The Three Biggest Mistakes in Testing for Business Purposes

It is clear from validation studies and ROI analyses that properly designed and administered selection testing systems can help companies get the right people in the right jobs, avoid hiring mistakes and help reduce turnover from selection errors. As you minimize hiring mistakes, you make more money. The numbers are clear and unambiguous. But it takes more than just a couple of brainstorming sessions with an "I know what it takes to be a good sales person!" mindset. Although psychology often suffers from being a soft science, proper test construction and validation relies on statistically rigorous processes and sophisticated quantitative analyses to help make predictions about behavior. When coupled with a good system of selection interviewing and background checks, well-validated and customized test batteries can add significant accuracy to your predictions about success on the job.

Of course, there are horror stories about companies that have gotten into legal trouble by the misuse of tests. But when done right, testing systems are quite defensible and will hold up to legal and ethical scrutiny. However, you need to do your homework. Using an inappropriate test and/or using the results of tests in an inappropriate manner will get you into trouble. Testing systems need to be developed with an eye towards fairness and current legal guidelines, and they need to be carefully validated to make sure you do not run into unanticipated difficulties.

Be careful about the three critical mistakes in the use of selection tests.

Although there are potentially costly risks in the use of pre-employment tests, they continue to be one of the most popular and effective tools to help managers make good hiring decisions. If you currently use them, or are considering doing so, here are three things you really need to know to help you avoid legal trouble and to help you hire better people.

ONE: Using a Bad Test

By "bad test," we mean one that has not been shown as valid and reliable. Validity refers to how well it does what it is supposed to do. Reliability refers to how consistently it measures whatever it measures, and how consistently it performs over time. A bad test may also be missing evidence that shows it's fair to different groups of test takers. There are specific tasks a test publisher should accomplish before making a test available. The statistical procedures are clear and well-known. Once these steps have been performed, they should be clearly explained in a technical manual for the user. If you cannot get such documentation from a publisher, do not use the test to help make decisions on people.

TWO: Using a Good Test, but One That is Not Related to the Job

By "good test," we mean one that is valid and job-related. If you are using a highly valid test of math skills, but using it to help select people for a job that does not require math, you are looking for trouble. However, if you have demonstrated that strong verbal skills are related to success as a writer, the use of a vocabulary test will help you select better employees for this position. Be sure you are choosing the right tests for the right positions.

If you do not have enough people in a particular job to perform a proper job analysis and relate the key job behaviors to test results, make sure the test meets the validity requirements outlined above. If you do have a sufficient number of people in the job under consideration, use the general guidelines for validation presented in the section of validation later in the course.

THREE: Using the Test the Wrong Way

Even if you have the most valid and reliable test in the world, you should not rely exclusively on it. Structured behavioral interviews, reference checks, work samples and general background checks should also inform your decision.

You should be consistent in your use of tests, interviews and all other components of your selection process. Always be sure to accommodate special needs as outlined in the ADA (Americans with Disabilities Act) and stay aware of current EEOC guidelines. You should periodically review the cutoff score used for the test to make sure it is set at the appropriate level and that you have no adverse impact. Your selection system should not discriminate on anything other than job performance.

Of course, there are a wide range of options for selection/assessment tests, but you should do your homework to make sure the instrument or instruments you select have been scientifically developed and validated for use in business settings. The eTest Personality Inventory was developed specifically for use for business applications, not adapted or re-purposed from clinical roots. It has been shown to help reduce turnover and to predict performance in a wide range of settings. But no matter what assessment instruments you use, you are on much more solid legal ground by customizing and validating the tests for your own environment. See the section on validation for more information on this topic.

Quiz Questions: