from different perspectives. Supervisors usually possess greater information about job requirements and performance outcomes. On the other hand, peers often see a different, more realistic view of the employee's job performance because people often behave differently when the boss is present. Using peer ratings to supplement supervisory ratings may thus help to develop a consensus about an individual's performance. It may also help eliminate biases and lead to greater employee acceptance of appraisal systems.
Potential problems may limit the usefulness of peer ratings, however, especially if they are used in lieu of supervisory ratings. First, the company must consider the nature of its reward system. If the system is highly competitive, peers may perceive a conflict of interest. High ratings given to a peer may be perceived as harming an individual's own chances for advancement. Second, friendships may influence peer ratings. A peer may fear that low ratings given to a colleague will harm their friendship or hurt the cohesiveness of the work group. On the other hand, some peer ratings may be influenced by a dislike for the employee being rated.
Some organizations use self-ratings to supplement supervisory ratings. As one might expect, self-ratings are generally more favorable than those made by supervisors and peers and therefore may not be effective as an evaluative tool. However, self-ratings may be used for employee development. Their use may uncover areas of subordinate-supervisor disagreement, encourage employees to reflect on their strengths and weaknesses, lead to more constructive appraisal interviews, and make employees more receptive to suggestions.
Once the appropriate performance dimensions have been established for jobs, the organization must determine how best to measure the performance of employees. This raises the critical issue of which rating form to use. In the vast majority of organizations, managers rate employee job performance on a standardized form. A variety of forms exist, but they are not equally effective. To be effective, the form must be relevant and the rating standards must be clear. Relevance refers to the degree to which the rating form includes necessary information, that is, information that indicates the level or merit of a person's job performance. To be relevant, the form must include all the pertinent criteria for evaluating performance and exclude criteria that are irrelevant to job performance.
The omission of pertinent performance criteria is referred to as criterion deficiency. For example, an appraisal form that rates the performance of police officers solely on the basis of the number of arrests made is deficient because it fails to include other aspects of job performance, such as conviction record, court performance, number of commendations, and so on. Such a deficient form may steer employee behavior away from organizational goals; imagine if police officers focused only on arrests and neglected their other important duties.
When irrelevant criteria are included on the rating form, criterion contamination occurs, causing employees to be unfairly evaluated on factors that are irrelevant to the job. For example, criterion contamination would occur if an auto mechanic were evaluated on the basis of personal cleanliness, despite the fact that this characteristic has nothing to do with effective job performance.
Performance standards indicate the level of performance an employee is expected to achieve. Such standards should be clearly defined so that employees know exactly what the company expects of them. For instance, the standard "load a truck within one hour" is much clearer than "work quickly." Not only does the use of clear performance standards help direct employee behavior, it also helps supervisors provide more accurate ratings; two supervisors may disagree on what the term "quickly" means, but both attribute the same meaning to "one hour."
To meet the standards described in the previous section, a firm must use an effective rating form. The form provides the basis for the appraisal, indicating the aspects or dimensions of performance that are to be evaluated and the rating scale for judging that performance. Human Resources (HR) experts have developed a variety of instruments for appraising performance. A description of the most commonly used instruments, along with their strengths and weaknesses, is given in the following paragraphs. A summary of these instruments appears in Exhibit 1. It should be noted, however, that companies can create additional types of instruments. For instance, they can rate employees on job task performance using graphic or behavior rating scales.
Most appraisal instruments require raters to evaluate employees in relation to some standard of excellence. With employee comparison systems, however, employee performance is evaluated relative to the performance of other employees. In other words, employee comparison systems use rankings, rather than ratings. A number of formats can be used to rank employees, such as simple rankings, paired comparisons, or forced distributions. Simple rankings require raters to rank-order their employees from best to worst, according to their job performance. When using the paired comparison
|Implicit personality theory||X|
approach, a rater compares each possible pair of employees. For example, Employee 1 is compared to Employees 2 and 3, and Employee 2 is compared to Employee 3. The employee winning the most "contests" receives the highest ranking. A forced distribution approach requires a rater to assign a certain percentage of employees to each category of excellence, such as best, average, or worst. Forced distribution is analogous to grading on a curve, where a certain percentage of students get As, a certain percentage get Bs, and so forth.
Employee comparison systems are low cost and practical; the ratings take very little time and effort. Moreover, this approach to performance appraisal effectively eliminates some of the rating errors discussed earlier. Leniency is eliminated, for instance, because the rater cannot give every employee an outstanding rating. In fact, by definition, only 50 percent can be rated as being above average. By forcing raters to specify their best and worst performers, employment decisions such as pay raises and promotions become much easier to make.
Employee comparison systems are plagued with several weaknesses. Because the rating standards for judging performance are vague or nonexistent, the accuracy and fairness of the ratings can be seriously questioned. Moreover, employee comparison systems do not specify what a worker must do to receive a good rating and, thus, they fail to adequately direct or monitor employee behavior. Finally, companies using such systems cannot compare the performance of people from different departments fairly. For example, the sixth-ranked employee in Department A may be a better performer than the top-ranked employee in Department B.
A graphic rating scale (GRS) presents appraisers with a list of dimensions, which are aspects of performance that determine an employee's effectiveness. Examples of performance dimensions are cooperativeness, adaptability, maturity, and motivation. Each dimension is accompanied by a multi-point (e.g., 3, 5, or 7) rating scale. The points along the scale are defined by numbers and/or descriptive words or phrases that indicate the level of performance. The midpoint of the scale is usually anchored by such words as "average," "adequate," "satisfactory," or "meets standards."
Many organizations use graphic rating scales because they are easy to use and cost little to develop. HR professionals can develop such forms quickly, and because the dimensions and anchors are written at a general level, a single form is applicable to all or most jobs within an organization. Graphic rating scales do present a number of problems, however. Such scales may not effectively direct behavior; that is, the rating scale does not clearly indicate what a person must do to achieve a given rating, thus employees are left in the dark as to what is expected of them. For instance, an employee given a rating of 2 on "attitude" may have a difficult time figuring out how to improve.
Graphic rating scales also fail to provide a good mechanism for providing specific, non-threatening feedback. Negative feedback should focus on specific behaviors rather than on the vaguely defined dimensions the GRSs describe. For example, if told that they are not dependable, most employees would become angered and defensive; they would become less angry and defensive if such feedback were given in behavioral terms: "Six customers complained to me last week that you did not return their phone calls."
Another problem with GRSs concerns rating accuracy. Accurate ratings are not likely to be achieved because the points on the rating scale are not clearly defined. For instance, two raters may interpret the standard of "average" in very different ways. The failure to clearly define performance standards can lead to a multitude of rating errors (as noted earlier) and provides a ready mechanism for the occurrence of bias. U.S. courts consequently frown on the use of GRSs. One court noted that ratings made on a graphic rating scale amounted to no more than a "subjective judgment call," and ruled that such rating scales should not be used for promotion decisions because of the potential bias inherent in such a subjective process.
A behaviorally-anchored rating scale (BARS), like a graphic rating scale, requires appraisers to rate employees on different performance dimensions. The typical BARS includes seven or eight performance dimensions, each anchored by a multi-point scale. But the rating scales used on BARS are constructed differently than those used on graphic rating scales. Rather than using numbers or adjectives, a BARS anchors each dimension with examples of specific job behaviors that reflect varying levels of performance.
The process for developing a BARS is rather complex. Briefly, it starts with a job analysis, using the critical incident technique. This involves having experts generate a list of critical incidents—or specific examples of poor, average, and excellent behaviors—that are related to a certain job. The incidents are then categorized by dimension. Finally, a rating scale is developed for each dimension, using these behaviors as "anchors" to define points along the scale.
When initially formulated, BARS were expected to be vastly superior to graphic rating scales. HRM experts thought the behavioral anchors would lead to more accurate ratings because they enabled appraisers to better interpret the meaning of the various points along the rating scale. That is, rather than having the rater try to pinpoint the meaning of a vague anchor such as "excellent," the rater would have improved accuracy by having a critical incident as an anchor. As we shall see, however, this expectation has not been met. Perhaps the greatest strength of BARS is its ability to direct and monitor behavior. The behavioral anchors let employees know which types of behavior are expected of them and gives appraisers the opportunity to provide behaviorally-based feedback.
The superiority of BARS over graphic rating scales has not been substantiated by research. In fact, the great majority of studies on this topic have failed to provide evidence that justifyies the tremendous amount of time and effort involved in developing and implementing BARS. The failures of BARS may lie in the difficulty raters experience when trying to select the one behavior on the scale that is most indicative of the employee's performance level. Sometimes an employee may exhibit behaviors at both ends of the scale, so the rater does not know which rating to assign.
A behavior observation scale (BOS) contains a list of desired behaviors required for the successful performance of specific jobs, which are assessed based on the frequency with which they occur. The development BOS, like BARS, also begins with experts generating critical incidents for the jobs in the organization and categorizing these incidents into dimensions. One major difference between BARS and BOS is that, with BOS, each behavior is rated by the appraiser.
When using BOS, an appraiser rates job performance by indicating the frequency with which the employee engages in each behavior. A multi-point scale is used ranging from "almost never" to "almost always." An overall rating is derived by adding the employee's score on each behavioral item. A high score means that an individual frequently engages in desired behaviors, and a low score means that an individual does not often engage in desired behaviors.
Because it was developed more recently, the research on BOS is far less extensive than that on BARS. The available evidence, however, is favorable. One study found that both managers and subordinates preferred appraisals based on BOS to both BARS and graphic rating scales. The same study found that equal employment opportunity attorneys believed BOS is more legally defensible than the other two approaches.
Because raters do not have to choose one behavior most descriptive of an employee's performance level, the problem noted earlier regarding BARS does not arise. Moreover, like BARS, BOS is effective in directing employees' behavior because it specifies what they need to do in order to receive high performance ratings. Managers can also effectively use BOS to monitor behavior and give feedback in specific behavioral terms so that the employees know what they are doing right and which behavior needs to be corrected. Like BARS, however, a BOS instrument takes a great deal of time to develop. Moreover, a separate instrument is needed for each job (since different jobs call for different behaviors), so the method is not always practical. Developing a BOS for a particular job would not be cost-efficient unless the job had many incumbents.
Accurate ratings reflect the employees' actual job performance levels. Employment decisions that are based on inaccurate ratings are not valid and would thus be difficult to justify if legally challenged. Moreover, employees tend to lose their trust in the system when ratings do not accurately reflect their performance levels, and this causes morale and turnover problems. Unfortunately, accurate ratings seem to be rare. Inaccuracy is most often attributable to the presence of rater errors, such as leniency, severity, central tendency, halo, and recency errors. These rating errors occur because of problems with human judgment. Typically, raters do not consciously choose to make these errors, and they may not even recognize when they do make them.
Leniency error occurs when individuals are given ratings that are higher than actual performance warrants. Leniency errors most often occur when performance standards are vaguely defined. That is, an individual who has not earned an excellent rating is most likely to receive one when "excellent" is not clearly defined. Why do appraisers distort their ratings in an upward or downward direction? Some do it for political reasons; that is, they manipulate the ratings to enhance or protect their self-interests. In other instances, leniency and severity come about from a rater's lack of conscientiousness. Raters may allow personal feelings to affect their judgments; a lenient rating may be given simply because the rater likes the employee.
Severity error occurs when individuals are given ratings that are lower than actual performance warrants. Severe ratings may be assigned out of a dislike for an individual, perhaps due to personal bias. A male appraiser may, for example, underrate a highly-performing female employee because she threatens his self-esteem; a disabled employee may receive an unduly low rating because the employee's presence makes the appraiser feel embarrassed and tense; or an appraiser may provide harsh ratings to minorities out of a fear and distrust of people with different nationalities or skin color. Alternately, a severe rating may be due to the very high standards of a rater, or to "send a message" to motivate employees to improve.
When raters make leniency and severity errors, a firm is unable to provide its employees with useful feedback regarding their performance. An employee who receives a lenient rating may be lulled into thinking that performance improvement is unnecessary. Severity errors, on the other hand, can create morale and motivation problems and possibly lead to discrimination lawsuits.
Central tendency error occurs when appraisers purposely avoid giving extreme ratings even when such ratings are warranted. For example, when rating subordinates on a scale that ranges from one to five, an appraiser would avoid giving any ones or fives. When this error occurs, all employees end up being rated as average or near average, and the employer is thus unable to discern who its best and worst performers are. Central tendency error is likely the result of administrative procedures. That is, it frequently occurs when an organization requires appraisers to provide extensive documentation to support extreme ratings. The extra paperwork often discourages appraisers from assigning high or low ratings. Central tendency errors also occur when the end points of the rating scale are unrealistically defined (e.g., a 5 effectively means "the employee can walk on water" and a 1 means "the employee would drown in a puddle").
Appraisals are also subject to the halo effect, which occurs when an appraiser's overall impression of an employee is based on a particular characteristic, such as intelligence or appearance. When rating each aspect of an employee's work, the rater may be unduly influenced by his or her overall impression. For example, a rater who is impressed by an employee's intelligence may overlook some deficiencies and give that employee all fives on a one-to-five scale; an employee perceived to be of average intelligence may be given all threes. The halo effect acts as a barrier to accurate appraisals because those guilty of it fail to identify the specific strengths and weaknesses of their employees. It occurs most often when the rating standards are vague and the rater fails to conscientiously complete the rating form. For instance, the rater may simply go down the form checking all fives or all threes.
Most organizations require that employee performance be assessed once a year. When rating an employee on a particular characteristic, a rater may be unable to recall all of the employee's pertinent job behaviors that took place during that rating period. The failure to recall such information is called memory decay. The usual consequence of memory decay is the occurrence of recency error; that is, ratings are heavily influenced by recent events that are more easily remembered. Ratings that unduly reflect recent events can present a false picture of the individual's job performance during the entire rating period. For instance, the employee may have received a poor rating because he or she performed poorly during the most recent month, despite an excellent performance during the preceding eleven months.
In the management phase of performance appraisal, employees are given feedback about their performance and that performance is either reinforced or modified. The feedback is typically given in an appraisal interview, in which a manager formally addresses the results of the performance appraisal with the employee. Ideally, the employee will be able to understand his or her performance deficiencies and can ask questions about the appraisal and his or her future performance. The manager should give feedback in a way that it will be heard and accepted by the employee; otherwise, the appraisal interview may not be effective.
The appraisal interview may also have an appeals process, in which an employee can rebut or challenge the appraisal if he or she feels that it is inaccurate or unfair. Such a system is beneficial because it:
The downside of using an appeals system is that it tends to undermine the authority of the supervisor and may encourage leniency error. For example, a supervisor may give lenient ratings to avoid going through the hassle of an appeal.
Management by objectives (MBO) is a management system designed to achieve organizational effectiveness by steering each employee's behavior toward the organization's mission. MBO is often used in place of traditional performance appraisals. The MBO process includes goal setting, planning, and evaluation. Goal setting starts at the top of the organization with the establishment of the organization's mission statement and strategic goals. The goal-setting process then cascades down through the organizational hierarchy to the level of the individual employee. An individual's goals should represent outcomes that, if achieved, would most contribute to the attainment of the organization's strategic goals. In most instances, individual goals are mutually set by employees and their supervisors, at which time they also set specific performance standards and determine how goal attainment will be measured.
As they plan, employees and supervisors work together to identify potential obstacles to reaching goals and devise strategies to overcome these obstacles. The two parties periodically meet to discuss the employee's progress to date and to identify any changes in goals necessitated by organizational circumstances. In the evaluation phase, the employee's success at meeting goals is evaluated against the agreed-on performance standards. The final evaluation, occurring annually in most cases, serves as a measure of the employee's performance effectiveness.
MBO is widely practiced throughout the United States. The research evaluating its effectiveness as a performance appraisal tool has been quite favorable. These findings suggest that the MBO improves job performance by monitoring and directing behavior; that is, it serves as an effective feedback device, and it lets people know what is expected of them so that they can spend their time and energy in ways that maximize the attainment of important organizational objectives. Research further suggests that employees perform best when goals are specific and challenging, when workers are provided with feedback on goal attainment, and when they are rewarded for accomplishing the goal.
MBO presents several potential problems, however, five of which are addressed here.
Lawrence S. Kleiman
Revised by Marcia J. Simmering
Gomez-Mejia, Luis R., David B. Balkin, and Robert L. Cardy. Managing Human Resources. 4th ed. Upper Saddle River, NJ: Prentice Hall, 2004.
Grote, Richard C. The Performance Appraisal Question and Answer Book: A Survival Guide for Managers. New York: AMACOM Books, 2002.
Kleiman, L.S. Human Resource Management: A Tool for Competitive Advantage. Cincinnati: South-Western College Publishing, 2000.
Latham, G.P., and K.N. Wexley. Increasing Productivity Through Performance Appraisal. 2nd ed. Reading, MA: Addison-Wesley, 1994.
Noe, Raymond A., John R. Hollenbeck, Barry Gerhart, and Patrick M. Wright. Human Resource Management: Gaining a Competitive Advantage. 5th ed. Boston: McGraw-Hill/Irwin, 2006.