C. Strategy 2: Select the Right Raters
A second way that firms have tried to improve the accuracy of
performance ratings is by focusing on who might conduct the ratings and
which of these sources is more likely to be accurate.
To lessen the impact of one reviewer, and to increase participation in the
process, a method known as 360-degree feedback has grown popular.
oThis method assesses employee performance from five points of
view: supervisor, peer, self, customer, and subordinate.
oThe flexibility of the process makes it appealing to employees at
all levels within an organization; most companies using the system
report that their employees are satisfied with its results.
oRegardless of the positive responses from those who have
implemented the 360-degree feedback system, today most companies
still use it only for evaluation of their top-level personnel and for
employee development rather than for appraisals or pay decisions.
oSome research indicates that the different types of raters are very
similar in their evaluations. Given the extra costs, why ask for multiple
views the critics argue.
oResponding to this criticism, recent research shows a distinct
impact of rating source-raters do provide differing views.
Supervisors as Raters
oSome estimates indicate that more than 80 percent of the input
for performance ratings comes from supervisors.
oSupervisors assign what work employees are to perform, which
makes them knowledgeable about the job and the dimensions to be
rated. Also, supervisors have considerable prior experience in rating
employees, thus giving them some pretty firm ideas about what level of
performance is required for any given level of performance rating.
oOn the negative side, supervisors are particularly prone to halo
and leniency errors.
Peers as Raters
oOne of the major strengths of using peers as raters is that they
work more closely with the ratee and probably have an undistorted
perspective of typical performance, particularly in group assignments.
Balanced against this positive are at least two powerful negatives:
Peers may have little or no experience in conducting appraisals,
leading to rather mixed evidence about the reliability of this rating
source.
In a situation where teamwork is promoted, placing the burden of
rating peers on co-workers can either create group tensions or yield
ratings second only to self-ratings in level of leniency. One
exception to this leniency effect comes from top performers, who it
seems give the most objective evaluations of peers.
Self as Rater
oSome organizations have experimented with self-ratings. Self-ratings are
done by someone who has the most complete knowledge about the ratee’s
performance.
Unfortunately, though, self-ratings are generally more lenient
and possibly more unreliable than ratings from other sources One
compromise in the use of self-ratings is to use them for developmental
rather than administrative purposes.
Customer as Rater
oThis is the era of the customer. The drive for quality means more
companies are recognizing the importance of customers.
oCompanies either survey customers or hire mystery customers to pretend
like a customer, observe the processes, and rate the customer service
process.
Subordinate as Rater
oThe notion of subordinates as raters is appealing since most employees
want to be successful with the people who report to them. Hearing how
they are viewed by their subordinates gives them the chance to both see
their strengths and their weaknesses as a leader and to modify their
behavior.
oThe difficulty with this type of rating is in attaining candid reviews and
also in counseling the ratee on how to deal with the feedback.
oResearch shows that subordinates prefer to give their feedback to
managers anonymously. If their identity is known they give artificially
inflated ratings of their supervisors.
D. Strategy 3: Understand How Raters Process Information
A third way to improve performance ratings is to understand how raters
think—what else influences ratings besides an employee’s performance?
Research exploring how raters process information about the
performance of the people indicates the following kinds of processes occur.
oThe rater observes the behavior of a ratee.
oThe rater encodes this behavior as part of a total picture of the ratee, i.e.,
the rater forms stereotypes.
oThe rater stores this information in memory, which is subject to both
short-term and long-term decay.
oWhen it comes time to evaluate a ratee, the rater reviews the performance
dimensions and retrieves stored observations/impressions to determine
their relevance to the performance dimensions.
oThe information is reconsidered and integrated with other available
information as the rater decides on the final ratings.
Quite unintentionally, this process can produce errors, and they can
occur at any stage.
Errors in the Rating Process
oStudies show that performance actually does play an important role,
perhaps the major role, in determining how a supervisor rates a
subordinate.
oEmployees who are technically proficient and who do not create problems
on the job tend to receive higher ratings than these who are weaker on
these dimensions.
oOn the negative side, performance-irrelevant factors appear to influence
ratings, and they can cause errors in the evaluation process.
Common Errors in Appraising Performance: Criterion Contamination
oCriterion contamination, or allowing non-performance factors to affect
performance scores, occurs in every company and every job, and probably
affects everyone sometime during their careers.
oOne survey of 1,816 organizations reported that only 4.6 percent of the
managers were rated below average.
oThe truth is that raters tend to make mistakes. Recognizing and
understanding the errors, such as the nine most common noted in Exhibit
11.10, is the first step to communicating and building a more effective
appraisal process.
oThe potential for errors causes employees to lose faith in the performance
appraisal process. Employees, quite naturally, will be reluctant to have pay
systems tied to such error-ridden performance ratings.
oThere are several factors that lead raters to give inaccurate appraisals:
Guilt
Embarrassment about giving praise
Taking things for granted
Not noticing good or poor performance
The halo effect (seeing one good attribute and leaping to positive
impressions on remaining attributes)
Dislike of confrontation
Spending too little time on preparation of the appraisal
oCompanies and researchers alike have expended considerable time and
money to identify ways job performance can be measured better.
Errors in Observation (Attention)
oGenerally, researchers have varied three types of input information to see
what raters pay attention to when they are collecting information for
performance appraisals.
oFirst, it appears that raters are influenced by general appearance
characteristics of the ratees. Males are rated higher than females (other
things being equal).
oRace also matters in performance ratings. Both in layoff decisions and in
performance ratings, blacks are more likely to do worse than whites.
oResearchers also look at change in performance over time to see if this
influences performance ratings. Both the pattern of performance
(performance gets better versus worse over time) and the variability of
performance (consistent versus erratic) influence performance ratings,
even when the overall level (average) of performance is controlled.
oWorkers who start out high in performance and then get worse are rated
lower than workers who remain consistently low. Not surprisingly,
workers whose performance improves over time are seen as more
motivated, while those who are more variable in their performance are
tagged as lower in motivation.
Errors in Storage and Recall
oResearch suggests that raters store information in the form of traits. More
importantly, they tend to recall information in the form of trait categories.
oFor example, a rater observes a specific behavior such as an employee
resting during work hours. The rater stores this information not as the
specific behavior but rather in the form of a trait, such as “That worker is
lazy.”
oThe entire rating process may be heavily influenced by the trait categories
that the rater adopts, regardless of their accuracy.
oErrors in storage and recall also appear to arise from memory decay. At
least one study indicates that rating accuracy is a function of the delay
between performance and subsequent rating.
Errors in the Actual Evaluation
oThe context of the actual evaluation process also can influence
evaluations. Several researchers indicate that the purpose of an evaluation
affects the rating process.
oFor example, performance appraisals sometimes serve a political end.
Supervisors have been known to deflate performance to send a signal to an
employee—“You’re not wanted here.”
oSupervisors also tend to weigh negative attributes more heavily than
positive attributes.
oIf the purpose of evaluation is to divide up a fixed pot of merit increases,
ratings also tend to be less accurate. Supervisors who know ratings will be
used to determine merit increases are less likely to differentiate among
subordinates than they are when the ratings will be used for other
purposes.
oAlso, being required to provide feedback to subordinates about their
ratings yields less accuracy than a secrecy policy. Presumably, anticipation
of an unpleasant confrontation with the angry ratee persuades the rater to
avoid confrontation by giving a rating higher than is justified.
oHowever, when raters must justify their scoring of subordinates in writing,
the rating is more accurate.
E. Strategy 4: Training Raters to Rate More Accurately
Most research indicates rater training is an effective method for reducing
appraisal errors.
Rater training programs can be divided into three distinct categories.
oRater-error training—goal is to reduce psychometric errors (leniency,
severity, central tendency, halo), by familiarizing raters with their
existence.
oPerformance-dimension training—exposes supervisors to the
performance dimensions to be used in rating (e.g., quality of work, job
knowledge), thus making sure everyone is on the same page when
thinking about a specific performance dimension.
oPerformance-standard training—provides raters with a standard of
comparison or frame of reference for making appraisals (what constitutes
good, average, and bad).
Several generalizations about ways to improve rater training can be
summarized:
oStraightforward lecturing to ratees about ways to improve the quality of
their ratings generally is ineffective.
oIndividualized or small-group discussion sections are more effective in
conveying proper rating procedures.
oWhen these sessions are combined with extensive practice and feedback
sessions, rating accuracy significantly improves.
oLonger training programs (more than two hours) generally are more
successful than shorter programs.
oPerformance-dimension training and performance-standard training
generally work better than rater-error training, particularly when they are
combined.
oThe greatest success has come from efforts to reduce halo errors and
improve accuracy.
Leniency errors are the most difficult form of error to eliminate.
oThink about the consequences to a supervisor of giving inflated ratings
versus those of giving accurate or even deflated ratings. The latter two
courses are certain to result in more complaints and possibly reduced
employee morale.
oThe easy way out is to artificially inflate ratings. Unfortunately, this
positive outcome for supervisors may come back to haunt them: With
everyone receiving relatively high ratings there is less distinction between
truly good and poor performers.
III. Putting It All Together: The Performance Evaluation Process
First, some of the key elements in the total process that from day one make for a
good appraisal outcome are outlined below:
oFigure out what your culture and strategy are. This should help figure out the key
factors for measure.
oMake sure job descriptions are up to date and employees know what is expected
of them.
oPeriodically have informal discussions with employees about progress and any
barriers that need removed.
oWhen evaluating performance, focus on behavior, not the person.
Second, employees need to be involved in every stage of developing
performance dimensions and building scales to measure how well they perform on
these dimensions.
oIn cases where this occurs, employees have more positive reactions to ratings,
regardless of how well they do. They are happier with the system’s fairness and
the appraisal accuracy.
oManagers also respond well to this type of ‘due process’ system. They have
higher job satisfaction and less reason to distort appraisal results to further their
own interests.
Third, raters should be trained in the use of the appraisal system and all
employees should understand how the system operates and what it will be used for.
Fourth, raters should be motivated to rate accurately.
oOne way to achieve this is to ensure that managers are rated on how well they
utilize and develop human resources. A big part of this would be evaluation and
feedback to employees.
Fifth, raters should maintain a diary of employee performance, both as
documentation and to jog memory.
Sixth, raters should conduct a performance diagnosis to determine in advance if
performance problems arise because of motivation, skill deficiency, or external
environmental constraints.
Finally, feedback to employees should be timely.
IV. Equal Employment Opportunity and Performance Evaluation
Equal employment opportunity (EEO) and affirmative action have
influenced HR decision making for more than 40 years. While there are certainly
critics of these programs, at least one important trend can be traced to the civil rights
vigil in the workplace.
oEEO has forced organizations to document decisions and to ensure they are firmly
tied to performance or expected performance.
Performance appraisals are subject to the same scrutiny as employment tests.
This interpretation of performance evaluation as a test, subject to validation
requirements, was made in Brito v. Zia Company.
oZia Company used performance evaluations based on a rating format to lay off
employees.
oThe layoffs resulted in a disproportionate number of minorities being discharged.
oThe court held that Zia, a government contractor, had failed to comply with the
testing guidelines issued by the Secretary of Labor, and that Zia had not
developed job-related criteria for evaluating employees’ work performance to be
used in determining employment promotion and discharges which is required to
protect minority group applicants and employees from the discriminatory effects
of such failure.
The courts stress six issues in setting up a performance appraisal system.
oCourts are favorably disposed to appraisal systems that give specific written
instructions on how to complete the appraisal. Presumably, more extensive
training in other facets of evaluation would also be viewed favorably by the
courts.
oOrganizations tend to be able to support their cases better when the appraisal
system incorporates clear criteria for evaluating performance. Performance
dimensions and scale levels that are written, objective, and clear tend to be
viewed positively by courts in discrimination suits. In part, this probably arises
because behaviorally oriented appraisals have more potential to provide workers
feedback about developmental needs.
oThe presence of adequately developed job descriptions provides a rational
foundation for personnel decisions. The courts reinforce this by ruling more
consistently for defendants (companies) when their appraisal systems are based
on sound job descriptions.
oCourts also approve of appraisal systems that require supervisors to provide
feedback about appraisal results to the employees affected. Absence of secrecy
permits employees to identify weaknesses and to challenge undeserved appraisals.
oThe courts seem to like evaluation systems that incorporate a review of any
performance rating by a higher-level supervisor.
oThe courts consistently suggest that the key to fair appraisals depends on
consistent treatment across raters, regardless of race, color, religion, sex, or
national origin.
The focal question then becomes: Are similarly situated individuals treated
similarly? This standard is particularly evident in a court case involving performance
appraisal and merit pay.
oA black male filed suit against General Motors, claiming race discrimination in
both the timing and the amount of a merit increase.
oThe court found this case without merit.
oGeneral Motors was able to show that the same set of rules was applied equally to
all individuals.
Experts note that firms approaching performance appraisal primarily as a way to
defend against discrimination claims may actually create more claims.
oDocumentation of performance to discourage claims only causes poor employee
relations.
oA better strategy is to follow the guidelines discussed earlier, as they permit both
effective performance reviews and a strong foundation in case legal issues arise.
V. Tying Pay to Subjectively Appraised Performance
The central issue involving merit pay is, “How do we get employees to view
raises as a reward for performance?”
From a pragmatic perspective, organizations frequently grant increases that are
not designed or communicated to be related to performance.
oMany companies view raises not as motivational tools to shape behavior but as
budgetary line items to control costs. Frequently this results in pay increase
guidelines with little motivational impact.
Two types of pay increase guidelines with low-motivation potential provide
equal increases to all employees regardless of performance.
oGeneral increases—provide equal increase to all employees regardless of
performance. These increases are typically found in unionized firms. A contract is
negotiated that specifies an across-the-board, equal increase for each year of the
contract.
oAcross-the-board increases—provide equal increase to all employees regardless
of performance. These increases are often linked to cost-of-living changes. When
the Consumer Price Index (CPI) rises, some companies adjust base pay for all
employees to reflect the rising costs.
oThe third form of guideline is Seniority increases—comes somewhat close to
tying pay to performance. These increases tie pay increases to a preset
progression pattern based on seniority.
In practice, tying pay to performance requires three things:
oFirst, a company needs some definition of performance.
Merit increases may be linked to employee ability and willingness to
demonstrate key competencies. For example, showing more of the following
behaviors might be tied to higher merit increases:
Competency: Customer Care
1. Follows through on commitments to customers in a timely manner
2. Defines and communicates customer requirements
3. Resolves customer issues in a timely manner
4. Demonstrates empathy for customer feelings
5. Presents a positive image to the customer
6. Displays a professional image at all times
7. Communicates a positive image of the company and individuals to
customers
oSecond, the company needs some continuum that describes different levels from
low to high on the performance measure.
oThird, the company needs to decide how much of a merit increase will be given
for different levels of performance.
Decisions about these three requirements lead to some form of merit pay
guide. In its simplest form a guideline specifies pay increases permissible for
different levels of performance (Exhibit 11.11).
A more complex guideline ties pay not only to performance but also to position
in the pay range. Exhibit 11.12 illustrates such a system for a food market firm.
oThe percentages in the cells of Exhibit 11.12 are changed yearly to reflect
changing economic conditions. Two patterns are evident in this merit guideline:
In a pay-for-performance system, lower performance is tied to lower pay
increases.
Pay increases at a decreasing rate as employees move through a pay
range. For the same level of performance, employees low in the range receive
higher percentage increases than employees who have progressed further
through the range.
In part this is designed to forestall the time when employees reach the
salary maximum and have their salaries frozen. In part, though, it is also a
cost-control mechanism tied to budgeting procedures.
A. Performance- and Position-Based Guidelines
Given a salary increase matrix, merit increases are relatively easy to
determine.
As Exhibit 11.12 indicates, an employee at the top of his or her pay
grade who receives a “competent” rating would receive a 2 percent increase in
base salary.
A new trainee starting out below the minimum of a pay grade would
receive an 8 percent increase for a “superior” performance rating.
B. Designing Merit Guidelines
Designing merit guidelines involves answering four questions.
oFirst, what should the poorest performer be paid as an increase?
Wage cuts tied to poor performance are very rare.
Most organizations give no increases to very poor performers,
perhaps as a prelude to termination if no improvements are shown.
oSecond, how much should average performers be paid as an increase?
Most organizations try to ensure that average performers are kept
whole relative to cost of living. This dictates that the midpoint of the
merit guidelines equal the percentage change in the local or national
consumer price index (CPI).
oThird, how much should the top performers be paid?
In part, budgetary considerations answer this question.
There is growing evidence that employees do not agree on the
size of increases that they consider meaningful.
oFinally, matrixes can differ in the size of the differential between different
levels of performance. A larger jump between levels would signal a
stronger commitment to recognizing performance with higher pay
increases.
Most companies balance this, though, against cost
considerations.
Exhibit 11.13 shows how a merit grid is constructed when
cost-constraints (merit budgets) are known.