a multi-criteria model for predicting corporate...
TRANSCRIPT
A QUALITATIVE EXPLORATION OF THE RELATIONSHIP BETWEEN PERFORMANCE APPRAISAL AND THE DEVELOPMENT OF INSPIRED
BUSINESS LEADERS
GEORGE P. SILLUPSt. Joseph’s University
Haub School of BusinessAssistant Professor, Department of Pharmaceutical Marketing
Fellow, Pedro Arrupe Center for Business Ethics5600 City Avenue
Philadelphia, PA 19131-1395(610) 660 – 3443
Fax: (610) 660 – 1229E-Mail: [email protected]
May 31, 2006
The author wishes to acknowledge the financial support from the Saint Joseph’s
University Pedro Arrupe Center for Business Ethics and the intellectual guidance from
John McCall and Stephen J. Porth
CJBE 2006 1
A Qualitative Exploration of the Relationship between Performance Appraisal and
the Development of Inspired Business Leaders
Abstract
A review of the literature and summary of structured interviews with employees
of four Fortune 100 companies (Aetna, IBM, Johnson & Johnson and Wyeth) were
conducted to understand whether, and if so how, performance appraisal (PA) influences
the development of inspired business leaders (IBLs) in the US. IBLs were defined as
employees who contributed to the fiscal success and establishment and maintenance of
corporate social responsibility of their respective companies by exceeding their
individual annual objectives. The structured interviews explored the way PA, a
systematic process that requires descriptions of job-relevant strengths and weaknesses
and uses observation, judgment and solicited feedback to assess performance, helped
and/or hindered development of an IBL. Qualitative findings infer that PA often
attributes success to outcomes of behavior (e.g., units sold) rather than behavior leading
to job success (e.g., motivating others to sell units). Furthermore, PA may impair
development of IBLs by shortcomings related to the PA system’s implementation, such
as not training those conducting PA or not allowing performance assessors enough time
to conduct PA properly. Results suggest that PA and its implementation can influence
development of IBLs and merits further investigation using quantitative analyses.
CJBE 2006 2
A Qualitative Exploration of the Relationship between Performance Appraisal and
the Development of Inspired Business Leaders
Table of Contents
Connecting Performance Appraisal and Inspired Business Leaders page 4
Review of the Literature about Performance Appraisal Systems page 5
Research Methodology page 25
Results page 28
Discussion page 31
Concluding Remarks / Implications for Further Research page 33
References page 35
Tables and Figures
Table 1: Behavioral Checklist Example page 11
Table 2: Graphic Rating Scale Example page 12
Table 3: Successful Uses of Performance Appraisal Data page 18
Table 4: Profile of Respondents page 28
Table 5: Frequency of Interview Responses page 29
Figure 1: Balanced Scorecard Example page 34
N.B. Contributors to the content of this paper include employees of Aetna, IBM, Johnson &
Johnson and Wyeth Pharmaceuticals as well as Susan Givens-Skeaton for her input about
managing human resources and introduction to Wayne Casio’s research and Christopher Field for
his critical review.
CJBE 2006 3
A Qualitative Exploration of the Relationship between Performance Appraisal and
the Development of Inspired Business Leaders
Connecting Performance Appraisal and Inspired Business Leaders
Performance Appraisal (PA) emerged as a concept with the emergence of “big” business
in the early twentieth century in the US when employees were considered a means of
production. Karl Marx’s work can be said to pioneer the moral critique of capitalism and
subject capitalistic business enterprise to moral scrutiny as well as encourage big business
to place some value on an individual employee’s contributions rather than only on the
production of its products.1 But progress to raise this concept to an issue for
consideration was slow; big business was busy building products for consumption around
defining events of the first half of the last century (World War I, the Great Depression,
the New Deal, World War II).
Following World War II, big business began to use PA, usually informally and not as an
assessment for all employees. However, the importance of PA was enhanced over the
last three decades of the twentieth century. US big business has encountered competition
from new domestic and global competitors and began to connect individual employee
contributions with improving their overall performance. Additionally and importantly,
“ethical” treatment of employees was enhanced by the civil rights and environmental
movements along with questions about American business practice from direct and
indirect involvement with the Nixon campaign. These events caused a revolution of
rising moral expectations for American business and created a new way of doing
business that resulted in numerous legislative acts benefiting the employee, e.g., Civil
Rights Act of 1964 at the national level, Massachusetts’ Right-to-Know Law at the state
level and Collective Bargaining at the union contract level.
Despite this enhanced treatment of employees, US business employees still work “at
will” without contractual assurance of her/his job in most instances (e.g., non-union
employees) based on the right to property.2 An employee is expected to contribute to the
business’s success where success is defined by well-known financial metrics (e.g., ability
to generate revenue).3 However, emphasis on the well-known financial metrics often fails
CJBE 2006 4
to place value on other attributes of performance, such as integrity, or nurture business
practices that help to develop inspired business leaders (IBLs).
Recognizing this potential shortcoming, PA systems have been embellished with
questions to assess performance attributes and/or company values (e.g., dependability)
over the past 20 years by US companies. These embellishments add complexity and may
require more time for implementation. Consequently, businesses may not be allowing
sufficient time for implementation of the PA system they are using. When you consider
that the basis of PA is people making judgments about other people in an organized
setting, insufficient time for PA can be unfair to an employee in the short term and
jeopardizes his/her development as an IBL in the long term.4 A review of the literature
about PA systems and the potential impediments to using and implementing them is in
order to answer the following research questions:
Does implementation of an effective PA system help to develop IBLs?
What are the key impediments to using a PA system? This includes impediments
before and/or during implementation.
How can implementation be improved?
What are the implications of this research for the development of IBLs?
Review of the Literature about PA Systems
PA is the systematic description of the job-relevant strengths and weaknesses of an
employee or an employee team. PA combines two processes, observation and judgment,
to assess job performance on the basis of objective and subjective indices. Objective
indices, such as sales data, are appealing because they measure a quantifiable endpoint.
However, they often measure factors beyond an employee’s control, or outcomes of
behavior (e.g., number of units sold), rather than the job performance behavior itself
(e.g., what did it take to sell the units). Subjective indices measure the job performance
behavior.
PA is an integral part of performance management (PM), which requires a willingness
and commitment to improve the every day performance of the employee. PM systems
provide instantaneous, real-time information that describe the difference between
employees’ current and desired courses by giving timely feedback about performance
CJBE 2006 5
while constantly focusing attention on implementing strategies that will help the business
grow.
To work successfully, PM requires that managers do three things well:
Define performance (through goals, measure and assessments),
Facilitate performance (by identifying obstacles to good performance and providing
resources to accomplish objectives),
Encourage performance by providing a sufficient number of rewards in a timely and
fair manner that people value.5
As an integral part of PM, PA helps employees improve their work performance by
realizing their full potential by providing information to employees and managers for use
in making work-related decisions. PA is also a feedback process, an organizational
intervention, a measurement process as well as an intensely emotional process. Above
all, it is an inexact, human process, which, not surprisingly, can be subject to bias during
implementation. However, PA can fail for a number of reasons that can infiltrate and/or
compromise PA systems before implementation:
Setting fair performance standards,
Meeting requirements of effective PA systems,
Assessing the types of PA systems/determining which one to use,
Identifying who should rate performance,
Overcoming judgmental biases in rating,
Establishing training programs for raters,
Determining how often PA should be done,
Understanding how PA fits with Total Quality Management.
Setting Fair Performance Standards
To understand how these failures can occur, it is necessary to consider of how a business
sets its performance standards (PS). As mentioned earlier, PA involves observation and
judgment. Observation processes include the detection, perception and recall or
recognition of specific behavioral events. Judgment processes include the categorization,
integration and evaluation of information.6 Observation and judgment help PA make
distinctions among people, especially among people in the same job. PS provide the
critical link in the process between job analysis, which identifies the components of a
CJBE 2006 6
particular job or job description, and PA.
Ultimately, it is management’s responsibility to establish PS at levels of performance
deemed acceptable or unacceptable for each of the job-relevant, critical areas of
performance identified through job analysis. For some jobs (e.g., production or sales), PS
can be set on the basis of quantitative performance measures, such as the number of units
sold. For others, such as new product development, setting PS is considerably more
subjective and is frequently a matter of manager and employee agreement.7
PS are essential in all types of goods-producing and service organizations because they
help ensure consistency in supervisory judgments across individuals in the same job.
Unfortunately, it is often the case that charges of unequal treatment and unfair
discrimination arise because no clear PS exist.8 Once PS are in place, PA can be done by
gathering job performance information using observation and by evaluating the adequacy
of individual performance using judgment.
Meeting Requirements of Effective PA Systems
The requirements of effective PA Systems are relevance, sensitivity, reliability,
acceptability and practicality. Relevance implies there are clear links between up-to-date
PS for a particular job and an organization’s objectives and between the critical job
elements identified through job analysis and the dimensions to be rated. In short,
relevance is determined by answering the question “What really makes the difference
between success and failure in a job according to the customer?” The customer may be
internal (e.g., your manager) or external (the recipient of business’s product or service).
In all cases, it is important to pay attention to what the customer believes is important
(e.g., on-time delivery).9
Sensitivity implies that a PA system is capable of distinguishing effective from
ineffective performance. If it cannot, and the best employees are rated no differently
from the worst employees, then the PA is not useful. PA that lacks sensitivity will not
help employees develop and will undermine the motivation of both managers (who will
be likely to view PA as pointless paperwork) and employees. Such measures include
acknowledgement of different endpoints of factors in one’s work environment.
Reliability is the third requirement of sound PA and refers to consistency of judgment.
For any given employee, PA by raters working independently of one another should
CJBE 2006 7
agree closely. Raters with different perspectives (e.g., managers, peers and employees)
often view the same employee’s job performance very differently.10 To provide reliable
data, each rater must have an adequate opportunity to observe what the employee has
done and the conditions under which he or she has done it; otherwise unreliability may
be confused with unfamiliarity. Note that there has been no mention of the validity or
accuracy of appraisal judgments because there is no thoroughly objective “truth” in PA.
By making PA systems relevant, sensitive and reliable, the resulting judgments can be
considered valid as well.
In practice, acceptability is often considered the most important requirement of all. Any
PA system must have the support of those who use it, or else employees will undermine
it. Unfortunately, many businesses have not put enough effort into garnering the front-
end support and participation of those who will use the PA system. This is also true for
practicality which requires that PA systems are easy for managers and employees to
understand. Consequently, PA systems often do not work because they were designed
with limited input from managers and even less input from the employees.11
In a broader context, relevance, sensitivity and reliability are simply technical
components of a PA system designed to make decisions about employees. Using a PA
system that ensures consistent evaluation of employees’ performance will help the PA
system’s acceptability. However, because some degree of error is possible in all
employment decisions, determining the optimal PA system to use will result in the
greatest benefit for the business and its employees.
Assessing the Types of PA Systems / Determining Which One to Use
There are two primary types of PA systems, Behavior-Oriented Rating Methods and
Results-Oriented Rating Methods. The type selected should depend on what the system
needs to accomplish; this requires a strategy for the management of performance. In a
study about work motivation, a fairly well established principle, “things that get
rewarded get done,” was reinforced. At least one author has termed this “the greatest
management principle in the world.”12 So, a fundamental issue for managers is “What
kind of behavior do I want to encourage in my employees?” If employees are rewarded
for generating short-term results (e.g., sales during business quarter), then they will work
towards short-term results. If they are rewarded for completing long-term results, (e.g.,
CJBE 2006 8
generating repeat business), then employees will aspire for those things. To be most
effective, however, the manager must think strategically so when all her/his employees’
objectives are completed, the business gains competitive advantage in their market (e.g.,
faster delivery to customers, higher quality at lower cost).13 With the strategic intent of
PA in mind, a review of key Behavior-Oriented and Results-Oriented Rating Methods
follows.
Behavior-Oriented Rating Methods
There are several types of Behavior-Oriented Rating Methods. Narrative Essay is the
simplest type of PA rating system.14 In this method, a rater describes in writing an
employee’s strengths, weaknesses and potential, together with suggestions for
improvement. This approach assumes that a candid statement from a rater who is
knowledgeable about an employee’s performance is just as valid as more formal and
complicated rating methods. When done well, Narrative Essays can provide detailed
feedback to employees regarding their performance. However, when they are totally
unstructured and vary widely in length and content, comparisons across other employees
or departments is almost impossible. Also, comparisons across other employees or
departments are not possible because different essays touch on different aspects of each
employee’s performance and do not objectively compare employees relative to one
another.
Ranking is the next Behavior-Oriented Rating Method. Simple Ranking requires only
that a rater order all employees from highest to lowest or from “best” employee to
“worst” employee. Alternation ranking requires that a rater initially list all employees.
From this list he or she first chooses the best employee (No. 1), then the worst employee
(No. n), then the second best (No. 2), then the second worst (No. n — 1), and so forth,
alternating from the top to the bottom of the list until all employees have been ranked.
Both simple and alternation ranking implicitly require a rater to compare each ratee with
every other ratee but a systematic ratee-to-ratee comparison is not a feature of simple or
alternation ranking.
To get a ratee-to-ratee comparison, Paired Comparisons need to be used. Each employee
is compared with every other employee, usually in terms of an overall category such as
“current value to the organization.” The number of pairs of ratees to be compared may
CJBE 2006 9
be calculated from the formula [[n(n-1)]÷2]. For example, if 10 individuals were being
compared, [10(9)] ÷ 2 or 45 comparisons would be required. The rater’s task is simply
to choose the “better” of each pair and each employee’s rank is determined by counting
the number of times she or he was rated superior. However, these comparisons are made
on an overall basis and their results may be questioned because they lack behavioral
specificity from making comparisons in terms of a single overall suitability category (that
is, “Who is better?”).15. On the other hand, paired comparisons are easy to explain and are
helpful in making personnel decisions. They also provide useful data in validation
studies, for they effectively control leniency, severity and central tendency bias.
Forced Distribution is another method of comparing employees with one another. As the
name implies, the overall distribution of ratings is forced into a normal, or bell-shaped,
curve under the assumption that a relatively small portion of employees is truly
outstanding, a relatively small portion is unsatisfactory and everybody else falls in
between (five categories based on a normal distribution curve). Its primary advantage is
that it controls leniency, severity and central tendency biases rather effectively. However,
it assumes that ratees conform to a normal distribution. This has the potential to
introduce a great deal of error if a significant group of ratees is either outstanding or
unsatisfactory.
Another Behavior-Oriented Rating Method is the Behavioral Checklist in which raters
are provided a series of statements that describe job-related behavior. The rater’s task is
simply to “check” which of the statements describes the employee’s job-related behavior.
In this approach, raters are more reliable as reporters of job behavior than as evaluators
of job behavior.16 Table 1 below is an example of the Likert method of a summated
rating for one declarative statement, “She or he follows up on customer complaints.” It
is followed by categories, such as “always,” “very often,” “fairly often,” “infrequently”
and “never.” The rater checks the response category that he or she thinks best describes
the employee’s performance for that category. Each category is weighted, for example,
from 5 (always) to 1 (“never”) and an overall numerical rating (or score) for each
employee is then derived by summing responses that were checked for each item.
Table 1: Behavioral Checklist ExampleJob-RelatedBehavior
Always (5)
Very Often (4)
Fairly Often (3)
Infrequently(2)
Never(1)
CJBE 2006 10
Good follow-up on customer complaints
There can be several problems with summated rating scales, e.g., selection of response
categories is often arbitrary. Additionally, behavioral checklists have no control for
leniency but do not have a significant problem with halo effect because they only yield
an overall summary rating. It is also difficult for a rater to give diagnostic feedback based
on checklist ratings because they are not cast in terms of specific behaviors. On balance,
practicality probably accounts for their widespread popularity.
Critical Incidents are brief reports by managers of employees’ performance that are
effective or ineffective in accomplishing the objectives of their jobs. For example, a store
manager in a retail computer store observed a salesperson encourage a customer try a
new word processing package by telling her to sit down at the computer and write a
letter. The finished product highlighted typographical and spelling errors. The
salesperson then applied a “spelling checker” to the written material. Pleased, the
customer purchased the new word processing package plus a spelling checker.
This example of a Critical Incident brings attention to the ways in which situations
determine job behavior and on ways of doing the job successfully that may be unique to
the employee described. As a result, they can provide the basis for training programs.
Critical Incidents also lend themselves nicely to PA interviews because managers can
focus on actual job behaviors rather than on vaguely defined traits. They judge
performance, not personality. Ratees receive meaningful feedback and can see what
changes in their job behavior will be necessary for them to improve. In addition, when a
large number of critical incidents are collected, abstracted and categorized, they can
provide a rich database about job and organizational problems in general and are
particularly well-suited for establishing objectives for training programs. However,
managers may find that recording incidents for their employees burdensome, especially if
they are doing so on a weekly basis. Moreover, Critical Incidents alone do not permit
comparisons across individuals or departments but Graphic Rating Scales can overcome
this problem.17
Graphic Rating Scales (GRS) are used by many organizations.18 Many variations of
CJBE 2006 11
graphic rating scales exist but, basically, they differ in three ways in the amount of
structure they provide:
1.) Degree to which the meaning of the response categories is defined; in the table below,
what does “appearance” mean?
2.) Degree to which the individual who is interpreting the ratings (e.g., higher-level
reviewing manager) can tell clearly what response was intended (place an “x” in the
box for the intended response)
3.) Degree to which the performance dimensions are defined for the rater; see example in
table below, what does “satisfactory” mean?)
GRS may not yield the depth of narrative essays or critical incidents but they are less
time-consuming to develop and administer. Additionally, results can be expressed in
quantitative terms. GRSs also consider more than one performance dimension and,
because the scales are standardized, facilitate comparisons across employees; see Table 2
below.
Table 2: Graphic Rating Scale ExampleResponse Categories(below)
Performance Dimension(to the right)
Unsatisfactory – does not meet job requirements
Satisfactory – meets job requirements
Outstanding – exceeds job requirements
Appearance – dresses appropriately for job (e.g., steel-toe work shoes)
X
Behaviorally Anchored Rating Scales (BARS) are a variation of simple graphic rating
scales. Their major advantage is that they define dimensions to be rated in behavioral
terms and use critical incidents to describe various levels of performance. As a result,
BARS provide a common frame of reference for raters. However, they require
considerable effort to develop; separate BARS are needed for dissimilar jobs.
Consequently, this approach may not be practical for many companies.19 Nevertheless,
the participative process required to develop them provides information that is useful for
other organizational purposes, such as developing PS so employees know exactly what
“good performance” means in the context of their jobs.
CJBE 2006 12
Results-Oriented Rating Methods
There are two key Results-Oriented Rating Methods. Management By Objectives (MBO)
is a well-known process of managing that relies establishing objectives for an
organization as a whole, for each department, for each manager within each department
and for each employee. MBO is not a measure of employee behavior; rather, it is a
measure of each employee’s contribution to the organization.20 To establish objectives,
the people involved should do three things:
1.) Meet to agree on the major objectives for a given period of time, e.g., every year
2.) Develop plans for how and when the objectives will be accomplished
3.) Agree on the metrics to determine whether the objectives have been met.
Progress reviews are held regularly until the end of the period for which the objectives
were established. At that time, those who established objectives at each level meet to
update the results and to agree on the objectives for the next period.21
MBO is often considered a complete system of planning and control to include a
philosophy of management. In theory, MBO promotes success in each employee because,
as each employee succeeds, so that employee’s manager, department and organization
also succeed. But this is true only to the extent that the individual, departmental and
organizational goals are aligned. Very few applications of MBO have actually adopted a
formal “cascading process” to ensure such a linkage. An effective MBO system takes
from three to five years to implement and, relatively few firms are willing to make that
kind of commitment, so it is not surprising that MBO systems do not work as well as
they could.22
Work Planning and Review is the other results-oriented rating method and is similar to
MBO. However, Work Planning and Review places greater emphasis on the periodic
review of work plans by both manager and employee in order to identify goals attained,
problems encountered and the need for training.23 As a result, Work Planning and
Review is popular in the US where emphasis is placed on objective outcomes but is less
popular in other parts of the world like Japan and France where greater emphasis is
placed on the psychological and behavioral sides of performance, such as effort put into a
job or cooperative spirit.
From review of many rating formats, it is clear that they focus on employee behaviors,
CJBE 2006 13
either by comparing the behaviors of employees (Behavior-Oriented Rating Methods) or
by looking at the results of employees’ work, e.g., dollar volume of sales, (Results-
Oriented Rating Methods). The type of system needs to be consistent with what the
system needs to accomplish as well as make it more difficult to reinforce less
quantifiable performance measures. As mentioned earlier, use of Work Planning and
Review system deemphasizes assessment of the non-financial attributes of performance.
Identifying Who Should Rate Performance
Raters are critically important to the success of any PA system because the employee
ratings depend on the judgments of the raters. Raters must accept the importance of PA
as an organizational objective, as an effective means for achieving that objective and as a
personal goal.24 In short, diligent attention to PA must be seen as an integral part of the
rater’s job, not as “make-work” harassment. Probably the best way to do this is to base
part of the rater’s own performance appraisal on the quality of his or her ratings of
others. To rate performance effectively, the most fundamental requirement for any rater
is that he or she have an adequate opportunity to observe an employee’s job performance
over a reasonable period of time (e.g., six months), which suggests several possible
raters.
Immediate Manager
Surveys show that up to 98% of appraisal programs ask the immediate manager (IM) to
take sole responsibility for PA because IMs are probably most familiar with each
employee’s performance and have the broadest opportunity to observe the employee.25
Furthermore, the IM is probably best equipped to relate the employee’s performance to
what the department and organization is trying to accomplish. Because the IM is also
responsible for reward (and punishment) decisions in the overall PM process, it is not
surprising that feedback from IMs is more highly re1ated to performance than that from
any other source.27 However, when job circumstances dictate that the IM is not always in
the best position to observe performance (e.g., teaching), several other raters can be used
to provide a better picture of the employee’s total performance.
Peers
In some jobs, such as outside sales, and in some environments, such as self-managed
work teams, the IM may not observe an employee’s actual job performance routinely or
CJBE 2006 14
only indirectly, through written reports. In these situations, assessment by peers can
provide a perspective on performance that is very valuable. Peer assessment can be
accomplished in at least three ways:
Peer Nominations - most useful for identifying persons with extremely high or low
levels of knowledge, skills, abilities, other characteristics;
Peer Rating - most useful for providing feedback;
Peer Ranking - best for discriminating various levels of performance from highest to
lowest on each dimension.
Assessment by peers reaches valid conclusions regarding reliability, validity and freedom
from biases but perceived friendship bias is a caveat.27 Employees react negatively to
negative assessment by their peers and produced significantly lower performance of their
groups. Meanwhile positive peer rating feedback produced non-significantly higher
performance levels of their groups on a subsequent task.28
To reduce potential friendship bias while simultaneously increasing the feedback value of
the information provided, those responsible for the implementation of the peer
assessment must specify exactly what the peers are to evaluate, e.g., the quality of their
help on technical problems.29 Another approach is to require input from a number of
colleagues (similar to 360-feedback). For example, at Harley-Davidson, salaried workers
have five colleagues critique their work.30
Direct Reports
Direct reports offer a somewhat different perspective on a manager’s performance. They
know directly the extent to which the manager delegates, plans, organizes, communicates
as well as their leadership style.31 PA by direct reports requires considerable trust and
openness. To protect direct reports against retribution, ratings are collected and
combined into an overall rating so the manger’s performance is not distorted by an
anomalous opinion. Additionally, the manager does not know the individual responses
of their direct reports. Any organization contemplating use of direct report ratings
should pay careful attention to the intended purpose of the ratings. Findings indicate that
their use for salary administration or promotion is not as good as their use for guided
self-development.32
Self-Appraisal
CJBE 2006 15
The opportunity to participate in the PA, particularly if appraisal is combined with goal
setting, improves the employee’s motivation and reduces her or his defensiveness during
the appraisal.33 On the other hand, comparisons with appraisals by managers, peers and
direct reports suggest that self-appraisals tend to show more leniency, less variability,
more bias and less agreement with the judgments of others.34 To some extent, these
disagreements may stem from the tendency of self-raters tend to place greater emphasis
on personal skill and technical competence while managers stress output and results.
The validity of self-appraisals can be improved by incorporating four research-based
suggestions.35
1.) Instead of asking individuals to rate themselves on an absolute scale (e.g., a scale
ranging from “poor to average”), provide a relative scale that allows them to compare
their performance with that of others (e.g., “below average,” “average,” “above
average”).
2.) Provide multiple opportunities for self-appraisal because the skill being evaluated can
improve with practice.
3.) Provide reassurance of confidentiality—self-appraisals will not be “publicized.”
4.) Focus on predicting future behavior.
Customers
In some situations, external consumers can provide a unique perspective about job
performance, e.g., subscribers to a cable television service. Although the customers’
objectives cannot be expected to correspond with the organization’s objectives, the
information that customers provide can be useful input for employment decisions, such
as those regarding promotion, transfer and need for training. Customers’ input can also
be used to assess the impact of training or as a basis for self-development. At General
Electric, customers of senior managers are interviewed formally and regularly as part of
the managers’ appraisal process. Their evaluations are important but at the same time,
they also build commitment, because customers are giving time to help improve General
Electric’s emplloyees.36
Multi-Rater or 360-Degree Feedback
Many organizations, including Alcoa, Lockheed-Martin and DuPont, now use input from
managers, peers and direct reports to provide a perspective on performance from all
CJBE 2006 16
angles (360 degrees).37 Multi-rater feedback can be helpful but can be accompanied by
problems resulting from:38
Ambiguous Objectives,
Selective Use - only problem employees get 360-degree feedback,
Changing the ground rules after the process has begun,
Unclear instructions to raters,
Untrained Raters,
Failing to develop an action plan following feedback,
Lack of follow-through-need plan for each employee.
Computer Monitoring
Many employees spend a lot of time unsupervised by their managers but technology is
enabling continuous supervision of several million employees today, e.g., grocery check-
out clerks.39 Rating using computers is neither good nor bad; how managers use the data
determines its acceptance in the workplace. Once employees are informed that their
performance will be monitored by technology, it can provide some valuable insights.
However, if it is used to identify only problems, rating using computers may seem like a
modern-day version of Modern Times, the Charlie Chaplin movie in which the hapless
hero was tyrannized by automation. If used to identify positive performance, it can be an
incentive. For example, bank tellers like the opportunity to earn up to 25 percent more
than their base pay if their output is high.
Research indicates that data from multiple sources (e.g., self, managers, peers, direct
reports) are desirable because they provide a complete picture of an employee’s effect on
others.40 Moreover, data from four studies revealed that there is consistency in the way
people in different positions evaluate performance; this suggests that there is consistency
across raters.41 Note that contextual differences of the raters may place different emphasis
on specific aspects of job performance that can also affect results of PA and suggest
implementing a combination of systems, e.g., manager, peer and self; see Table 3
below.42
Table 3: Sources and Uses of Performance Appraisal DataData Use Data Source
Manager Peer Direct Reports Self CustomersPersonnel X X X
CJBE 2006 17
DecisionsSelf Development
X X X X X
Personnel Research
X X X
Overcoming Judgmental Biases in Rating
Rater judgments are subject to various types of biases: leniency and severity, central
tendency, halo effect, contrast error and recency error. In the traditional view,
judgmental biases result from some systematic measurement error on the part of a rater.
As such, they are easier to handle than random errors. However, to many managers, these
biases are not errors at all. Rather, they are discretionary actions that help them manage
people more effectively.43 With this in mind, a review of the most commonly observed
judgmental biases, along with ways of avoiding them, follows.
Leniency and Severity
The use of ratings rests on the assumption that human observers are capable of some
degree of precision and objectivity. Based on this assumption, it is believed that
employees doing PA can be objective about certain aspects of the person rated. However,
this assumption is the one most often violated. Raters subscribe to their own sets of
assumptions (which may or may not be valid). Many employees have encountered raters
who seemed either inordinately easy (lenient) or inordinately difficult (severe).
Numerous reasons sustain this tendency but perhaps one of the most compelling is
psychological commitment.44
Psychological commitment is when an individual with input into a personnel decision has
a commitment in favor of reaffirming their previous decision when asked to provide a
later decision the same issue. The direction of the bias depends on the opinion of the
rater, either for or against the original decision. Individuals who agree with the previous
decision tend to provide lenient ratings, whereas those who disagree with the decision
tend to provide severe ratings.45
Leniency and severity biases can be controlled or eliminated in two ways:
Allocating ratings into a forced distribution, in which ratees are apportioned
according to an approximately normal distribution;
Reducing ambiguity in the rating scales themselves.
CJBE 2006 18
The two alternatives differ in their theoretical underpinnings. Forced distribution
removes a substantial amount of control from the rater by telling him or her that “10% of
your ratees must be rated poor or outstanding.” Improvement in scale definitions is based
on the notion that the rater should be helped by specifying clearly what is to be rated and
what the various scale points mean. In this way, he or she will be better able to distin-
guish the various levels of absolute vs. relative performance.
Central Tendency
When political considerations predominate, raters may assign all their subordinates rat-
ings that are neither too good nor too bad. They avoid using the high and low extremes of
rating scales and tend to cluster all ratings about the center of all scales. “Everybody is
average” is one way of expressing the central tendency bias. The unfortunate conse-
quence, as with leniency or severity biases, is that most of the value of PA is lost. The
ratings fail to distinguish between superior and inferior employees and fail to isolate
areas in which an individual employee can improve.46 Central tendency biases can be
minimized by specifying clearly what the various definitions mean. In addition, raters
must be convinced of the value and potential uses of merit ratings if they are to provide
meaningful information.
Halo Effect
Halo bias is perhaps the most pervasive bias in PA and has been identified over 80 years
ago.47 A rater who is subject to the halo bias assigns ratings on the basis of a global
impression of the ratee. An individual is rated either high or low on many factors
because the rater knows (or thinks he/she knows) that the individual is high or low on
some specific factor. According to this theory, the rater fails to distinguish among
specific levels of performance due to the interaction of a rater and a situation. For
example, a rater may extrapolate one observation and project it into all facets of the
employee’s performance.48
A common assumption is that increased observation of performance behavior will reduce
halo bias and thus improve the validity of subsequent ratings. This is not true. Evidence
indicates than when raters have a greater opportunity to observe ratees’ performance and
become more familiar with the behavior to be rated, halo actually increases.49 Training
raters to adhere to objective PA is the most practical remedy to overcome this bias.
CJBE 2006 19
Contrast Error
Contrast Error results when a rater compares several employees with one another rather
than with an objective PS.50 If the first two workers are unsatisfactory while the third is
average, the third worker may well be rated outstanding because in contrast to the first
two his “average” level of job performance is magnified. Likewise, “average”
performance could be downgraded unfairly if the first few workers are outstanding. In
both cases, the “average” worker receives a biased rating. Here again the answer to this
bias is heightening raters’ awareness through training and through a broadening of their
opportunities to evaluate a wide array of employees’ performance.
Recency Error
Recency Error results when a rater assigns his or her ratings on the basis of the
employee’s most recent performance. It is most likely to occur when appraisals are done
only after long periods. Recency error can also penalize an employee who has a
substandard performance during the last quarter before PA is conducted. While there is
no easy correction to this, more frequent PA and/or training to raise the level of
awareness of the rater so he/she can avoid it.
Each of the available PA methods for rating job performance attempts to reduce bias in
some way, although no method is completely bias-free. Biases may be associated with
raters (e.g., lack of first-hand knowledge of employee performance), ratees (e.g., sex, job
tenure), interaction between raters and ratees (e.g., race and sex) or judgmental biases. As
mentioned, a key to reducing bias is training.
Establishing Training Programs for Raters
This is one of those areas that seems as though it should be straight forward but training
is not that easy. Many training programs are designed without consulting those who are
rating performance to determine what skills raters need. Generally, there are three broad
objectives for training:
Eliminate judgmental biases such as leniency, central tendency, halo, contrast effects,
Improve observational skills by teaching which performance attributes should receive
attention,
Improve ability to communicate appraisal information in a constructive manner.51
When these three broad objectives are met, raters improve in the objectivity of their
CJBE 2006 20
ratings by showing less leniency and halo effect. However, their accuracy tends to
decrease.52 The key to correcting this problem is emphasis on improving the raters’
observational skills using a step-wise training process to emphasize how to observe
behavior more accurately, not how to rate.53 Moreover, research indicates that rater
training aimed at improving the judgmental skills of raters is clearly worth the effort to
improve PA.54
Determining How Often Should PA Be Done
Traditionally, formal PA is done once, or at best, twice a year. Research, however, has
indicated that this is far too infrequent.55 Unless raters keep a diary, they will face
considerable difficulties remembering what employees did over the previous six-to-12
months. Consequently, several companies, such as Southern California Gas, add
frequent, informal “progress” reviews between the annual ones.56 Another useful
alternative is to conduct PAs upon the completion of projects or upon the achievement of
important milestones. For example, at several Johnson & Johnson companies, employees
get formal reviews every six months or at the end of each project.57
This an approach has merit because PAs are more likely to send clear messages to
employees about where they stand. It helps to guarantee no “surprises” in PA.
Companies with monthly or quarterly PA (as opposed to annual assessments)
outperformed competitors without such systems on every financial and productivity
measure used in study, e.g., profits, and got positive feedback from employees about the
fairness of the PA system.58 On the other hand, frequent PA can be quite time consuming
and may make employees feel as though their performance is always being formally
assessed. An alternative is to calibrate all PAs during one time of the year (e.g., last
quarter of fiscal year) and make less formal, ongoing observations the rest of the time.
Understanding How PA Fits with Total Quality Management
Total Quality Management (TQM) emphasizes the continuous improvement of products
and processes to ensure long-term customer satisfaction. Its group problem-solving focus
encourages employee empowerment by using job-related expertise and the ingenuity of
the workforce. Cross-functional teams develop solutions to problems, often shortening
the time taken to design, develop or produce products and services. Because a team may
not include a manager, the dividing line between labor and management often becomes
CJBE 2006 21
blurred in practice, as employees begin to solve organizationa1 problems themselves.59
Thus, adoption of TQM generally requires cultural change within the organization as
management reexamines its past methods and adopts this philosophy.60
If the “fathers of TQM,” Joseph M. Juran and W. Edwards Deming, had their way, PA
that ties individual performance to salary adjustments would be eliminated. They posit
that most PA systems are based on a faulty assumption that employees have significant
control over their own performance—that is, they can improve if they choose to do so by
putting forth the necessary effort. However, everything done in an organization is done
within the framework of one or more systems (e.g., accounting, purchasing). These
systems provide limits on processes, employees and even managers. In a well-designed
system, it will be nearly impossible to do a job improperly. Conversely, a poor system
can thwart the best efforts of the best employee. Because employees have little
opportunity to change systems, such as outdated technology that makes it impossible to
meet current standards, they become frustrated and demoralized.61 Bottom line is that, as
a basis for implementing a “pay-for-performance” philosophy, PA is a meaningful tool
only if workers have a significant control over the variables that affect their individual
performance.62
Harmonizing PA with employee autonomy and performance within a business
organization seems like a daunting task. However, three practical suggestions can help
ensure fair and equitable treatment of all employees.63 First, let internal or external
customer expectations generate the performance expectations for employees. Then, have
employees assess their performance against those expectations. Using this baseline,
employees and managers can develop their own continuous improvement targets.
Comparing actual performance against expected performance helps avoid detrimental
competition because employees are compared against their own benchmarks rather than
against each others’ performance.
Second, include results-expectations that identify actions to meet or exceed those
expectations. Here, the direct report and manager work together to set expectations in
conjunction with the business objectives. Third, emphasize behavioral skills that make a
real difference in achieving quality performance and total customer satisfaction, e.g.,
“attention to detail.” These continuous-improvement skills are as important to TQM as
CJBE 2006 22
are results-oriented expectations. This form of PA helps all employees achieve excellent
customer satisfaction as the way to do business.
Behavioral Barriers to Implementing a PA System
With a better appreciation of reasons that can compromise the fair and equitable
treatment of employees before implementation, a review of the challenges that arise
when implementing a PA system is in order. To begin, even when pre-implementation
potential reasons are controlled, there will still be challenges when it comes to
implementation. Tantamount to containing these challenges is serious commitment by
executives combined with “buy-in” by the employees.64
PA may encounter challenges due to mistrust by managers and employees about the uses
of information generated by the PA system. This mistrust can elicit behavioral barriers
to implementation, which may be political or interpersonal. Political barriers stem from
deliberate attempts by managers and/or raters to enhance or to protect their self-interests
when multiple courses of action are possible. For example, a manager may wish to retain
an employee who has demonstrated outstanding performance by submitting a
substandard or inaccurate PA as opposed to an honest PA that would lead the employee’s
promotion. Furthermore, interpersonal barriers arise from the actual face-to-face
encounter between employee and manager.65
Political considerations are organizational facts of life. Evidence indicates that PAs take
place in an organizational environment that is anything but completely rational,
straightforward or dispassionate. It appears that accuracy in PA is less important to
managers than motivating and rewarding their employees. Many managers will not allow
fair and accurate ratings to cause problems for themselves and attempt to use PA to their
own advantage.66 In this type of climate, personal values and bias can replace
organizational standards. Unfairly low ratings may be given to highly valued employees
or favored treatment may be given to other employees. In extreme cases, a manager may
promote a difficult employee to get rid of them.
Even if political tension can be lessened, significant interpersonal barriers may also
hinder PA. Due to lack of communication, employees may think they are being judged
according to one set of standards, when their managers actually use different standards.
Furthermore, managers often delay or resist making face-to-face appraisals. This can
CJBE 2006 23
appreciably reduce the validity of the ratings because, rather than confront substandard
performers with low ratings, managers often find it easier to “damn with faint praise.”
that usually translates into average or above average ratings for inferior performers.72
Finally, some managers complain that formal PA interviews often interfere with the
more constructive coaching relationship that should exist between managers and direct
reports by placing them in the role of judge – but judging is one of the key roles of a
manager.
CJBE 2006 24
Research Methodology
PA can improve employee’s work performance by helping them to realize their full
potential as IBLs. PA accomplishes this by providing information to employees and
managers for use in making work-related decisions. However, because PA is an inexact
human process, it may only effectively evaluate how employees meet objective
performance measures (e.g., assessing the number of units sold). However, PA may fail
to evaluate attributes that lead to improved performance, e.g., integrity or other attributes
especially important to the development of IBLs. As the literature review indicates, PA
can fail for a number of reasons that can infiltrate and/or compromise it before
implementation and/or during implementation.
To assess how these shortcomings potentially impact the relationship between PA and the
development of IBLs, a 30-45 minute structured interview was developed that included
nine designated areas and one open-ended question:
1.) Does your PA system have clear PS?
2.) How well does your PA system meet the requirements of an effective PA system?
a. Relevance – identify what makes a difference
b. Sensitivity – determine effective/ineffective performance
c. Reliability – platform for consistent judgment
d. Acceptability – have the support of those using it
e. Practicality – easy to understand
3.) What type of PA system does you company use?
a. Behavior-Oriented
i. Narrative essay
ii. Ranking
iii. Checklist
iv. Critical incidents
v. Graphic rating scale
b. Results-Oriented
i. Management By Objectives
ii. Working Plan and Review
4.) Who rates PA?
CJBE 2006 25
a. Immediate Manager
b. Peers
c. Direct Reports
d. Self-appraisal
e. Customers
f. Multi-Rater (360˚-feedback)
g. Computer Monitoring
5.) Are you aware of any biases in rating?
a. Leniency and severity
b. Central tendency
c. Halo effect
d. Contrast error
e. Recency error
6.) Have you had “rater” training? When?
7.) How often is PA conducted?
8.) Is your company doing some variation of TQM? If so, how does PA fit with TQM?
9.) Have you encountered any barriers to PA?
a. Lack of top-down support
b. Lack of employee buy-in
c. Political barriers
d. Interpersonal barriers
10.)Are there any other factors that help/hinder implementation of PA?
The structured interview was implemented by the author to 18 managers from four
Fortune 100 companies – Aetna Life and Casualty (Aetna), International Business Machines
(IBM), Johnson & Johnson (J&J) and Wyeth Pharmaceuticals (Wyeth)-between June and
December 2004. Each manger had to give permission for his/her responses to be included. This
approach attained a total of 16 respondents; two did not have enough time.
All 16 of the complete interviews yielded answers to all the questions. The responses were
identified and categorized from each interview, such as biases in rating. Regardless of the
positions expressed during the interview, I probed to see if the respondent would provide
further detail. For example, when answering Question #2 about meeting the
CJBE 2006 26
requirements of their PA system, I probed to see if they could identify whether they met
all five requirements (Questions #2 a-e). Finally, because perceptions may be influenced
by whether the respondent was conducting the PA or having his/her own performance
being appraised, I asked them to consider both perspectives as well as provided
opportunity to include their own observations (Question #10).
CJBE 2006 27
Results
The 16 interviews from the four companies were divided as follows: Aetna = 4,
IBM = 2, J&J = 6 and Wyeth = 4. The 16 respondents’ departments were Marketing,
Sales, Manufacturing, R&D, Customer Service (CS), Finance, Human Resources (HR)
and Regulatory Affairs (RA). A breakout can be observed in Table 4 below.
Table 4: Profile of Respondents
Companies Number of Respondents(Gender) Departments/Years Experience
Aetna 4 (3F, 1M) Marketing = 2 (7, 10)
HR = 2 (20, 18)
IBM 2 (2F) HR = 1 (22)
CS = 1 (11)
J&J 6 (1F, 5M) Finance = 2 (15, 17)
Marketing = 2 (19, 23)
Sales = 2 (9, 15)
Wyeth 4 (2F, 2M) R&D = 1 (24)
RA = 2 (16, 20)
Finance = 1 (20)
Total 16 (8F, 8M) Marketing = 4
Finance = 3
HR = 3
Sales = 2
RA = 2
CS = 1
R&D = 1
Analysis of the structured interviews revealed use or results-oriented PA systems on an
annual basis and had similar responses from the four companies for both the male and
female respondents (see Table 5 below). Top-down support was identified as most
important for the successful implementation of the PA system. No matter the
department, all respondents had their PA conducted by their immediate managers and
CJBE 2006 28
used multi-rater input (360-feedback). Ten of the 16 used self-appraisal as preparation
for PA. There were two different responses. One was Wyeth for TQM because Wyeth
was the only company with a respondent from Manufacturing, the only department
implementing TQM. The other was Aetna where four customers were asked for
feedback.
Table 5: Frequency of Interview Responses
Interview Questions Responses as Rater Responses as Ratee
1.) Clear PS Yes = 16 Yes = 16
2.) Meets requirements of effective PA system
Yes = 16 Yes = 16
3.) Type of PA system Behavior-Oriented = 0Results-Oriented = 16 with non-objective performance attributes included
Behavior-Oriented = 16Results-Oriented = 16
4.) Who rates PA Immediate Manager = 16Peers = 0Direct Reports = 0Self-appraisal = 10; employee preparation for PA at J&J and WyethCustomers = 4 at AetnaMulti-Rater = 16; 360˚ Computer Monitoring = 0
Immediate Manager = 16Self-Appraisal = 10 as preparation Multi-Rater = 12
5.) Bias in rating None observed = 16 No = 16 but two suspected bias in their reviews
6.) Rater training Yes = 14; informal from their managersNo = 2; new hires at company
N/A
7.) How often conducted Annually = 16 with mid-year update
Annually = 16 with mid-year update
8.) TQM fit with PA Yes = 2; Manufacturing at Wyeth
Yes =1; Manufacturing at Wyeth
9.) Behavioral barriers No = 16; all noted the importance of top-down support
No = 16
10.) Other factors Not enough time to implement = 15
Not enough time to implement = 15
Number of responses does not always equal the number of respondents; 15 respondents rate multiple employees
N = 16 N = 16
Taken as a whole, the interviews were positive regarding PA as a way to gain recognition
for performance and suggesting it as a link to becoming an IBL. Although, none of the
respondents gave me their individual PA rating, they said they “exceeded” the annual
CJBE 2006 29
objectives. Interestingly, none of the respondents suspected bias when they were raters
but two suspected bias (political) when they were being rated by their managers. It was
also interesting to learn that the vast mast majority (15 of 16 respondents) provided
unsolicited responses to Question #10 (Other) and indicated that they did not have
enough time to implement the PA system. Finally, it is surprising that none of the
respondents received formal training about the PA systems they were using. Two
received no informal training because they were new hires with considerable experience
in conducting PA. Those interviewed were steeped in experience with a mean of 16.6
years; the most inexperienced respondent had seven years experience.
CJBE 2006 30
Discussion
The results of this research reinforce that effective implementation of a PA system is an
optimal way to assess an employee’s performance. While this is important, the
development must also introduce the employee to aspects of performance that go beyond
the objective measures evaluated in results-oriented PA systems if the PA system is to
help develop IBLs. These include the performance attributes added to complement the
behavior-oriented PA systems being implemented by the companies in this study.
Examples of this can viewed in three of the four companies participating in this study,
Johnson & Johnson, IBM and Aetna. They have matched their words with their deeds
and integrated concerns for employees’ rights into the very fabric of their PA systems
and corporate cultures. Research from a quarter century ago about IBM and Aetna found
that taking such measures has not been overly costly or impeded implementation of their
PA systems. Despite more recent problems (e.g., IBM connected to the Holocaust), both
companies continue to receive strong employee approval for their policies when they ask
about them on company attitude surveys.
Enough time to implement a PA system seems to be a determinant for the success of the
system. Furthermore, when time is constrained, raters seem to revert to familiar objective
measures, such as financial metrics, and “tell and sell” approach rather than an
information exchange.73 Additionally, raters also become more vulnerable to biases,
(e.g., recency effect). Although most companies require that PA results be discussed
with employees, they do not consider PA as an individual objective. By not making PA
an individual objective, companies seem to missing an opportunity to help improve their
PA system and establish the basis for development of IBLs.
Implementation practices that are being done should be continued. These include:
Annual reviews with a mid-year update – generally feedback has maximum impact
when it is given as close as possible to the action and when the employee already has
relatively accurate perceptions of her or his performance.
Have employees prepare – self-appraisal is very useful. Employees who spend more
time prior to PA interviews (e.g., analyzing the problems they encounter and the
quality of their performance) are more likely to be satisfied with the PA process and
more likely to meet their personal objectives.
CJBE 2006 31
Encourage Participation - A perception of ownership by the employee that her/his
ideas are genuinely welcomed by the manager is related strongly to the employee’s
satisfaction with PA, especially top-down support.
Some practices that are not being done should be done routinely:
Training for Raters – Training enables raters to observe behavior more accurately and
fairly and not on how to rate or not rate employees’ performance.
Set PA as an Objective – Objectives that are set emphasize the importance of meeting
the objective.
CJBE 2006 32
Concluding Remarks / Implications for Further Research
Looking at more companies and more raters/ratees from different departments with less
experience than those interviewed in this study would provide the basis for quantitative
assessment of these qualitative observations. Based on a more robust dataset, interesting
observations, such as how many managers who “exceed” their annual objectives rate
their employees similarly.
Emphasis of this study has been placed on the US but the companies are global and
consideration of their PA systems are working globally would be an interesting
comparison with the findings about the US. Based on research to date, there are
differences between the US and Europe and, especially, non-western cultures like the
countries in the Pacific rim where there is more focus on group rather than individual
performance. This suggests that US managers will need to modify the PA process that is
familiar to them when working with employees in order to make it more consistent with
the values of other cultures.
While support by top management is necessary for successful implementation, an
important part of that implementation is effective executive PA with the same “at will”
PA system that is being used for the organization (e.g., without a “golden parachute”).74
Similar to other employees, it should be taken seriously and provide the executives with
useful feedback and guidance. For example, PA for executives should be strategic and
include a flexibly framed description of the company’s mission and vision. It should also
reinforce the importance of performance attributes and behaviors, such as corporate
social responsibilities. This will go a long way toward creating an environment to
nurture the development of IBLs.
While there is no “correct” PA system to stimulate development of IBLs, one alternative
that enables a business to “manage by values” is the Balanced Scorecard (BSC) depicted
in Figure 1 on the next page.74 The BSC not only looks at the quantifiable measures of
performance (e.g., financial measures) but also puts emphasis on how the business is
regarded by employees and customers in the present and the future. From initial efforts at
Aetna where PA is an identified objective, it appears to be working to accomplish
objectives and inspire future IBLs.
Figure 1: Balanced Scorecard Example
CJBE 2006 33
CJBE 2006 34
Financial MeasuresOwners and Stakeholders’ view
Customer PerspectiveCustomers’ view
Future PerspectiveContinue improvement
Internal PerspectiveEmployees’ viewBalanced
Scorecard
References1 Newton, L.H. and Schmidt, D.P. Wake-Up Calls. Thomson/Southwestern, 2004.2 Weiss, J.W. Business Ethics a Stakeholder and Issues Management Approach.
Thomson/Southwestern, 2003. 3 Juran, J., 2004. Architect of Quality. McGraw-Hill. 4 Cascio, W.F. Applied Psychology in Personnel Management. Prentice Hall, 1991, 4th
edition.5 Thornton, G.C. and Zorich, S. Training to Improve Observer Accuracy. Journal of
Applied Psychology, 1980, vol. 65, pp. 351-354.6 Cascio, W.F. Applied Psychology in Personnel Management. Prentice Hall, 1991, 4th
edition, p. 78.7 Nathan, B.R. and Cascio, W.F. Technical and Legal Standards for Performance
assessment. In R.A. Berk (ed.) Performance Management. Johns Hopkins, 1986.8 Cascio, W.F. Managing Human Resources. Irwin McGraw-Hill, 1998, 5th edition, p.
305.9 Borman, W.C. Job Behavior, Performance and Effectiveness. In M.D. Dunnette and
L.M. Hough, (eds.), Handbook of Industrial and Organizational Psychology, 1991, Vol. 2, pp. 271-326.
10 Kerr, S. in Sherman, S. Stretch Goals: The Dark Side of Asking for Miracles. Fortune, 1995, Nov. 13, p. 31.
11 Cascio, W.F. Managing Human Resources. Irwin McGraw-Hill, 1998, 5th edition, p. 307.
12 Beatty, R.W., 1989. Competitive Human resource Advantage through Strategic Management and Performance. Human Resource Planning, 12, pp. 179-194.
13 Cascio, W.F. Applied Psychology in Personnel Management. Prentice Hall, 1991, 4th edition, p. 85.
14 Cascio & Bernardin, op. cit.15 Stockford, L. and Bissell, H. W., 1949. Factors Involved in Establishing a Merit
Scale. Personnel, 26, pp. 94-116.16 Cascio, W.F. Managing Human Resources. Irwin McGraw-Hill, 1998, 5th edition, p.
312.17 Landy, F. J. and Rastegary, H., 1988. Criteria for Selection. In M. Smith & Ison
(eds.), Advances in Personnel Selection and Assessment. New York, pp. 68-115.18 Cascio, W.F. 1998, op. cit.19 Campbell, J. P., Dunnette, M. D., Lawler, E. F. and Weick, K. F., 1970. B ehavior,
Performance, and Effectiveness. New York: McGraw-Hill.20 Kondrasuk, J. N., 1981. Studies in MEG Effectiveness. Academy of Man Review, 6,
419-430.21 Albrecht, K., 1978. Successful Management By Objectives: An Action Manual.
Englewood Cliffs, NJ: Prentice-Hall.22 Meyer, H. H., Kay, E. and French, J. R. P., 1965. Split Roles in Performance.
Harvard Business Review, 43, 123-129.23 Guion, R. M., 1986. Personnel Evaluation. In R. A. Berk (ed), Performance
Mentoring. Baltimore: Johns Hopkins University Press, pp. 345-360. .24 Bernardin and Beatty, 1984).
CJBE 2006 35
25 Becker, T. F. And Klimoski, R. J., 1989. A Field Study of the Relationship between Organizational Feedback Environment and Performance. Personnel Psychology,pp. 353-358.
26 Kane and Lawler, 1978; Lewin and Zwany, 1976, 27 DeNisi, Randolph and Blencoe, 1983.28 McEvoy, G. M. and Buller, P. F., 1987. User Acceptance of Peer Appraisals in Trial
Setting. Personnel Psychology, 40, 785-787.29 Labor letter, 1990, Oct. 16, The Wall Street Journal, p. Al.30 Reilly, R. R., Smither, J. W. and Vasibopoubos, N. L., 1996. A Longitudinal Upward
Feedback. Personnel Psychology, 49, 599-612. 31 Zedeck and Cascio, 1982.32 Campbell, D.J. and Lee, C., 1988. Self-Appraisal in Performance Evaluation:
Development versus Evaluation. Academy of Management Review, 13, pp. 302-314.33 Harris and Schaubroeck 1988; Thornton, 198034 Mabe and West, 1982; Campbell and Lee, 1988; Fox and Dinur, 198835 Hartel, C.E.J., 1993. Rating Format Revisited. Journal of Applied Psychology, 78,
pp. 212-217.36 Yammarino, F.J. and Atwater, L.E., 1997, Spring. Do Managers See Themselves as
Others SeeThem? Organizational Dynamics, 25, 4, pp. 35-44.37 Cheney, A. and Bremely, M., 1996. The Pitfalls of 360-Degree Feedback. St.Louis,
MO: Psychological Associates.38 Piller, C., 1993, July. Privacy in Peril. Macworld, pp 124-130.39 Wohlers & London, 1989. 40 Shapira & Zevulun, 1989. 41 Zammuto et al., 1982. 42 Longenecker et al., 198743 Bass, 195644 Schoorman, 1988.45 Cascio, W.F. Applied Psychology in Personnel Management. Prentice Hall, 1991, 4th
edition, p. 84.46 Thorndike, 192047 Murphy, K. R. and Anhalt, R. L., 1992. Is Halo Error a Property of the Rater or
Ratees - the Specific Behavior Observed? Journal of Applied Psychology, 77, 494-500.
48 Jacobs & Kozlowski, 198549 Sumer, H. C. and Knight, P. A., 1996. Assimilation and Contrast Effects in
Performance Ratings: Effects of Rating the Previous Performance on Rating Subsequent Performance. Journal of Applied Psychology, 81, 436-442.
50 Longenecker, C.0., Sims, H. P., Jr. and Gioia, D. A., 1987. Behind the Mask:Politics of Employee Appraisal. Academy of Management Executive, 1, 183-191.
51 Murphy, K. R. and Balzer, W. K., 1989. Rater Errors and Rating Accuracy. Journal of
Applied Psychology, 74, 619-624.52 Pulakos, E. D., 1986. The Development of Training Programs to Increase Accuracy
with Different Rating Tasks. Organizational Behavior and Human Decision Process 38, 76-91.
CJBE 2006 36
53 Nathan and Alexander, 198554 Sanchez, J. I. and DeLaTorre, P., 1996. A Second Look at the Relationship between
Rating and Behavioral Accuracy in Performance Appraisal. Journal of Applied Psychology, 81, 3-10.
55 Schellhardt, loc. cit.56 Ibid .57 Campbell and Garfinkel, loc. cit. – GP Sillup personal communication58 Juran, J. M., 2004. Architect of Quality. New York: CWL McGraw-Hill.59 Wiedman, loc. cit.60 Deming, W.E., 1986. Out of Crisis. Cambridge, MA; MIT Center for Advanced
Engineering Study.61 Wiedman, T.G., 1993, October. Performance Appraisal in a Total Quality
Management Environment. The Industrial-Organizational Psychologists, 31, 2, pp 64-66.
62 Total Quality and Performance Appraisal, 1992, October. Johnson & Johnson bulletin.
63 Cascio, W.F., 1998. Managing Human Resources. Irwin McGraw-Hill, 5th edition, p. 328.
64 Longenecker, Sims and Gioia, 1987.65 Ibid. (Longenecker et al., 1987).66 Benedict & Levine, 198867 McEvoy, G.M. and Cascio, W.F., 1990. The Unites States and Taiwan: Two
Different Cultures Look at Performance Appraisal. Research in Personnel and Human Resources Management (Supplement 2), pp. 201-219.
68 Cascio, W.F., 1998. Managing Human Resources. Irwin McGraw-Hill, 5th edition, p. 327.
69 Ibid.70 Peter Principle71 Cascio, W.F., 1998. Managing Human Resources. Irwin McGraw-Hill, 5th edition, p.
297.72 London & Bray, 198073 Drofman, P.W., Stephan, W. G. and Loveland, J., 1986. Performance Appraisal
Behaviors: Supervisor Perceptions and Subordinate Reactions. Personnel Psychology, 39, 579-597.
74 Porth, S.J. Strategic Management, 2003 adapted from Kaplan, R.S. and Norton, D.P., “The Balanced Scorecard-Measures that Drive Performance,” Harvard Business Review. January –February 1992, p. 72.
CJBE 2006 37