I. Prevalence of Use

It is very difficult to calculate how many honesty tests are given each year. There has been a large increase in the number of these tests given in recent years, caused in part by passage of the Employee Polygraph Protection Act of 1988, which barred the use of polygraphs in most employment settings. (There is also concern that numbers may be rising because of the increase in security consciousness following September 11th.) Adding to the difficulty of calculation is the huge rise in the number of companies offering tests in recent years (one 1994 study identified 46 separate publishers; another 1996 report estimated that the testing industry was growing at 20% per year). Perhaps most challenging is the fact that a large proportion of the companies that use honesty tests prefer not to publicize this information.

A couple of reports did attempt to estimate the number of tests given each year. A 1995 report quoted estimates that 7500 to 10,000 “businesses, schools, agencies and other settings” were using the tests for about 9 to 12 million people annually , while another report from 1996 quoted estimates that 5000 to 6000 employers were using the tests and that 2.5 to 5 million tests were given annually . In addition, a 1992 Ernst & Young survey found that 47% of retail executives questioned admitted using the tests during 1990, while 19% planned to use them in the future. Additionally, 20% of the members of the Society for Human Resource Management reported using honesty tests.

For the most part, honesty tests seem to be used to screen applicants for lower-level, unskilled or semi-skilled jobs in areas such as retail and in wholesale/warehouse operations as well as financial institutions, areas in which employees have access to cash or merchandise. Furthermore, test publishers are increasingly marketing the tests as being able to identify a wide selection of “counterproductive behaviors,” not merely theft.

II. Reasons for Use

There are a number of reasons for the widespread use of honesty tests and their increased popularity. One researcher suggested that the tests were in part a way for employers to reduce in a simple manner the large size of the pool of applicants for a particular position. Another possible reason for the significant number of companies using honesty tests is the widely held belief about the high level of employee theft, which takes place in companies. Calculating the actual level of employee theft is impossible to do because of the difficulty of obtaining reliable data. Detecting employee theft can be very difficult. Therefore, any calculation of theft levels depends entirely on estimates based solely on theft that is detected. Another difficulty in calculating rates and levels of theft include the difficulty of deciding how broadly to define theft and which behaviors to include in the definition. One last problem, which relates to efforts to compare employee theft to other types of theft such as shoplifting, is the difference in consequences between employees and others and in the ability of calculating and recovering previous theft.

(Among the estimates on employee theft are the following: the U.S. Chamber of Commerce estimates that 30% of business failures result from employee theft, pilferage and fraud cost companies $120 billion in 2000. The Association of Certified Fraud Examiners estimated that the average organization loses approximately 6% of its annual revenue to fraud and abuse, costing over $40 billion annually in the United States; and a study by the National Retail Federation from 1993 reported that employee theft accounted for 38% of total “shrinkage.”)

The increasing popularity of honesty tests is also due in part to the unwillingness of many employers to provide candid references about former workers because of a fear of libel suits resulting from negative comments made about former workers. Additionally, a significant number of applicants for lower level positions may be quite young and may have little or no prior employment history from which to get references. Employers are therefore forced to look for alternative sources of information on prospective employees. Lastly, some reports argue that employers might be using honesty tests as a potential defense in case of negligent hiring suits (despite the fact that the tests have never been used as a defense in a court of law and there is no evidence that such an argument would carry weight as the tests are not intended to measure tendencies toward violence).

III. Validity of Honesty Tests

The validity of honesty tests is very difficult to calculate for a number of reasons. To begin with, calculating the test’s validity requires a clear definition of the term honesty and what the tests are intended to calculate. Adding to this complication is the fact that many psychologists believe that honesty is not really a permanent trait but is situational, composed of a variety of factors which can vary from one situation to another (such as who it is that is being deceived, the motive for the deception, the item of deception, the method of deception, and the consequences for all affected). Additionally, questions have been raised about the extent to which character can be considered a valid predictor of an individual’s behavior.

Another problem in measuring validity comes from the difficulty of uncovering and measuring the quantity of theft (or calculating the “level of honesty” by those who pass the test (plus those thieves who avoid detection may in fact pose a higher risk). This calculation is obviously even more difficult for those who fail the tests as they are unlikely to be given the jobs they are applying for. An additional problem is that honesty tests have usually been compared with polygraph tests. This is a problematic comparison because polygraph tests were banned in part because of questions about their own validity. Furthermore, self-selection of those taking part in experiments on the honesty tests means that testers are more likely to be straightforward and honest in their responses than someone who is applying for a job.

Lastly, and perhaps most problematic, because the scoring systems on the tests are proprietary and testing companies are immensely protective of these systems, experiments into the validity of the tests have almost all been carried out by researchers affiliated with or hired by the testing companies. The evidence that does exist suggests that pass scores are frequently chosen arbitrarily and are often purposefully high in order to try to catch as many potential thieves as possible, in spite of the likely consequence of disqualifying a larger number of honest individuals. The only other source of information on the tests is, therefore, anecdotal, most of which comes from the testing companies.

In 1990, both the federal Office of Technology Assessment (OTA) and the American Psychological Association (APA) published papers on honesty tests, which involved a review of existing research on the tests. The literature they used reported that the false-positive error rates (i.e. mislabeled honest people) for honesty tests ranged between an average of 73% and 97%, while false negatives were around 58%. (Other reports have attributed a much higher validity to the tests - one report estimated a false positive rate of 38% while Dr. Arnold of the Reid Testing Company, one of the nation’s two largest testing companies, claims that his company’s test correctly identifies test takers 80 to 85% of the time, with mistakes occurring in both directions). While the APA felt that honesty tests were still better (because they are more scientific) than other methods used in the hiring process (arguing for the test’s “utility...for selection in some job families”), the OTA strongly disagreed, arguing for the need for further investigations and suggesting the possibility that Congressional involvement might be required. One reason for this difference may be that the APA, unlike the OTA, seemed to be focusing entirely on the scientific validity of the test for preventing employee theft, without taking into account some of the social consequences for those labeled as dishonest, potentially permanently, and thereby disqualified from numerous jobs.

IV. Reliability (i.e. Repeatability)

There is not much information on the reliability of honesty tests, but what does exist suggests that test scores are somewhat constant. However, one possible reason for this is that the tests may have been taken in a study setting, rather than a hiring setting, so that there was no reason for candidates to be more cautious or less forthcoming in their replies.

V. Types of Questions

The main areas covered in some or all tests include: admission of illegal or disapproved activities, opinions regarding illegal or disapproved behavior, description of individual’s own thought patterns and personality, and reactions to hypothetical situations.

Overt Tests (tests that ask straightforward questions directly related to one’s own behavior and attitudes towards illegal or disapproved behavior):

1) What total value in merchandise and property have you taken without permission from employers? (Multiple choice)

2) I have frequently associated with fellow employees who admitted they were stealing merchandise from the company. (True of False)

3) I am not an honest person and might steal or cheat (T or F)

4) I might help friends steal from my company (T or F)

5) I return quarters I find on the street to the police station (T or F - trick question to identify dishonest answering)

6) How often do you tell the truth?

7) Do you think you are too honest to take something that is not yours?

8) How much do you dislike doing what someone tells you to do?

9) Do you feel guilty when you do something you should not do?

10) Do you think it is stealing to take small items home from work?

11) What percentage of people you know are so honest they wouldn’t steal at all?

12) How easy is it to get away with stealing?

13) Do you believe most employers take advantage of the people who work for them?

Covert Tests (tests which seek to elicit information about applicant’s attitudes on areas which are believed to correlate in some manner to honesty):

1) How many employees take small things from their employers from time to time? (Multiple choice - percentages)

2) My philosophy is best described as:
a. the meek shall inherit the earth
b. it’s a dog eat dog world
c. no one owes you anything

3) What should be done if an employee occasionally smokes marijuana on the job? (Multiple choice)

4) Does it bother you when people ask stupid questions?

5) Do you start up conversations in a waiting room?

6) Should a person who writes a check he knows will bounce be refused a job in which honesty is important?

7) Eating right is important to my health (T or F)

8) I like to take chances (T or F)

9) On average, how often during the week do you go to parties?

10) I feel lonely even when I am with other people (all, most, some time, almost never, never)

11) How often do you blush?

12) How often do you make your bed? (Every day, never, etc.)

13) I have nightmares every few nights.

14) It would be better if almost all laws were thrown away

15) My soul sometimes leaves my body

16) I see things or animals or people around me that others do not see

17) Sometimes I am strongly attracted by the personal articles of others such as shoes, gloves, etc., so that I want to handle or steal them though I have no use for them.

VI. Desired Answers

A number of reports pointed to the fact that honesty tests appear to have a high correlation with such traits as conventionality, traditionalism, conformity, social closeness, sensitivity, self-discipline, tidiness, and disapproval of dishonest conduct. Successful test takers also tend to demonstrate a lack of cynicism - they do not view theft as commonplace. They also tend not to feel very alienated and are not very impulsive, neither taking risks nor seeking thrills. (Additionally, one psychologist said that he watched for candidates who made excuses for their own dishonest behaviors, which he saw as a sign of the individual’s dishonesty.) As one report put it, what the tests seem to be looking for are people that “see no evil, commit no evil, and are properly punitive toward those who do.” (A number of reports pointed out that the tests discouraged a forgiving attitude towards misdeeds, occasionally mentioning the case of the Minnesota nun who scored particularly low on a test because of her unwillingness to punish others.) However, the tests also claim to have failsafes such as trick questions aimed to catch out those who present themselves as totally pure and perfect. Receiving a very high score would also cause testers to seriously doubt the honesty of the applicant’s answers.

VII. Fakability/Coachability of Tests

The significant similarities between the different tests hints at the extent of their susceptibility to both faking and coaching, although, according to the reports, there does appear to be a significant difference between the overt and covert tests on this front, as the questions on the latter are much more varied and the desired answer is not always as obvious.

VIII. Other Issues Raised by the Reports

• Some of the reports pointed out that a number of the employers using the tests do not understand how they work and are unaware of their fallibility. They believe honesty tests possess a scientific ability to accurately discern between those who are honest and those who are dishonest. This leads to a number of problems, including using the tests as the key criterion in employment decisions and putting too much faith into employees identified as honest by the test.

• This lack of understanding points to the fact that honesty tests are designed to be both carried out and interpreted by professional psychologists and are intended for use as part of a wider evaluation process. One report pointed out that the lack of this professional involvement could actually affect the validity of the test results through such seemingly minor factors as the applicants’ mood, which could be altered by such elements as the tester’s phrasing of verbal instructions, the tone of his instructions, and his relationship with the person being tested. However, this professional involvement is often missing - perhaps more often than not. One reason for this is that most companies are unlikely to have psychologists on their staff and so would be forced to hire someone new for the purpose. Furthermore, using a professional psychologist for the task would be probably be more expensive than using someone without these skills, whereas the tests are used by employers and marketed by testing companies as a way of cutting costs. Additionally, the use of professional psychologists might, ironically, be discouraged by ADA regulations as a way of preventing the tests being labeled as improper pre-offer medical exams. Lastly, the results are also designed in a simplistic way that does not require a psychologist but which also makes the tests seem much more cut and dry than they in fact are. As one psychology study put it “[Test] Publishers typically offer very basic reports (containing scale scores and risk levels) with more detailed reports (e.g., specific cut scores by industry, critical items) offered at additional cost...integrity tests are not accompanied by detailed narrative reports.” This comment also points to the fact that these tests often create universal standards for all types of jobs in what they are looking for, rather than tailoring the tests to take into account what characteristics might be desirable for the particular position, company, etc.

• Another consequences of employers using honesty tests without fully understanding their purpose is generating a false sense of security for companies about the honesty of their employees. This may lead them to feel that they do not need to spend money on other safeguards to protect against those who were mistakenly evaluated as honest by the tests. Additionally, the tests can also lead companies to reach an overly simplistic understanding of honesty, which does not take into account the possible situational aspects. This allows them to disregard the possible need to reexamine their own policies and change any aspects that might lead employees to be less concerned with the well-being of the company or which may even contribute to theft, by causing employees to feel resentment towards the company - some reports even suggest that forcing employees to take the tests could in fact contribute to this feeling of resentment.

• Directly connected to the issue of creating a false sense of security is the fact that these tests depend upon a naive reliance on the test-taker’s self-reporting of his own previous dishonest behavior, a situation which might lead to some individuals being penalized for being too honest or conscientious while other are rewarded for their dishonesty. This, in turn, points to the ironic dilemma that those who admit to their previous dishonesty may be more honest than those who do not. Furthermore, the reliance on self-reporting depends upon the assumption that past crimes are in fact a sign of future untrustworthiness, leaving no room for the possibility that an individual might have reformed themselves, a particularly questionable attitude if the individual is admitting to their previous misbehavior.

• Another issue raised by the reports is that the tests themselves could cause their own set of problems. To begin with, as mentioned earlier, the tests could cause resentment among applicants and employees by creating an atmosphere of distrust in the workplace and treating all prospective and present employees as if they are potential thieves. This resentment could erode the loyalty and morale of employees. Using the tests could therefore backfire, causing employees to steal more or behave in other, more passive-aggressive manners, such as putting less effort into their jobs or into paying attention to the company’s interests. Additionally, some researchers argue that being forced to take the tests might put off some more able or talented individuals from working for the company. Furthermore, the test’s high correlation with mainstream values could lead some talented or creative types from being disqualified.

• One serious concern about honesty tests is that certain honest individuals could be repeatedly mislabeled as dishonest by the tests and therefore become blacklisted and unable to get hired. This could occur first if the tests are reliable (repeatable results) as the test companies claim, so that a person would always be judged in the same way by the tests. Compounding this difficulty is the fact that most companies do not reveal to applicants that they have been rejected because of a poor honesty test score or even let them know their score, so that the individual does not even know he is blacklisted and therefore has no opportunity to try to change his position, either by changing his answers, appealing the test, arguing, or attempting to demonstrate the fallibility of the tests, a method which could lead to systematic changes in either the use or the legality of testing. Another problem comes from the fact that the testing industry is dominated by a couple of companies who, especially with the advent of internet testing, could in fact keep a database of “dishonest” applicants, so that the company would not even have to test applicants and, if they did, the testing company could identify those who had significantly changed their answers. Lastly, some companies might choose to share results on different applicants among one another as a way of saving money, a legal procedure. In fact, one study estimated that “sales of pre-employment data are growing as much as 75% per year for some information companies.”

• Some reports question whether the tests should in fact be viewed as discriminatory either under Title VII of the Civil Rights Act (which prohibits employment discrimination on factors such as race and gender) or the Americans with Disability Act. One issue raised by researchers questioning the tests’ legality under Title VII is the fact that most of the tests were tested to be normed on a white, middle-class population, which failed to take into account the possibility that different groups might have different standards or views on certain issues. In fact, there is evidence to suggest that certain groups tended to score poorly on some of the models upon which most of the tests were based. Although there have been some changes in recent years, tests still apply a single standard across all groups. Others question whether honesty tests should be considered illegal pre-offer medical exams under the ADA, in part due to concerns that the tests could possibly be used to detect an applicant’s hidden psychological disabilities, the original intention of many of these tests. As mentioned earlier, some companies have tried to get around this issue by choosing not to involve professional medical personnel, a factor which would make the test’s legality more tenuous. However, by avoiding being considered medical exams, honesty tests can be required prior to the offer stage and there is less of a necessity for questions to be job-related and consistent with business necessity.

• The possibilities that the tests might be discriminatory or that certain individuals might be blacklisted are particularly worrying in light of the fact that the tests are often targeted towards lower-level, unskilled, or semi-skilled workers so that they risk exacerbating already existing social inequalities by failing to give these individuals an opportunity to find employment and improve their lot. Furthermore, the target group’s poverty level also places them among the most politically powerless and most socially marginally, making them among the least able to defend themselves and stand up for their rights by refusing to take the tests, demanding greater political scrutiny of the tests, or trying to prove the tests’ discriminatory effects.

• Some researchers are concerned that the tests may in fact constitute an invasion of the applicant’s privacy as a number of the tests touch upon personal issues. (They also raise the possibility that individuals could unknowingly incriminate themselves.) The only investigation of an honesty test upon these grounds took place in Soroka v. Dayton Hudson Corporation (1993), in which the California State Appellate Court found that certain questions asked by the honesty test in question violated the defendants’ right to privacy. The court also ruled that employers could only ask questions which violated applicants’ privacy if the questions were related to the position for which the individuals were applying. The first limitation with this finding is that it merely required the testing companies to remove the offending questions. The other problem is that the right to privacy does not exist on the federal level outside of public settings, a category which does not cover employers, even public employers. One last privacy-related concern relates to the safety of test scores and answers either within the testing company or once they are given to the employer.

• The last major concern about honesty tests is that they are almost completely unregulated (unless one counts the minimal voluntary self-regulation that came about out of fear that the tests might be put under governmental scrutiny). The only exceptions are Massachusetts which has banned the tests and Rhode Island which requires that employers not use the tests as the sole criterion for their employment decisions. While other states have tried to pass regulation, none have so far succeeded. This lack of regulation causes a number of problems. First, because the companies are immensely protective of their product, the tests cannot be suitably and fully tested in an unbiased atmosphere on concerns about their validity or for any discriminatory effects. It also allows employers to give the tests in unsatisfactory conditions and allows the results to be interpreted by untrained personnel, who may not understand the concerns about the tests’ validity and are not aware of the role tests are intended to play in the hiring process. The situation has also allowed for the creation of new tests of dubious validity and allows companies to engage in false or exaggerated advertising relying solely on anecdotal evidence.

• Finally, one researcher pointed out that governmental inaction in regulating honesty tests, which is based in part on the lack of significant research about honesty tests, effectively results in the government presumptively siding with test supporters. This places the burden of proof on job applicants, the party least able to gather or publicize information about the tests. Such a system also removes any incentive for test publishers to allow the tests to be examined independently and in greater detail.

IX. Recommendations Offered by Various Articles and Reports about Honesty Tests

• Screening employers prior to selling tests and limiting test use to those company workers either deemed competent by companies or willing to undergo training.

• Requiring test publishers to provide employers with notices that make them aware of the fact that the tests lack independent verification of validity and which warn them that the tests may be over inclusive and may misclassify people as dishonest (and honest). This would allow employers to make more informed decisions in choosing to use the tests, possibly leading some companies to stop using the tests and thereby perhaps eliminating honesty testing without direct government prohibition. Furthermore, requiring these notices might encourage many test publishers to seek independent verification of their product so as not to have to include the notices with their tests.

• Requiring employers to give applicants notices and/or informed consent forms which describe the purpose of the test, the employer’s reasons for using it, and an announcement that the test taker will not be given results. Alternatively, requiring employers to give applicants access to the results (and an explanation of their part in the process).

• Having Congress declare a moratorium on the use of honesty tests until they receive independent verification.

• Creating a regulatory agency.

• Creating protections and incentives that would give prior employers an incentive to be more forthcoming about past employees.

• Researching the correlation between certain traits and job performance in particular fields.

• Requiring that tests be job-related, i.e. demanding that test publishers explain the purpose of their product, which behaviors it aims to predict, what areas it is applicable to, its limitations, and what proof exists of the test’s validity.

• Requiring that tests not be the sole criterion in an employer’s hiring decision.

• Demanding that test publishers stop using such universal standards in their scoring system. Getting rid of the pass/fail system, explaining the results, adapting tests/scoring systems to make them more relevant and applicable to different industries.

• Mandating strict confidentiality of test results.