Note: The following is based upon a portion of an earlier work.
For the complete version of the original publication, see:

Downing, L. L. (1994).  Criterion shaped behaviour: Pitfalls of performance appraisal.
International Journal of Selection and Assessment, 2, 1-21.

PART 3

CSB RELATED PROBLEMS WITH CRITERION
CONTAMINATION AND CRITERION DEFICIENCY

PROBLEMS OF CRITERION DEFICIENCY
        A measure will be Criterion Deficient to the extent that it fails to assess one or more of the behaviors or characteristics of the Ultimate Criterion, i.e., of the hypothetical, "ideal" performance.  Deficiency CSBs will include reductions in any factors of ideal performance that are not measured.  To know what these will be requires a list of Ultimate Criterion Factors, and a list of Actual Criterion Factors.  A hypothetical list, pertaining to teacher evaluation, will be used for purposes of illustration (see Figure 2)

Actual Criterion Factors Ultimate Criterion Factors
1. Average score of students on national 
     objective test.

2. Percentage of students who eventually
    graduate.

 1. Preparation of students to take objective tests.

2. Preparation of students for eventual graduation.

3. Teaching students to write effectively.

Figure 2: Hypothetical Situation Illustrating Criterion Deficiency
 

        We will assume that the Actual Criterion Measure adequately assesses Factor 1, and Factor 2.  These are also part of the Ultimate Criterion, and are therefore Criterion Relevant.  Increases in these factors are desirable, and to the extent that teachers believe they can change their behavior to increase them, and that valued outcomes will be contingent upon such increases, and easier means of increasing scores are not available, these desirable Relevant CSBs should increase
        Factor 3 of the Ultimate Criterion is not assessed, however, and efforts to increase it would not be expected to result from imposition of this evaluation system.  In fact, to the extent that time and energy spent on increasing the students' performances on objective tests and on preparing them for eventual graduation interferes with prior efforts to teach them to write effectively, these efforts may actually decrease.
        Let us look more closely at these various factors and at how and why some get measured and others do not.  In this case, it seems that those factors most readily measured by objective means were included in the Actual Criterion Measure, while the one most difficult to measure by objective means, ability to write effectively, did not get measured at all.  Factors that are not measured are usually not left out because they were deemed to be less important.  They are more likely excluded because no measure of them was available, or measures of them were too costly or time consuming, or perhaps because measures of them were too subjective, and hence, in our terms, contaminated, by bias, prejudice, etc.
        This simple illustration is not an example of a rare failure of the system to measure an important, or even a critical, factor of performance.  It is commonly the case that the most important factors are the most difficult to adequately assess, and are consequently, the most likely to be excluded.  Few would contest that ideal teacher performance should produce students who are able to write effectively; to think critically, creatively, and independently; and to do oral presentations of material effectively.  If ideal teachers produce increases in such student behaviors as these, then the Ultimate criterion of teacher performance would assess them.  It is rare, however, that such factors are measured by becoming part of the Actual Criterion, and rarer still that they are measured adequately.  It is even less likely that such factors will be measured on objective tests, such as are most likely to be used to provide a board-based national standard of student, or of teacher, performance.
        What we will find is that these important but difficult to measure factors, when they do get measured, are not often done so by objective tests, but by some form of subjective rating.  This may be done by the principal, who presumably has access to a broad base of information concerning a teacher's effectiveness.  The potential for bias in such ratings may render them relatively useless for purposes of evaluating teachers.  Elimination of such bias is necessary to reduce problems of Criterion Contamination, a problem to be addressed in the following section of this paper.  Our current concern with Criterion Deficiency is that the problems of bias in the measurement of important factors may result in those factors not being measured at all.  Furthermore, if they are not measured, Deficiency CSBs will result not only in an absence of increase, but in an actual decrease in these desirable factors.  Attempts to develop adequate unbiased measures of such aspects of teacher performance may be devised, but a recurrent tradeoff will become evident in such efforts.  Methods effective at reducing Criterion Deficiency will nearly always increase subjective influences, thereby increasing Criterion Contamination, and Contamination CSBs.
        The most common method for eliminating Criterion Deficiency is to use some form of Subjective, global, rating scales.  Such ratings are best done by a person who is broadly knowledgeable of the many facets of performance that make up the Ultimate criterion.  These ratings could take the form of a simple question, asked of one's peers, or supervisor, such as, "How effective is _____ at being a sixth grade teacher?"  The response might be on a seven-point scale from 1 equals "extremely ineffective," to 7 equals "extremely effective."  Such a measure allows for all of the factors that make up the ideal performance of a sixth grade teacher to have an impact on the performance evaluation.  To ensure that the rater did take all major factors into account, numerous scales, one for each factor judged to be important, might be used, with a total or average of these serving as the performance evaluation.  It can be reasonably argued that such measures are low in Criterion Deficiency, and it is for purposes of reducing Criterion Deficiency that such measures are typically employed.  Certainly, in terms of Criterion Deficiency, subjective ratings are better than most available objective methods of performance evaluation.  This is especially likely to be the case in evaluation of professionals, students, or others whose roles are complex, and who ideal performances involve numerous dimensions that are difficult or even impossible to assess by more objective methods.
        Unfortunately, the benefits to be derived from such subjective ratings are nearly always offset, to some degree, by a host of problems.  Attempts to achieve national goals and meet national standards will require that evaluations be standardized.  While this is fairly easy to do with objective measures, such as multiple choice tests, it is nearly impossible to do with procedures requiring firsthand, long-term experience with and observation of the person being evaluated.  To do so would require equivalent amounts and types of observations by all raters of all those being rated, as well as an invariant interpretation by all raters of the meaning of and weights to be attached to all relevant performance factors.  To do this would require standard educational practices and intensive training of raters, the first of which is probably undesirable and the second of which is probably impractical or impossible.  Therefore, it can be expected that any national system will tend to minimize or eliminate such subjective methods, the value of which is in their ability to reduce Criterion Deficiency, and will opt instead for the most objective and easily standardized methods, for which Criterion Deficiency and Deficiency CSBs will be inevitable.  In a national, standardized program for evaluation of teachers, we would expect that those performance factors that cannot be readily measured by objective tests, e.g., teaching students to write effectively, will not get measured; and to the extent that positive incentives become attached to other factors, teaching students to memorize facts, that the level of those unmeasured factors will be expected to decline.

PROBLEMS OF CRITERION CONTAMINATION
        A measure will be Criterion Contaminated to the extent that scores can be influenced by increasing factors that are not part of the Ultimate Criterion.  Contamination CSBs will include efforts to increase these factors, and more importantly will include decreases in Criterion Relevant and Criterion Deficient Factors resulting from such efforts.  To illustrate these Criterion Contamination effects, let us devise a second set of hypothetical lists (see Figure 3).
 

Actual Criterion Factors Ultimate Criterion Factors
1. Average score of students on national 
     objective test.

2. Percentage of students who eventually
    graduate.

3. Supervisor ratings of the writing
    effectivenessof the teacher's students.

    a. Reflecting actual quality of students' 
        writing.
    b. Reflecting how much the rater likes the
        teacher.

  1. Preparation of students to take objective
     tests.

2. Preparation of students for eventual
     graduation.

3. Teaching students to write effectively.

Figure 3: Hypothetical Situation Illustrating Criterion Contamination

        We will assume that the Actual Criterion Measure adequately assesses Factors 1 and 2, which overlap with the Ultimate Criterion and are, therefore, Criterion Relevant, leading to desirable Relevant CSBs.  We will also assume that teaching students to write effectively, Factor 3, has been assessed by the supervisor's rating, and is no longer Criterion Deficient.  Unfortunately, in assessing Factor 3 with a subjective rating by the supervisor, we have unintentionally added a contaminating bias factor by means of which it is possible for teachers to increase their scores through other, possibly easier, and incompatible Contamination CSBs.  In our example, if efforts to be liked by the supervisor prove to be an easier and more reliable means of improving one's rating, and hence one's total score on the Criterion Measure, than do efforts to enhance Factors 1, 2, and 3a, and if those efforts take time and energy away from the pursuit of 1, 2, and 3a, we will expect a decrease in these Relevant Factors, as well as decreases in any desired but unmeasured Deficiency Factors such as were discussed previously.
        The effects of Criterion Contamination on reductions in Criterion Relevant and Criterion Deficient Factors of the Ultimate Criterion can be devastating, but even were these effects somehow miraculously avoided, increases in Contamination Factors may have other detrimental effects on the organization.  The variety of undesirable consequences that can be expected as a result of Criterion Contaminated measures requires some elaboration.  In the following section will be presented many such anticipated pitfalls of evaluation systems, and attempts to shed some light on when each type of effect is most likely to occur.

Undesirable Consequences From the Use of Criterion Contaminated Measures.

        It is generally the case that Contamination CSBs are more of a problem when using subjective rating methods than when using objective testing methods to evaluate performance.  This is a result of the fact that raters are highly susceptible to influences that systematically bias their evaluations of others, while such evaluator biases do not usually influence scoring of objective tests.  Nevertheless, Contamination does also occur with the use of objective tests.  Some of the major types of Contamination CSBs that can influence objective test scores are presented here.

        Contamination Effects on Objective Tests: Contamination CSB's that are most likely when using objective tests fall into three categories:

         1. Cheating.  One category of Contamination CSBs includes the numerous versions of cheating, all of which refer to efforts to increase Actual Criterion Factors without increasing Ultimate Criterion Factors.  The Contaminating Factors in this case include all possible illegitimate means of obtaining a high score.  These include copying from someone else's test, obtaining a key of correct answers before the test, concealing written answers on "cheat sheets," written on parts of one's body, or hidden in the restroom, and having someone else take the test under the name of the person to be evaluated.  Most students and teachers have also encountered more original methods tailored for special testing situations.  Most of these methods are more common when objective tests are being employed, but some would suffice for nearly any type of evaluation system.

        2. "Cutthroat Methods."  A special variety of Contamination CSBs is applicable to any evaluation system in which one's score, or the value of consequences contingent upon a given score, is determined by one's performance relative to that of others.  A ranking system, and "grading on a curve," are examples of such systems.  One illegitimate way to improve one's score is to actively undermine the performance of others whose performance is being evaluated.  This can be accomplished in many ways.  I have known of students in a competitive law school to steal, mutilate, or "misplace" law books in the library that were essential to the adequate performance of classmates on papers or exams.  I taught at one competitive undergraduate liberal arts college where students who used such tactics were openly referred to as "throats," short for cutthroats.
 Where teachers are being evaluated for raises, promotions, or tenure, one would hope that such potentially destructive undermining of colleagues is rare, and perhaps it is so, but many teachers seem to know of at least one instance where even more brutal internecine CSBs occurred.

        3. Loss of Mutual Cooperation.  More common, perhaps, than cutthroat tactics, but possibly more damaging to the organization, is the widespread tendency for competitive evlauation systems to discourage interdependent cooperation of group members.  A system in which members believe that their own self-interest is incompatible with the interests of their colleagues is not likely to be either satisfying to individuals or productive for the organization.  This can be a particularly pernicious effect in schools, where it is often the case that fellow teachers are a teacher's most valuable resource.  The cost to an organization of undermining the potential for cooperation between members is far from a trivial consideration, yet seldom are such anticipated problems taken into account by those advocating new methods for evaluating teachers and making them more accountable.

        The use of "merit money" in schools will almost always be subject to the above effects, even if the evaluation system does not appear to be of the ranking variety.  This is because the amount of money to be distributed amongst teachers is always limited, and however performance evaluations were initially done, for purposes of allocating merit money some form of ranking must be devised from them.  Any system in which teachers perceive that their own raise or bonus will be less if their colleagues perform well, than if they perform poorly, is subject to the types of undesirable consequences described.
        The prominence of efforts such as those described above is a result of the fact that these Contamination CSBs are perceived as requiring less effort, or are believed to have a high expectancy of success at achieving valued outcomes.  The negative experiences of guilt or shame for having violated laws, rules or norms designed to further social, institutional or organizational goals, should be expected to offset these gains for individuals.  To the extent that they are personally subject to such emotions these will be costs, or negative utilities.  The unfortunate derivation from this set of premises is that those most likely to "succeed" in such a system are precisely those most lacking in such scruples.  As a nation that has experienced Watergate, Iran-Contra, the HUD scandal, the savings and loan crisis, and other debacles too numerous to name, one might question the advisability of imposing more systems in which "good guys" are destined to finish last.

        Contamination CSBs in Subjective Rating Systems.  Many of the Contamination CSBs common to the use of objective tests, such as some forms of cheating or of undermining the performance of others, are also potential problems for use of subjective rating systems.  Because rating systems typically depend upon long-term exposure to and knowledge of the person being rated, however, some forms of cheating are less likely than with objective tests.  The types of contamination that are most often important in the use of subjective measures primarily involve rater bias.

        Rater Bias. The amount of research that has bben conducted on the sources and varieties of rater bias is staggering.  This research is usually presented as a demonstration of sources of invalidity in subjective rating systems of evaluation, all of which undermine the usefulness of such methods for purposes of assessment (cf. Smithers, 1988).
        From the perspective of Criterion Shaped Behavior Theory, this research provides a basis for understanding the numerous means by which the individual being evaluated can, by increases in Contamination CSBs, increase scores on the Actual Criterion Measure.  This class of CSBs all involve the active exploitation of those biases that cause ratings to reflect factors other than those contained in the Ultimate Criterion.  Little has been formally written about these effects, nor do other theories show the functional relationships between the various types.  Anyone who has ever worked in an organization where performance evaluations were used as a basis for distribution of valued outcome, however, will immediately recognize many of them.  They are often known by colorful names which reflect the unacceptable or distasteful connotations associated with them, or with those who use them.
        One such category of Contamination CSBs contains Asskissing, Brown-nosing, Bootlicking, and Sucking-Up-To-The-Boss.  In more polite language these are called Ingratiation, or Flattery.  At the extreme they may involve granting of major Favors, giving of expensive Gifts, or Sleeping With The Boss.  Where a quid pro quo is involved, these may become Bribery.  All of these involve providing the evaluator with something of value.  This may be the illusion of superiority, the semblance of power, the enhancement of ego, sexual compliance, or more material commodities including gifts, or money.  Making any of these available to an evaluator is a Criterion Shaped Behavior if it results from an expectation of more favorable ratings on a performance evaluation measure.  Where a subjective element exists in the method used, as is nearly always the case with the use of ratings, research points to several sources of such a bias.

Exploiting the Interpersonal Attraction Bias

        Ingratiation Tactics.  Considerable research shows that raters consistently favor those who they personally like in their performance ratings (Berscheid and Walster, 1978).  Much of this will occur unintentionally, for raters who have formed a positive impression of an individual will be susceptible to a "Halo Effect."  People perceive and interpret new information in ways supportive of earlier impressions.  They selectively perceive, selectively store, selectively retrieve, and selectively interpret information in a biased fashion directed to sustaining those early impressions.  People also are biased by a belief that all positive characteristics are positively correlated with each other (Bruner and Tagiuri, 1954), and any early attribution of positive traits or behaviors, e.g., being likeable, increases expectations that the liked person will also possess other positive traits, e.g., being a good teacher.  Anything one can do to increase how much he or she is liked by the rater will potentially lead to receiving a higher score on a subjective measure of performance.
         The tactics listed earlier, Brown-nosing, Bootlicking, and other unsavory categories of ingratiation (Jones, 1964) can all be viewed as Contamination CSBs designed to enhance one's scores on the Actual Criterion Measure by increasing one's interpersonal attractiveness to the rater.  The means by which this can be accomplished are delineated in any recent textbook on social psychology (c.f., Brehm and Kasin, 1990).  A broad generalization of the many separate influences is that we like others to the extent that they have been associated with our own positive experiences (Byrne, 1971).  If one has good feelings because of, or merely in association with, a person, he or she will tend to like that person.  If one can enhance such good feelings in a rater, by flattery, by gift-giving, etc., one can increase liking and consequently how well one is rated on subjective measures of performance.  To the extent that it is easier to increase one's score by use of such Contamination CSBs, one will be less likely to expend the effort required for real improvements in performance, i.e., Relevant CSBs.

        Exploiting the Similar-To-Me-Bias.  Perhaps the most thoroughly documented causal factor in interpersonal attraction is perceived similarity.  People like best others who they perceive to have attitudes, beliefs, lifestyles, and group memberships similar to their own (Byrne, 1971).  Because similarity is not merely a correlate of attraction, but a cause of it, anything that increases an evaluator's perception that the person being rated is similar to himself, or herself, is likely to produce an increase in that evaluator's performance ratings (Wexley and Yukl, 1977).  Efforts to increase perceived similarity are, therefore, Contamination CSBs.
        Such behaviors will not always be undesirable, in and of themselves.  In fact, encouraging the subordinate to become more like the supervisor is the essence of role-modeling and mentoring, both of which can be quite desirable to the organization.  The problem is that the Similar-To-Me-Bias will operate also for traits or characteristics that are irrelevant to or are even negatively related to the Ultimate Criterion, and that efforts to increase perceived similarity may interfere with Criterion Relevant behaviors.
        If the boss spends all weekend watching football games and smoking cigars, for example, it is not evident that increases in such behaviors will necessarily benefit the organization.  And if the supervisor is a member of the local Elk's lodge, it is to be expected that subordinates whose promotions are believed to be dependent upon being liked by that supervisor will be motivated to join also.  In this last example, it is ironic that the legal system has at times recognized the barring of women from such organizations as unfair discrimination.  While the legal argument typically focuses on women having reduced opportunities to engage in effective business transactions, in our terms women have been unfairly prevented from certain opportunities to exploit the personal biases of those engaged in subjective ratings of their performances.  Were there no Criterion Contamination, and no Contamination CSBs, such discrimination would be less of a factor.

        Exploiting the Reciprocity of Liking Bias.  A norm of reciprocity is a powerful influence on a wide variety of interpersonal behaviors.  Some view the norm as universal in its application (Gouldner, 1960).  One consequence of the operation of this norm is that individuals reciprocate liking, just as they reciprocate favors, gifts, insults and threats.  It is not simply that we like others who like us, but, at least to some extent, we like others because they like us; or more specifically because we believe they like us.  Ingratiation often involves the attempt to improve one's rating by pretending to like the rater.  This Contamination CSB may be relatively harmless, for it probably does not interfere much with actual performance, but as is the case with many CSBs, it can have other detrimental impact on the organization.  For example, in rewarding the disingenuous subordinate, less deceitful employees may feel inequitably treated, and resulting dissatisfaction and social friction may indeed interfere with productivity.

        Exploiting the Reciprocity of Behavior Bias.  The norm of reciprocity (Gouldner, 1960) requires that anything received from another be returned to them, if not in kind, at least in value.  In theories of social exchange (Thibaut & Kelley, 1959), the commodities of exchange include both material ones, such as money, and social or symbolic ones, such as approval or status.  As a norm, reciprocity reflects both an expectation and a social requirement.  Violation of the norm invokes both privately felt unease and socially mediated negative sanctions.  As a result, one powerful means of increasing the score one is given on a subjective rating, a valued outcome, is to previously give something of value to the rater.  What one gives can be anything valued by the rater, including approval, status, money, sex, etc.  Having received something of value, the rater will feel an obligation to return something of value to the giver.  One such commodity is an inflated, i.e., biased, score on the performance evaluation.
        An apple for the teacher, a donation to a congressman, pretended admiration of the boss, a case of Scotch for the purchasing agent's Christmas present, inviting the principal to dinner, and giving your unused NBA tickets to the superintendent are all potential means of influencing another's subjective decision relevant to one's own valued outcomes.  All of these are Contamination CSBs, for all can produce an increase in one's performance score without increasing factors which constitute the Ultimate Criterion.
 The undesirability of such Contamination CSBs is both a result of decreases in desirable behaviors in the relevant and deficiency sectors, and of inherently undesirable effects arising from such CSBs themselves.  While some such effects may be trivial, others involve wholesale attempts to subvert the official system and to replace it by one based upon such concerns as who owes who how much, rather than upon whose performance best promotes the goals of the organization.
        The reciprocity norm can function with or without the explicit conscious awareness by the rater of having been so influenced.  In extreme cases, however, such awareness does exist, and the rater knowingly participates in the illicit practice of exchanging favorable ratings for valued outcomes.  Though the distinction between being a dupe and being a conspirator is often difficult to define, bribery or quid pro quo exchanges are usually considered to be unethical or illegal, and evaluators found guilty of such conscious participation are subject to punishment or dismissal.  In fact, many forms of Contamination CSBs are viewed as unethical, or illegal, including the previously mentioned varieties of cheating.  Clearly, a system in which such behavior is successful at increasing performance scores used for allocation of valued outcomes is subject to many undesirable consequences.
        Reciprocity occurs for exchange of negative as well as positive outcomes.  People dislike, hate, or aggress against others who are believed to dislike, hate, or aggress against them.  In a system where everyone evaluates everyone else, one may fear that to give others negative evaluations may increase the chances of being negatively evaluated in return.  An expectation of reciprocity may lead a teacher to give inflated grades to students who, later in the semester, will be asked to evaluate the teacher.  I believe one would find a relationship between the increased use of student evaluation of teachers, in the 1970s, and the widely recognized trend toward higher grade point averages of students, in that same period of time, known as grade inflation.
        In peer review, giving a negative evaluation to a colleague's performance may rightly be expected to result in that colleague giving the rater a lower evaluation in return, hence contaminating the rating process with a well known positivity bias.  As with all of the other reciprocity effects, this may occur with or without conscious awareness of the evaluator.

        Additional Contamination CSBs.  Of the many potential sources of Criterion Contamination, all of which are problematic from an assessment point of view, some more easily lend themselves to the shaping, or manipulating, of rater's behaviors than do others.  For example, the "similar to me" bias may be exploited by taking up the boss's hobby, or by adopting the rater's political views, but cannot be exploited by becoming the same sex, race, age, ethnic group, etc.  As such, Contamination CSBs only occur related to biasing factors amenable to change.  So far we have looked primarily at those involving liking and reciprocity, but other means, not involving the relationship to the rater, but rather the limitations of the rater as an information processor, also exist.
        Exploiting the "Opportunity Bias."  From the assessment point of view it would be contaminating to evaluate the performance of two individuals on the same basis, if in fact they were provided with unequal opportunities to perform well.  For example, rating principals in terms of how many of the school's students go on to college, would be clearly unacceptable if one had a school in a poverty neighborhood, and the other had a wealthy suburban school.  Rating teachers based on performance of their students poses the same difficulty when teachers have classes differing in student ability, in class size, in content difficulty, etc.  Where the evaluation system has not eliminated such opportunity bias, it is sometimes possible to exploit it to one's own advantage through Contamination CSBs.  As a  college professor, I have observed a reluctance on the part of colleagues to volunteer to teach the most demanding courses in the year prior to a tenure decision, or to not teach the high enrollment introductory course until after one's promotion, or to not teach courses that meet at 8:00 A.M., or on Friday afternoon, because students will like them less well and hence rate the professor less favorably.  For the student, taking easy courses from lenient professors is a tried and true means of improving one's grade point average.  Efforts directed at maximizing performance scores by improving one's opportunities are a type of Contamination CSB.  Where such efforts detract from the pursuit of organizational goals, they are undesirable, and where they reward the whining and insistent complainer, and punish the cooperative team player, they promote the least desirable behaviors in those being evaluated.
        Exploiting Information Processing Errors.  Raters not only fail to adequately take into account the differences in opportunities afforded to those being evaluated, but they are subject to an array of well documented errors which systematically bias performance evaluations.  Some of these also can be exploited.  For example, ambiguity, confusion, lack of information, and uncertainty all increase the impact of a "central tendency bias," i.e., a tendency to give someone an "average" score.  An average score is good for an inadequate performer, hence such an individual will be motivated to create conditions in which the bias will operate.  An inadequate teacher may not invite colleagues or supervisors to sit in on classes, may not give students a chance to evaluate the course, and may otherwise, through what I call the "politics of obfuscation," make an informed evaluation of performance impossible.  Under such circumstances, a bias towards the average is more likely, and is clearly desirable from the point of view of the inadequate performer.  The other side of this is that the best performers will be eager to provide such information to insure that they are not rated as average, but as above average.  Both are Contamination CSBs, and both may take away from time or effort that might have been expended in improving actual performance.
        Among other information processing biases that can be exploited are the primacy effect, and the previously mentioned halo effect.  If it can be arranged such that one's best performances are made known early, the overall rating will tend to be higher than if one's poorest performances were made known first, even though all performances are made known prior to evaluation.  Likewise, an early positive impression based upon something other than performance, e.g., physical attractiveness, will increase subsequent ratings on other dimensions, including performance.  I tell my students, concerning applying for a job, to make their best qualities known first.  If they write well but are physically unattractive, they should introduce themselves first by letter.  If they have a lovely voice they might phone.  If they are physically attractive they should first appear in person.  These impression management tactics are Contamination CSBs dependent for their effectiveness on information processing biases of evaluators.
        One other bias capable of exploitation is the "availability heuristic," according to which one's overall impression is shaped inordinately by relevant instances most readily accessible to memory.  One can make a successful performance more accessible by attaching it to other memorable information, by "showcasing" it in some dramatic fashion, or by constant reminders of its occurrence.  One instance of prior success, so highlighted, may be more persuasive than ten unmemorable failures.  Getting one's successes in the paper, "publicity," or otherwise through "grandstanding," making them readily available to memory, while obscuring one's failures, all qualify as examples of Contamination CSBs.

        The importance of Contamination CSBs is that where they are perceived to be more efficient at achieving valued outcomes than are Relevant CSBs, they will lead to a reduction in desired behavior.  Furthermore, widespread occurrence of most of the described behaviors in an organization is likely to have other potentially devastating effects.  The lack of cooperation, the interpersonal enmity, the prevalence of cheating, stealing, and brown-nosing, the poor morale, and the job dissatisfaction often reported in educational as well as in other institutions can perhaps be attributed to the contamination CSBs likely to develop when highly valued outcomes are made contingent upon measures that are criterion contaminated.  Without evidence of powerful benefits also achieved by such performance evaluation, one might conclude that these are very high prices to pay.

REFERENCES
LINK TO PART 4
RETURN TO MAIN PAGE