Which Test Should we Use?

by Jeff Schroer - HRScience Copyright 2004

Psychometric assessment products exist in great variety and quantity. Many businesses have employed assessments to support their recruitment, retention, development, and succession planning strategies. However, few are aware of the U.S. Department of Labor's position regarding the use of tests and assessments in areas of employment.

It may surprise some to learn the USDOL actually supports the use of a sound testing and assessment strategy. In the white paper, "Testing and Assessment: An Employer's Guide to Good Practices," the USDOL recognizes that, "employers face the challenge of attracting, developing and retaining the best employees." The study goes on to say that a solid assessment strategy can "maximize chances for getting the right fit between jobs and employees."

They provide 13 basic principles that employers should follow when considering and/or designing an assessment initiative. The principles are summarized here. (For a complete copy of the 80-page document (.pdf) , go to www.davekahle.com/dfiles/empTestAsse.pdf).

1. Use assessment tools in a purposeful manner. They are most beneficial when used properly and for the purpose for which they are designed. Misuse or improper use could be harmful or possibly, illegal.
   
   
2. Use the whole-person approach to testing. No test is perfect. Complex behaviors are at work. Use a test, or combination of tests, that give as much information as possible about the behaviors that are most important to your organization.
   
   
3. Use tests that are unbiased and fair to all groups. Tests that deliberately or inadvertently discriminate prevents the employer from achieving a qualified and diverse work group.
   
   
4. Use tests that are reliable. Will the same person get the same results each time they take the test? Reliability ("r") is expressed as a statistical coefficient between 0.0 and 1.0. r = 0.90 or above is "excellent": 0.80 - 0.89 is "good": 0.70 - 0.79 is "adequate".
   
   
5. Tests must be valid for the purpose they are being used. Validity is the most important criteria for selection of a proper test instrument. Validity is expressed as a statistical coefficient. A v-score of .35 or higher means the test is "very beneficial" in determining the presence of desired characteristics. 0.21 - 0.34 means the test "is likely to be useful" to the employer. A v-score of 0.11-0.20 means the usefulness of the information derived will "depend on the circumstances" under which the test is being used. A test is "unlikely to be useful" under any circumstances when v= 0.11 or less.
   
   
6. Tests must be appropriate for the target population. Tests designed to assess nurse practitioners is likely to be inappropriate when applied to the construction trades.
   
   
7. The instructions and other documentation must be comprehensive and easy to understand. The results needed are directly related to the ability of the person being tested to understand the instructions for taking the test and the test questions themselves.
   
   
8. If the test requires proctoring and/or administration, those performing this function must be properly trained. Some instruments require an extensive certification process to administer, proctor or score tests.
   
   
9. It may be necessary to provide consistent and uniform testing conditions to obtain consistent results. Classrooms, conference rooms or other facilities may be necessary to isolate test takers from other distractions to assure the integrity of test results.
   
   
10. Provide reasonable accommodations for people with disabilities. As stated earlier, no group should be disadvantaged by the test or the conditions under which the tests are taken.
    
    
11. Test security must be maintained if the results are to be useful. Tests should never be accessible to the general population.
    
    
12. Test results must be maintained in a confidential manner. Tests taken over the Internet or other computer based methods that require usernames and passwords are often most effective at preserving the confidentiality of the results.
    
    
13. Accurate interpretation of results is necessary. It does little good to interpret good data poorly. Ensure that all test reports are, like the documentation, easy to understand.