MARCH 16, 1991 STATEMENT ON THE AMERICAN PSYCHOLOGICAL ASSOCIATION REPORT ON HONESTY TESTS

The American Psychological Association (APA) recently released their report on written honesty tests. This report is much more positive than the earlier report from the Congressional Office of Technology Assessment, and is being touted in the press as a vindication of honesty testing.

This, however, is not the case:

1. The report does not conclude that honest testing - as actually practiced -- is valid.

   It concludes that honesty testing is as accurate as interviews and reference checks -- as long as "cutting scores" are not used. (Honesty tests produce a continuous range of scores. A cutting score is an arbitrary point on this range. All people scoring above this arbitrary point are said to be honest; all those scoring below are considered dishonest. This unnecessary practice is what makes honesty tests a pass/fail exercise.)

   In reality, however, virtually all honesty tests employ cutting scores.
   
   

2. APA accepts as valid any test that has a correlation coefficient of .30 or higher.

   This is not a very demanding standard. A test with a coefficient of .30 is wrong far more often than it is right.

   While such a low correlation may be adequate for academic analysis, it is not good enough to be used when making real world decisions about peoples' jobs.
   
   

3. It is questionable that honesty tests meet even this low standard.

   A true measure of honesty testing's accuracy can only be obtained by "predictive validity" studies in which test scores are compared with subsequent dishonest behavior. APA was willing to judge honesty tests by how well their scores compared to admissions of dishonesty, results of other tests, and even polygraphs.
   
   

4. APA is not an unbiased body. A significant number (but by no means a majority) of APA members make their living designing or administering honesty tests.

   It is interesting to note that OTA, which has no financial interest in the subject, reached a much more negative conclusion.

   I hope this will be useful if you are called upon to comment on the APA study. If you need more detail, don't hesitate to call.