APRIL 16, 1990 STATEMENT

The American Civil Liberties Union is alarmed by the large and growing use of written honesty tests in the employment setting.

Many of these tests contain invasive questions which violate people's right to privacy. We have seen tests with questions about sexual practices, religious beliefs, bathroom habits, and intimate feelings between family members. While these questions are more common on general personality tests, there appears to be some problem in this regard on honesty tests as well.

Employers have a legitimate interest in information concerning a potential employee's abilities and prior job performance. Many of the invasive questions currently asked, however, have no connection to job performance.

The ACLU believes that people are entitled to be judged on their work. They should not be forced to disclose intimate details of their lives to get a job.

We are also concerned about the accuracy of honesty tests. Our research indicates that most tests reject 50 percent of an applicant pool. While this may include most of the potential thieves, it also includes many honest people. If one assumes that one in five people is seriously dishonest, and that three quarters of those who are dishonest would be detected, a typical honesty test would screen out 15 potential thieves and 35 honest people.

We recognize that employers need to make decisions about who to hire and that no selection system is perfect. However, a test which penalizes two innocent people for every guilty person it detects is far too arbitrary to be used when something as important as a job is at stake. We would all be shocked if employers were to make hiring decisions by flipping a coin. For the honest person honesty tests are equally arbitrary. A coin flip would reject 40 of the 80 honest applicants. An honesty test would reject 35.

These false accusations of dishonesty may not be random. Many tests purport to detect not only dishonest behavior, but those with unacceptable attitudes about honesty. Anecdotal evidence suggests that those of us who are very forgiving (such as the famous Sister Terressa) or merely realistic in their expectations of others may fail honesty tests. It is interesting to note in this regard that some tests use different scales for managers and non-managers.

If this is true, and false positives are not random, then millions of innocent people, including the most honest among us, would become permanently unemployable if honesty tests were universal.

It would be very valuable to identify several experimental groups of clearly honest people with characteristics which the anecdotal evidence suggests may put them at risk (scrupulously honest, very forgiving, or extremely realistic) and have them take a series of different honesty tests to see if a pattern emerges.

The American Civil Liberties Union believes that the use of written honesty tests by employers is appropriate only where:

1. The test does not include questions concerning private matters which many people would prefer not to disclose.
2. The number of honest people rejected is substantially lower than the number of dishonest people detected.