There is a growing debate about the relevance of employee performance appraisals. On the one hand, performance ratings are considered by many academics and practicing managers as essential personnel management tools. They are used in recruiting and hiring, in compensation administration, in training and development, and as a motivational tool. On the other hand, a sizeable number of business managers believe that this practice may generate more problems than it solves. For one thing, the design of performance appraisal systems is usually flawed. A recent survey of Fortune 100 companies shows that practicing managers and performance appraisal researchers have very different concerns. For instance, the former tend to place more importance on the appraisals' processes and behaviors, while the latter attaches more value on its cognitive aspects.
Many academics and practicing managers regard performance appraisal as one of the most valuable human resource tools. It is a vital component in recruiting and hiring employees, where it is used to validate selection tests, and in staffing, where transfer, layoff, termination, or promotion decisions are made on the basis of appraisal results. In compensation administration, performance appraisal forms the basis for the administration of merit pay systems. Most important, performance appraisal can be used as a motivational tool for communicating performance expectations to employees and providing them with feedback. Finally, performance appraisal is indispensable in training and development activities to assess potential and identify training needs.
On the other hand, there appears to be a growing debate about whether the consequences of the performance appraisal are truly beneficial to many organizations. A significant number of practicing managers appear to be saying that performance appraisal may create more problems than it solves. W. Edward Deming, for example, has proposed that we should abolish production standards that specify numerical goals and eliminate all individual performance appraisal in favor of systems that evaluate performance at the unit or plant level.(1)
Regardless of one's perspective, performance appraisal systems are likely to be a subject of concern for managers and employees alike for some time to come. In fact, the trend in organizations appears to be toward merit or other performance-based pay plans, promising even more emphasis on the appraisal process. Despite the trend, and even though a stream of appraisal research has flowed unabated for years, performance appraisal, as commonly practiced, has remained a largely unsatisfactory endeavor. Performance appraisal systems often suffer from design flaws. Managers receive poor training in appraisal administration and are seldom rewarded for accuracy in appraisal. In addition, for reasons summarized in Table 1, both managers and employees tend to approach appraisal feedback sessions with fear and loathing.
Table 1 Why Managers and Employees Dislike the Appraisal Process
1. Neither rarely has any sense of ownership. They are not involved in the design or the administration of the system; they frequently are not trained to use the system, and their reactions to the system are seldom solicited and acted upon.
2. Managers do not like to deliver negative messages to people with whom they must work, and whom they often like on a personal basis; and employees do not like to receice them. Negative messages tend to generate defensive reactions and promote hostility rather than serve as performance feedback.
3. Both managers and employees recognize that delivering a negative message will adversely affect a person's career. Managers may be aware of the permanence of the "paper trail" that follows formal appraisal and are often hestitant to commit negative feedback to writing even when they do not like the individual.
4. There are few formal rewards for taking the appraisal process seriously and probably no informal rewards. There are many informal rewards for not delivering unpopular messages.
5. Managers hestitate to give unfavorable appraisals for fear that the appearance of unsatisfactory work by subordinates reflects badly on the manager's ability to select and develop employees. Lack of candor in employee evaluation is one way of "hiding dirty laundry."
A new detailed survey of Fortune 100 firms conducted by Cornell University researchers, Bob Bretz and George Milkovich, provides some perspective on the performance appraisal enigma. The survey reveals a great deal about the format, design, and process of performance appraisal as currently practiced by the country's largest employers. The results, stacked up against recent performance appraisal research, suggest that the interests of the practicing managers and those of the researchers have surprisingly little in common.
Two fundamentally different perspectives have served to create a gap between research and practice. Managers, for example, tend to focus on the processes and behaviors (fairness and usefulness) of performance appraisal, while researchers are more concerned with the cognitive aspects of the rating process. While research, on one hand, has done little to improve the usefulness of performance appraisal as a management tool, organizations, on the other hand, have been guilty of ignoring research findings that have potential to improve the appraisal process.
Performance Appraisal in Practice
The Bretz and Milkovich survey was designed to examine the specific performance appraisal practices of large American companies as a first step toward understanding why performance appraisal research has been of marginal benefit to the practitioner. The study was a detailed 20 page questionnaire soliciting responses from those in policy-making positions in Fortune 100 companies.(2) Information was requested on the appraisal systems design, characteristics, management results and the use of the results.
Performance Appraisal System Design
The survey results revealed that performance appraisal systems in large companies are designed primarily by personnel specialists. Despite recent interest in participative management and other employee involvement schemes, there appears to be little input into performance appraisal design by line managers and almost none by the employees who are the subjects of appraisal. While employees are often expected and encouraged to participate in decisions relating to production or operations as part of a total quality management system, they rarely are allowed any meaningful input into the performance appraisal system that directly affects them.
Characteristics of Performance Appraisal Systems
The dominant format for performance appraisal systems in large U.S. industrial companies, at least for executives, managers, and professional employees, continues to be an objective-based approach such as management by objectives (MBO). As Table 2 illustrates, 80% of the surveyed organizations report using objectivebased approaches for their executives and managers, and 70% use them for professional employees. Little use is made of the traditional or graphic rating scale or other systems for these exempt employees.
Table 2 Performance Appraisal Formats in Fortune 100 Firms (Percent using
format for employee group)
Management Graphic Behavioral other by Rating Anchored Formats Objective Scale Scales Manager 80 16 NA 4 (Exempt) Professional 70 23 NA 7 (Exempt) Non-Exempt 31 32 14 23
By contrast, nonexempt employees in large organizations are equally likely to be evaluated using MBO or a graphic scale (31% and 32% respectively). A few organizations use forced choice formats (7%), while the rest use a handful of other techniques or simply do not evaluate nonexempt employees. Surprisingly, when only hourly paid employees are considered, more than half (52%) of responding organizations do not conduct formal appraisals of any kind on this group.
Many organizations use forced distributions or ranking, particularly in conjunction with pay decisions, to supplement their main performance appraisal method, especially for managers and professional employees. About one-fourth of reporting organizations use forced distributions and ranking for managers and professionals, but only about 10% do so for either executives or nonexempt employees. As might be expected, given the prevalence of MBO, quantitative or objective measures of performance such as sales, costs, profits, or knowledge acquisition are frequently used to supply performance information.
Who decides the final performance appraisal rating in these large organizations? As shown in Table 3, the employee's immediate supervisor is the key evaluator of performance. His or her opinion provides one-half to three-fourths of the weight that determines the final appraisal. Second-level managers have significant impact on appraisal results for managerial and professional employees (their opinion carries 25% to 43% of the weight); but they provide little input on nonexempt employee appraisals.
Table 3 Relative Importance of Rating Sources and Information (weights
assigned to each source)
Immediate Higher Self Peer Other Supervisor Level Supervisor Executives 42 36 5 2 15 Managers 47 43 5 1 4 Professional 67 25 5 1 2 Non-Exempt 74 18 5 0 3
Some organizations actually do provide information from self and peer appraisals. This information, however, tends to be heavily discounted by practicing managers. Less than 5% of the information actually used to determine the final performance rating comes from such evaluations. Despite theoretical and practical arguments that peer and self appraisal can be useful, and despite claims in the press that this form of appraisal is "mushrooming,"(3) little evidence was found that it is used to any significant extent, nor that it significantly influences performance ratings, in these large industrial organizations.
Management of the Performance Appraisal System
Decisions regarding the strategic nature of performance appraisal (such as whether or not to actually have appraisals, or what the role of merit pay or incentives should be) are made at the corporate level in over two-thirds of these large companies. The remainder make these decisions at business unit level. Slightly more than half, however, make decisions about actual practices (such as which format to use or what type of training to undertake) at the business unit level. Consistent with the extensive use of MBO, exempt employees do tend to participate in setting their own performance standards. While executives, managers, and professional employees in over 80% of the firms are involved in setting their own objectives, only 16% of firms allow hourly employees to do so.
Despite the assumption that performance appraisal is valuable, relatively little time is expended on the activity. An average of eight hours per employee per year is spent on appraising executives and managers, six on professional employees, and less than four on nonexempts (these figures include observing and documenting performance, completing the actual evaluation, and conducting employee feedback sessions). Considering that the averages are inflated due to a handful of firms that spend considerably more time on appraisal, the typical firm devotes surprisingly little time examining employee performance and providing employee feedback.
Why isn't more time spent on appraisal? One reason is simply that managers are not commonly held accountable for how well they conduct performance appraisal on their subordinates. In only 22% of the Fortune 100 sample managers were evaluated on how well they conduct performance appraisals. Basic motivational theory as well as common sense suggests that managers will devote little effort to a somewhat unpleasant chore for which they are not held accountable.
Most organizations do conduct managerial training in how to use the performance appraisal system. Training programs reported included conducting appraisal interviews and providing feedback (in 90% of the organizations), instruction in using the appraisal forms (in 83%), setting performance standards (78%), recognizing good performance (66%), and avoiding rating errors (56%). Most of this training tends to occur at the initial development of the appraisal system however. Few companies conduct ongoing training. Since most of the appraisal systems are more than ten years old, most managers have not benefitted from recent training. In addition, almost all training focuses on the manager. Employees receive no training on how to use feedback and appraisal information to improve performance.
Performance Appraisal Outcomes
Organizations appear to be quite concerned with the perceived fairness of their appraisal systems. Employees may contest appraisal results through a formal appeals process in 26% of organizations, or through an informal appeals process in 64%. Only 10% report no appeals procedure.
When asked about their major concerns regarding performance appraisal, organizations consistently identified issues relating to fairness. The three most important issues are: (1) the acceptance of the appraisal system by those being rated; (2) whether employees believe the process is fair; and (3) whether employees believe the results are fair. Ironically, however, only one-third of the organizations in the sample conduct attitude surveys of either managers or employees to assess their perceptions of the fairness of appraisal processes or outcomes.
The next two most important issues expressed as areas of concern in performance appraisal deal with uncertainty regarding the type of feedback given and the degree to which the appraisal system is a useful tool for managing performance. Issues of accuracy in appraisal or appraisal error -- issues that are common topics of scholarly research -- are not considered important issues by the respondents.
How do America's largest companies use performance appraisal information? The most frequent uses, as seen in Table 4, are for employee development and administering merit pay. (The five main development uses include communicating work expectations, improving performance, determining employee potential, and counseling and developing employees.) Performance appraisal information is much less likely to be used to make decisions relating to layoffs, terminations, or transfers. Interestingly, feedback from respondents reveals that many have more faith in performance rankings than performance ratings for this type of decision, because the latter simply are not perceived to be sufficiently accurate.
Table 4 Important Uses for Performance Appraisal (rank order)
1. Improving work performance
2. Administering merit pay
3. Advising employees of work expectations
4. Counseling employees
5. Making promotion decisions
6. Motivating employees
7. Assessing employee potential
8. Identifying training needs
9. Better working relationships
10. Helping employees set career goals
11. Assigning work more efficiently
12. Making transfer decisions
13. Making decisions about layoffs and terminations
14. Assisting in long-range planning
15. Validating hiring procedures
16. Justifying other managerial actions
Ninety-two percent of the firms use appraisal information in determining merit pay increases. One-third of the organizations, however, also rank managerial and professional employees for purposes of assigning pay increases.
Do organizations feel that their performance appraisal system accomplishes merit pay objectives? The answer is equivocal. On a five-point scale (1 = very well, 3 = somewhat well, 5 = not at all) the average rating is a 2.5. While this could be interpreted that the performance appraisal system is accomplishing its merit pay objectives "fairly well," it is hardly a ringing endorsement. While 62% of the organizations claims that the pay-for-performance system does provide sufficient differentials between high and low performers, more than three-fourths report that skewed performance distributions are a problem; and over two-thirds state that the skewed distributions affect pay administration and the ability to reward the best performers.
Skewed performance distributions are, indeed, evident in the actual distributions of performance ratings. Most companies (60%) have appraisal systems that describe five performance levels, while 20% use more and 20% use less. What quickly becomes evident in the 80% of companies that have five or more levels is that only the highest three levels are actually used. Table 5 shows that only 2% of executives, 8% of managers, 9% of professional employees, and 7% of nonexempts are rated in the bottom two levels. Clearly, the norm has a "leniency bias," where employees are rated at the top end of the scale. This appears to hold even in organizations that use forced distributions. Interestingly, the 23% of organizations that do not feel that skewed performance distributions are a problem report even higher proportions of employees in the top performance levels than the 77% of organizations that recognize a skewing problem. It is evident that even though leniency exists, it is not perceived to be a problem by some managers.
Table 5 Performance Distributions in Fortune 100 Companies
Level 1 Level 2 Level 3 Level 4 Level 5 Executives 23 46 31 2 0 Managers 14 43 33 6 2 Professionals 13 43 34 8 1 Non-Exempts 14 44 34 6 1
1. Far exceeds objectives
2. Exceeds objectives
3. Fully meets objectives
4. Partially meets objectives
Research in Performance Appraisal
Comprehensive reviews on performance appraisal abound in the academic literature.(4) What is clear from the research conducted over the last five years is that the topics that occupy the time of academic researchers are not necessarily those that command the attention of practicing managers. A recent review of some 153 articles(5) published between 1985 and 1989 found that most of the research is conducted in the laboratory using student subjects, paper people, or video tapes. The issues addressed most frequently relate to cognitive processing, rater/ratee characteristics, and psychometric errors.
Currently, research on the impact of different sources of rater information, on rater training, and on appraisal format is limited. These are issues, however, that have been subject to research in the recent past and for which further investigation appears to be unnecessary. Research relating to issues of fairness, the consequences of performance appraisal on employee attitudes and behaviors, and the uses of appraisal information is virtually nonexistent even though these appear to be major concerns in the practitioner literature and to the practicing managers in the survey.
It is clear from the foregoing that performance appraisal continues to present a vexing human resource challenge. The Fortune 100 survey demonstrates the gap between the concerns of practicing managers and performance appraisal research. Practicing managers are quite concerned about issues of fairness in appraisal, but do not consider appraisal accuracy, rating errors, or an understanding of the cognitive processes used in the appraisal process to be major organizational concerns.
This is unfortunate, since much of the fairness issue can be addressed by current research. The cognitive process research is really about fairness and may eventually lead to fairer ratings, because of its attempts to control the effect of bias. One problem for managers is trying to understand and interpret research that does not seem realistic because it is done in the lab or based on student subjects. What is needed is a transition into the organizational environments that managers understand. However, the lack of access to organizational settings continues to hamper performance appraisal research. The best way to improve fairness may be for managers to facilitate the researcher's access to their organizations.
Nevertheless, research does begin to offer some suggestions for improving the effectiveness of the appraisal process. The following five points stress that more satisfactory performance appraisal results may be attained through employee participation in system design, by devoting additional resources to training, and by generally creating an organizational culture supportive of performance appraisal:
1. Get employees more involved in the design,
development, and administration of the
performance appraisal system. Participation
creates ego involvement and a sense of
commitment to the process.
2. Invest more heavily in training raters to use
the system. Train managers not only to
observe and document performance but also
to communicate information effectively and
deliver performance feedback.
3. Create an environment in which performance
information is viewed as a resource that
managers can use to develop subordinates.
Top managers must create a climate in which
accurate and timely performance appraisal is
expected of all managers, is taken seriously,
and is rewarded.
4. Make performance appraisal the responsibil-
ity of the ratee, not the rater. This is a
fundamental philosophical shift that takes the
burden to "be nice" from managers and frees
them to honestly "call it as they see it." As
part of this philosophy, employees must be
trained to use the feedback from the appraisal
process to manage their own careers.
5. Use multiple perspectives (multiple raters)
including peer evaluation to reduce the
reliance on a single source. This reduces
sampling error by increasing the number of
observations and reduces the effect of
possible idiosyncratic biases. Raters are more
comfortable, since they are no longer solely
responsible for what happens to the person as
a result of the rating.
Still, many real organizational concerns -- how performance appraisal is perceived by raters and ratees and the effects of different system designs or processes on employee attitudes and organizational performance -- have not been properly studied. Although scholarly research does not exist to serve only the needs of the practicing manager, these and similar questions must be addressed before performance appraisal can begin to silence its critics and fulfill its potential for making organizations, and the people in them, more effective.
(1.)See W. Edwards Deming, Quality, Productivity, and Competitive Position (Cambridge, MA: Center for Advanced Engineering Study, MIT, 1982).
(2.)The response rate from the survey was quite high at 70%. The conclusion made about these large organizations, however, may not generalize to U.S. industry as a whole as these Fortune 100 firms are arguably more likely to have formal appraisal systems than are smaller firms.
(3.)See Kiechell in Fortune, June 19, 1989, p. 201.
(4.)See Robert D. Bretz and George T. Milkovich, Performance Appraisal in Large Organizations: Practice and Implications, Center for Advanced Human Resource Studies, Cornell University School of Industrial and Labor Relations, Working Paper #89-17 for a complete review of performance appraisal research.
(5.)Bretz and Milkovich, above.