Michelle Maru
The issue of honesty in the workplace is one that is of great interest to employers of all sizes. Employee theft can range from the stealing of large office equipment to taking extended breaks. It has been estimated that 50 % of employees engage in some sort of workplace theft (Bergmann, Mundt, Illgen, 1990). Anonymous survey results from employees across various industries indicated that �substantial proportions of employees admit to the theft of money and merchandise� (Ash, 1991, p. 4.). Various methods of combating this problem and lowering the incidences of employee theft have been developed including both the polygraph and paper and pencil tests. However, it is important to note that neither of these methods were developed by the psychological community in order to predict employee honesty. In fact, �the development of paper and pencil tests�is a phenomenon that has taken place largely outside the mainstream of psychological testing� (Sackett and Harris, 1984, p. 222). This paper will present both the polygraph and the paper and pencil test methods of predicting employee theft as well as examining both privacy and ethical concerns that surround honesty testing.
The polygraph was originally developed by the criminologist Lombroso, but was later brought to the United States by Hugo Munstergerg (Ash, 1991). It was brought to the forefront in 1923 when it was used to assess the truthfulness of murder suspects� responses (Saxe, 1994). The intended use of the polygraph was for criminal cases such as this, but by the mid-1940s, it began to emerge in the workplace (Ash, 1991). There were two major workplace applications of the polygraph. The first is very similar to that of criminal investigations where there was �some specific act or event being investigated and a small number of individuals were believed, on the basis of some reasonable evidence, to be involved in the event� (Murphy, 1993, p. 88). The other way that the polygraph is utilized by employers is to �conduct a very general inquiry into the character of the individual� applying for a position (Murphy, 1993, p .89). The use of the polygraph for pre-employment screenings was widespread prior to the late 1980�s, a phenomenon unique to the U.S. (Murphy, 1993).
A common myth of the polygraph is that the machine determines the deceptiveness of the individual being tested. This is not true. There are several stages to the examination process that when analyzed together determine the truthfulness of responses given. The polygraph examination generally consists of a pre-examination interview, a demonstration designed to display the accuracy of the machine�s ability to detect deception, questioning of the examinee while noting physiological changes and finally, a post-examination interview (Murphy, 1993). The goals of the pre-examination interview are to obtain initial information from the examinee, explain procedures, and review the questions that will be asked during the questioning stage. These sets of questions usually consist of very general questions as well as questions designed to get the examinee to make immediate admissions (Murphy, 1993). The stim test, or demonstration designed to convince the examinee of the polygraph�s ability to detect deception, often involves deceiving the examinee in some way. While intuitively this seems inappropriate for a test designed to get at honesty, it is critical to the success of the examination that the examinee believe the polygraph works. If this is not the case, the individual taking the test will be much less likely to make damaging admissions (Murphy, 1993). The actual polygraph examination follows this stim test. Although various forms of polygraph questioning exist, the most common is referred to as the control question test (CQT). The theory behind this line of questioning is that a deceptive individual should show greater physiological arousal to relevant questions such as �Did you steal $500 from the cash register?� than to control questions such as �Have you ever stolen anything of value?� (Saxe, 1994). These questions are repeated several times and the examinee�s physiological responses are recorded. In it not uncommon for the examiner to leave the examinee alone following this questioning for extended periods of time in order to score the results. One reason for this is to give the examinee time to think about the responses just given (Murphy, 1993). The post-examination interview is the final step in the polygraph process and is considered to be the most important. It is during the post-examination interview that psychological coercion is most likely to occur in attempts to induce confession (Murphy, 1993). The results of the polygraph examination are determined in various ways. The easiest determination to make is when an admission occurs. Frequently, no further analyses are necessary. However, in the cases in which individuals do not confess, it is left to the discretion of the examiner as to the honesty of the examinee. It is not uncommon for an examiner to review all of the information gained through the various stages of the process and decide that there is not enough clear evidence to make a accurate determination as to the truthfulness of the examinee�s responses (Murphy, 1993).
The psychometric properties of the polygraph have had a great impact on the permissibility of examination results. Inter-rater reliability is often the most cited evidence for polygraph reliability. In criminal contexts, many inter-rater reliabilities range from .8 - .9. However, it is important to note that �as long as both interrogators fail the great majority of examinees, they will almost always agree� (Murphy, 1993, p.95). To further illustrate this point, Murphy (1993) states that an individuals who confesses will undoubtedly lead two different examiners to the same conclusion. Despite these criticisms, adjustments to reliability coefficients can be made to correct for some of these problems and it appears as though trained polygraph examiners do make reliable decisions in criminal investigations. However, the reliability of judgements made in employment settings has not been demonstrated as well due in part to the lack of methodically sound studies (Murphy, 1993).
When looking at the validity of the polygraph, two major problems must be taken into account. The first is that it is difficult to assess exactly what it is that needs to be validated. This relates back to the myth that the polygraph machine determines the truthfulness of the responses, when in fact it is the polygraph examiner who makes the determination. Thus, the question is whether the machine or the man needs to be validated (Murphy, 1993). The second major problem with assessing validity evidence is criteria contamination. This can result from the examiner inducing confession as was discussed in the post-examination interview process. When a confession occurs, examiners don�t review the polygraph data to ensure that they indicate guilt. Thus, the validity of the polygraph can�t be evidenced by these confessions (Murphy, 1993).
With these concerns in mind, the major evidence for the validity of the polygraph has demonstrated that it is somewhat valid in criminal investigations, but that there is not enough evidence to support its use in employment settings. In fact, there is only a small amount of available evidence to indicate that it makes employment decisions any better than those that could have been arrived at by chance alone (Murphy, 1993).
In December of 1988, the U.S. enacted the Employee Polygraph Protection Act. �This act substantially prohibits the use of the polygraph in pre-employment screening in the private sector, defers to state laws, which are more stringent than federal laws, and excepts from its provisions the use of the polygraph for screening applicants to sensitive positions in the private sector and applicants for law enforcement, security, intelligence and military positions in the public sector� (Ash, 1991, p.5). The Employee Polygraph Protection Act severely limits the employment use of the polygraph examination; however, specific circumstances in which the polygraph may be used is outlined in Murphy (1993, p.101). There are three major reasons for the passing of this act. First, there wasn�t enough scientifically acceptable research to indicate that the polygraph accurately predicted job applicant�s propensity toward theft. Second, the results of the polygraph seemed to depend upon the training and skill of the examiner. This may be due in part to a lack of standards for polygraph examiners. Finally, the polygraph examination process was not viewed favorably by applicants (Jones and Terris, 1991).
Following the passing of the Employee Polygraph Protection Act, employers needed to find alternative methods to assess the honesty of potential employees. One method many employers have turned to is the use of paper and pencil integrity tests. There are various tests available to employers, but they all seem to have some common characteristics. First, they measure attitudes related to at least one of the following psychological constructs: tolerance of those who steal, projection of other�s theft, acceptance of rationalizations for theft, general antisocial beliefs/behaviors and admissions of theft or related behavior (Jones and Terris, 1991). In addition, Murphy (1993) points out that all integrity tests attempt to uncover general thought patterns tied to dishonesty and that they examine responses to hypothetical situations. Sackett and Harris (1984, p. 223) offer the following examples of integrity test questions: �What percentage of people take more than $1.00 per week from their employer?; �Should a person be fired if caught stealing $5.00?� and �Do you know for certain that some of your friends steal from their employer?�
There are two major types of paper and pencil integrity tests. While the names used to refer to them may vary, the basic difference is in the formation of questions. Generally, there are overt or clear-purpose integrity tests, calling for such things as direct admissions of past theft, and there are covert or veiled purpose integrity tests, aimed more at assessing personality traits associated with honest or dishonest behavior. Covert integrity tests are generally �scored on sub-scales that measure constructs such as �employee deviance� and �reliability�� (Saxe, 1994, p.70). It is important to note however, that the distinction between the two tests isn�t always as easy to make as it may seem. �Many clear-purpose tests include items, scales, and the like that are not obviously related to integrity, while many veiled-purpose tests include items and scales that seem very similar to those characteristics of clear-purpose tests� (Murphy, 1993, p.117). One of the most common examples of a covert or veiled-purpose test is the Rorschach Inkblot Test. While much of the scoring of this test is subjective, �keys have been developed to differentiate between the delinquent or criminally prone and the non-delinquent��(Ash, 1991, p.7). In contrast, the most common overt or clear-purpose integrity test used today is the London House Personnel Selection Inventory. �Currently available in 12 forms, tests in the series measure honesty (all forms), drug avoidance, nonviolence, employee/customer relations, emotional stability, safety, work values, supervision and response validity (all forms)� (Ash, 1991, p.9). This measure is strongly supported by a wide base of ongoing validity research examining predictive as well as concurrent evidences. While some will argue that there is no clear definition of integrity or honesty, therefore tests designed to measure it are not valid; �research seems to show that many integrity tests provide information about the overall likelihood of dishonest behavior, and that a person with poor scores is more likely to behave in a dishonest fashion across a wide range of situations than a person with higher scores� (Murphy, 1993, ps. 177-8).
The possibility of applicants giving socially desirable answers or �faking� exists in paper and pencil honesty testing due to its self-reporting nature. The most common method of reducing the effects of faking is to include a distortion scale on the scoring of the test. Additionally, �instructions state that any attempts to provide inaccurate answers can be detected and invalidate the test results� (Jones and Terris, 1991, p.47). However, caution is in order when reading theses distortion scales. Distortion scales will indicate applicants who have not been truthful in their responses; however, it is not able to predict or infer what the response should have been (Murphy, 1993). In these instances Murphy (1993) recommends discarding the results as opposed to correcting them.
The psychometric properties of honesty tests, specifically validity, is difficult to assess partly because of the lack of good criteria (Murphy, 1993). Murphy (1993) describes 5 methods of assessing honesty test validity as well presenting possible difficulties with each of them. The first is called the contrasted groups method. Here, tests scores of dishonest people are compared to scores obtained by individuals who have not been shown to be dishonest. However, �it is probably much easier to discriminate between convicted felons and people in general than it is to discriminate among job applicants who may or may not engage in small-scale employee theft� (p.124). A second method, the background check method, correlates the results of the background check with the results of the honesty test. The problem with this is that �many of the most frequent and, in the long run, most costly forms of property and production deviance will not show up on a standard background check� (p.124). The admissions method of assessing validity uses confessions of prior theft to validate honesty tests. Unfortunately, �you never know whether all admissions are truthful or whether admissions of all instances of theft or other undesirable behavior can be obtained� (p.124). Predictive validity studies are also possible by using future theft as a criterion. The problem with this method is that employers would have to hire all applicants regardless of selection measures then wait and see what happens. This is just not practical in most cases. The final validity strategy is the time series method. Here, data is collected regarding such things as job performance and losses both before and after honesty testing is implemented, and the results are compared. While many time series studies show increases in job performance and reductions in shrinkage, �the simplest explanation is that the introduction of an integrity test heightens everyone�s awareness of employee theft, which will itself reduce theft opportunities� (p.125). While this information seems discouraging, the validity of paper and pencil honesty tests is relatively high. Measures such as the London House Personnel Selection Inventory, which has undergone extensive validity research, has been shown to be �a reliable and valid instrument for use in identifying potential thieves among job applicants� (Werner, Joy and Jones, 1991, p.54). Further, Bergmann et al. (1990, p.219) stresses that �even though validity is constantly being challenged, there is a consistency of positive finding across all paper and pencil honesty tests.� Due to the various validity challenges, Bergmann et al. (1990) recommends that employers: 1) verify test publisher�s claims of validity through independent research, 2) be sure that numerous validity techniques have been utilized and that they are ongoing, 3) evaluate the amount of support available from the test publisher and 4) be sure that validation information is provided, when requested, by the test publisher. Since the Employee Polygraph Protection Act of 1988, the use of paper and pencil honesty tests has continued to increase. A few of the many reasons for this is the inexpensive nature of paper and pencil tests, the lack of specialized equipment necessary and the legal nature of paper and pencil tests (Bergmann et all. 1990). Honesty tests can be administered to job incumbents as well as to job applicants. When administrated to incumbents honesty tests are usually a part of a focused investigation (Murphy, 1993). However, implementing an honesty testing program to job applicants has many potential benefits. Employers can expected to see reductions in screening costs as well as reductions in both crime rates and terminations due to employee theft (Werner et al., 1991). Additionally, Werner et al. (1991) points out that increases in profitability can occur from the use of honesty testing in at least three major ways. First, large amounts of money don�t need to be spent on surveillance cameras; second, large security forces are not necessary, thus saving the organization money; and third, the products as well as the supplies used to produced those products are not as likely to need replacing due to lower levels of employee theft. Further, honesty testing programs increase productivity (Werner et al, 1991). Some possible reasons why this increase is productivity is seen may be �that honest employees are less likely to engage in time theft by extending their work breaks without authorization �(and) that honest employees report less stress from their work and more job satisfaction��(Werner et al., 1991, p.58).
While honesty tests are designed to predict honesty, they also serve the important function of predicting job performance. Current research indicates that although individuals who don�t score well on honesty tests may not actually commit crimes of theft, they do in fact turn out to be poor employees as evidenced by absences, lates and general irresponsibility (Murphy, 1993). While it is clear that predictions of job performance result from honesty testing, it isn�t clear as to why. Murphy (1993) offers the explanation that the nature of honesty tests is to predict behavior which in and of itself is counterproductive. It then only seems logical that honesty tests would be accurate predictors of (poor) job performance. When examining the use of honesty testing in pre-employment settings, one must consider the examinee�s right to privacy. This has been a central criticism of both paper and pencil tests as well as of the polygraph. When looking at an individual�s right to privacy in the workplace, three major sources must be considered. First, �common law torts serve to protect applicants against unreasonably intrusive selection programs and any adverse publicity that could place an applicant in a false light before the public� (Jones, Ash, Soto and Terris 1991, p.231). Federal laws, specifically the U.S. Constitution, do not clearly state an individual�s right to privacy; however, these rights are inferred from the First, Fourth, Fifth, Ninth and Fourteenth Amendments (Jones et al., 1991). Finally, state constitutions and laws can offer additional protections above and beyond those already mentioned (Jones et al., 1991). In order to ensure that an individual�s right to privacy is being upheld, employers must take care when selecting an honesty test, when administering it and when scoring it. When selecting an honesty test, employers should be sure that it evaluates job�related attitudes and behaviors. These types of tests are �less likely to be perceived as invasive compared to tests that ask questions about issues that are unrelated to the workplace�(Jones et al., 1991, p.232). A second consideration when selecting an honesty test is to be sure that it complies with established guidelines. One such guideline is the Model Guidelines for Pre-employment Integrity Testing Programs that will be discussed in greater detail later in this review. When administering honesty tests it is important that administrators are properly trained and that the test is given in a peaceable and relaxed environment. (Jones et al., 1991). It is also imperative that applicants understand that the honesty test �is merely one part of an overall personnel selection program� (Jones et al., 1991, p.233). Additionally, Jones et al. (1991) point out that applicants need to sign an informed consent form indicating that they are familiar with the instructions and that they authorize the appropriate individuals to evaluate and review their responses. A final concern regarding the administering of honesty tests in the workplace is that all applicants need to be subjected to the exact same selection measures (Jones et al., 1991).
In scoring and interpreting the results of honesty tests, employers need to uphold the applicants right to fair and accurate scoring. In doing so, employers must understand that honesty tests do have some degree of classification error, therefore labeling of applicants based on their results is not acceptable (Jones et al., 1991).
Finally, employers need to secure not only the honesty test results, but the used test booklets and unused templates as well. Theses materials should only be accessible to company representative who have a legitimate business necessity to see them and they should never be disclosed to anyone else including prior or future employers (Jones et al., 1991). In addition to privacy concerns, honesty tests present ethical concerns for the employers administering them. However, Sackett and Harris (1984) point out that only invalid honesty measures are unethical and that �selection procedures with modest validities are the best psychology has been able to devise�� (p.242). Thus, when considered in relation to other measures such as interviews, references and application blanks, valid honesty tests add to the predictive power of the overall selection process.
Experts in the fields of psychological testing and personnel law developed the Model Guidelines for Pre-Employment Integrity Testing Programs. It outlines the responsibilities of both the test publishers and the test users. The five primary topics addressed by the Model Guidelines are: test development and selection; test administration; scoring, interpretation and accuracy; test fairness and confidentiality; and public statements and test marketing practices (Jones, Arnold and Harris, 1991).
With regards to test development and selection, the Model Guidelines indicate that test publishers �provide a specimen test, directions, answer sheets, manuals, administration guides, sample score reports and procedures for proper test use� (Jones, Arnold, et al. 1991, p.240). Further, they are responsible for describing what the test measures, reporting who the test is appropriate for, providing validity data and an acknowledgement that no test is a perfect predictor of job performance (Jones, Arnold, et al., 1991). Test users are responsible for reading and evaluating this material for its applicability to the group it intends to test as well as being able to demonstrate the business necessity of the honesty test selected (Jones, Arnold et al., 1991).
Honesty test publishers are required, according to the Model Guidelines, to provide training regarding the proper use of the test including, but not limited to, a standard set of administration instructions, test scale descriptions and restrictions on the use of the test (Jones, Arnold et al., 1991). Test users are responsible for using this information to �establish, implement, and enforce a policy on the proper use of the testing program� (Jones, Arnold et al., 1991, p.242). When examining scoring, interpretation and accuracy of the honesty tests, the Model Guidelines indicate that test publishers should have professionals available to assist clients with integrating the test into the organizations� selection process and establishing fair, non-discriminatory cut-off scores (Jones, Arnold et al., 1991). Test users are responsible for checking the responses to see that all items were answered and marked correctly as well as for accurately relaying those responses to the test publisher (Jones, Arnold et al., 1991). It is the test publishers who should ensure that honesty tests do not unfairly discriminate against any protected group of individuals; therefore they should avoid clearly invasive and/or offensive questions and offer the test in foreign languages (Jones, Arnold et al., 1991). However, it is the responsibility of the test users to maintain information needed to conduct adverse impact studies as well as to establish an inventory control method for all test booklets, both used and unused (Jones, Arnold et al., 1991).
When test publishers market their products, the Model Guidelines indicate that they must do so �fairly and accurately, avoiding misrepresentation, exaggeration, sensationalism and superficiality� thereby using actual research as a basis for their claims (Jones, Arnold et al., 1991, p.244). When faced with these claims, test users must solicit the aid of trained professionals if any doubt as to their validity exists (Jones, Arnold et al., 1991).
Honesty in the workplace in an important and complex issue that all employers face. While the use of honesty tests is not a foolproof method of predicting future behavior, the available research indicates that it is both a reliable and valid measure when incorporated into an organization�s overall selection process.
REFERENCES
Ash, P. (1991). A History of Honesty Testing. In J.W. Jones (Ed.), Preemployment Honesty Testing: Current Research and Future Directions (pp. 3-19). New York: Quorum Books.
Bergmann, T. J., Mundt Jr., D. H., & Illgen, E. J. (1990). The Evolution of Honesty Tests and Means for their Evaluation. Employee Responsibilities and Rights Journal, 3, 215-223.
Jones, J. W., Arnold, D., & Harris, W. G. (1991). Model Guidelines for Preemployment Integrity Testing: An Overview. In J.W. Jones (Ed.), Preemployment Honesty Testing: Current Research and Future Directions (pp. 239-246). New York: Quorum Books.
Jones, J. W., Ash, P., Soto, C., & Terris, W. (1991). Protecting Job Applicants� Privacy Rights When Using Preemployment Honesty Tests. In J.W. Jones (Ed.), Preemployment Honesty Testing: Current Research and Future Directions (pp. 229-238). New York: Quorum Books.
Jones, J. W., & Terris, W. (1991). Selection Alternatives to the Preemployment Polygraph. In J.W. Jones (Ed.), Preemployment Honesty Testing: Current Research and Future Directions (pp. 39-52). New York: Quorum Books. Murphy, R. (1993). Honesty in the Workplace. Pacific Grove, CA: Brooks/Cole Publishing Co.
Sackett, P. R., & Harris, M. M. (1984). Honesty Testing for Personnel Selection: A Review and Critique. Personnel Psychology, 37, 221-245.
Saxe, L. (1994). Detection of Deception: Polygraph and Integrity Tests. Current Directions in Psychological Science, 3, 69-73.
Werner, S. H., Joy, D. S., & Jones, J. W. (1991). Improving Corporate Profitability with Preemployment Integrity Tests. In J.W. Jones (Ed.), Preemployment Honesty Testing: Current Research and Future Directions (pp. 53-61). New York: Quorum Books.
