MARCH 26, 1990 STATEMENT ON PENCIL AND PAPER HONESTY TESTS

One of the disappointing employer responses to the Polygraph Protection Act has been an increase in the use of paper and pencil honesty tests.

Approximately 15% of all companies currently use honesty tests. These 5000 companies administer about 2.5 million tests annually. They are generally used as a pre-employment screen by retail firms concerned about employee theft.

Testing industry spokespeople report a dramatic increase in business since the Polygraph Protection Act took effect, and this is likely to become a major issue for the ACLU in the near future.

This is a preliminary report on the state of the art and the law. We will update you as our research progresses.

1. Technical Issues

   Honesty tests rely upon three different types of indicators to predict future dishonesty.
   
   

   1. Admissions of previous dishonest acts.
   2. Opinions about dishonest behavior.
   3. Other personal characteristics which are believed to be correlated with dishonesty (such as alienation).

   At least 35 studies have been conducted on the validity of paper and pencil honesty tests. While most of these studies were conducted by testing companies, independent reviews indicate that they were done correctly.

   The results of these studies are reasonably consistent, and tend to show:
   
   

   1. Written tests can consistently determine to which of two groups vastly different in honesty (for example convicted felons and college students) a test subject belongs.
      
      
   2. Tests can identify a group of job applicants which includes most (about 80%) of the thieves.
      
      
   3. To accomplish this, the cutoff point must be set so low that many honest people are also disqualified.

      In a typical application, assume a company has 100 job applicants, 25 of whom are dishonest enough to steal from the company. Testing can screen out 20 of these thieves, but only by rejecting 50 candidates (including 30 honest applicants--a false positive rate of 67%).

   Examples of people who have been victimized by these tests are hard to obtain, because employers are advised by testing companies not to tell applicants who fail an honesty test the reason for their rejection.

   The only instance we are aware of involves Sister Terressa, a nun from Minneapolis, who was rejected for a part time job with B. Dalton Bookstores because she failed the written honesty test (the Reid Report). She was told that her score was the worst they had ever seen.

   Of particular concern is the indication that false positives are not random. People who are abnormally honest tend to fail honesty tests because they admit to dishonesty for acts which the average person (and the test) consider too trivial to count. For example, someone who took a legal pad home in their briefcase might answer "yes" to a question asking if they had ever stolen from their employer. People who are realistic/jaded about the degree of honesty to be expected of others (frequently from extensive supervisory experience) or unusually forgiving of human imperfection frequently fail because they are viewed as condoning dishonesty. Honest individuals who fail for these reasons might well be unemployable if honesty tests were universal.

   While we would reject such tests because of the high percentage of innocent people screened out, an employer might well consider it valid because most of the potential thieves can be eliminated. This is especially true where there are many more applicants than job openings.

   Both the Congressional Office of Technology Assessment and the American Psychological Association are conducting studies of paper and pencil tests which are due out this spring. Even our inside sources are unable to predict what these studies will conclude. We will advise you of the results when the studies are released.
   
   

2. Legal Issues
   
   
   1. Federal Statutes

      The Polygraph Protection Act does not cover paper and pencil tests. This is clear not only from the statutory language, but also from the legislative history.
      
      

   2. State Statutes
      
      

      At least 15 states have lie detector statutes. These fall into 4 general classes.
      
      

      1. Statutes which clearly restrict paper and pencil tests.

         Massachusetts - Bans the use of written honesty tests.
         Rhode Island - Prohibits employers from using written honesty tests as the primary basis for employment decisions.
         
         

      2. Statutes that restrict the use of "lie detectors" generally.

         Alaska
         New Jersey

         A respectable argument can be made that paper and pencil tests are covered by these statutes. Unfortunately, neither statute contains a definition of "lie detector." The common tendency to use "lie detector" as a synonym for "polygraph" is an obvious counter argument. A challenge under these statutes would probably turn on the legislative history.
         
         

      3. Statutes which ban polygraphs and "other similar tests".

         California
         Delaware
         Idaho
         Maryland
         Washington
         Wisconsin

         The argument here is essentially the same as the second category, but weakened by the requirement that the test be "similar to" a polygraph. None of these statutes define similarity.

         The only reported case interpreting this type of statute is State v. Century Camera, Inc.1, in which the Minnesota Supreme Court ruled that it applied only to tests that measure physiological changes, not written instruments.
         
         

      4. Statutes which prohibit the use of lie detectors involving "instrumentation or mechanical methods".

         Iowa
         Michigan
         Pennsylvania
         Vermont

         Since many tests are scored by computers, it could be argued that they meet this definition. This is a long shot at best.
         
         

   3. State Constitutional Privacy Provisions

      10 States have constitutional provisions protecting individual privacy:

      Alaska	Florida	Louisiana	South Carolina
      Arizona	Hawaii	Montana	Washington
      California	Illinois	Rhode Island*	Wisconsin*

      *General privacy statute instead of constitutional clause. Litigation has recently been started in California claiming that invasive questionnaires violate this clause. While this will have more impact on personality tests than honesty tests because of their greater use of invasive questions2, the outcome will still be important. We will keep you advised.
      
      

   4. Tortious Invasion of Privacy

      There is good case law in most states that the tort of invasion of privacy is applicable to employers.

      We are currently aware of case law holding that invasive questions by public employers are actionable in tort3. Further research will be required to determine to what extent this is true in the private sector (all we have now is some interesting dicta).4

      Also requiring further research is the relationship between this tort law and the doctrine of employment at will.

      We will send you an updated memo when this segment of the legal research is complete.
      
      

   5. Discrimination

      Many people have raised the concern that paper and pencil tests are racially or sexually discriminatory. Almost every test publisher, however, has data showing virtually identical passing rates by race, and slightly higher passing rates for women than for men.

      Because of this, numerous discrimination charges fired against such tests with the EEOC have been found to be without probable cause.

      We will keep you advised as this issue develops and our legal research progresses.

      If you would like more detail, backup material, etc. on what we already know, please let me know.

ENDNOTES

1. 309 N.W. 2d 735 (Minn. 1981)
2. Written tests which purport to measure other personality traits raise similar, but not identical issues, and will be covered in a subsequent report.
3. Thorne v. City of El Segundo, 726 F. 2d 459 (1983) (and other cases)
4. See Cort v. Bristol - Myers, 431 N.E. 2d 908 (1982)