How effective are tests and Interviews as selection devices?


The proven relationship between a selection device and some relevant criterion>>>

Any selection device that a manager uses such as application forms, tests, interviews, or physical examinations must also demonstrate validity. Validity is base on a proven relationship between the selection device used and some relevant measure. For example, a few pages ago, we mentioned a fire fighter applicant who was wheel chair bound. Because of the physical requirement of a fire fighter’s job, someone confined to a wheelchair would be unable to pass the physical endurance tests. In that case, denying employment could be considered valid, but requiring the same physical endurance tests for the dispatching job would not be job related. Thus, the law prohibits managements from using any selection device that cannot be shown to be directly related to successful job performance. That constraint goes for entrance tests, too; management must be able to demonstrate that, once on the job, individuals with high scores on this test outperform individuals with low scores. Consequently, the burden is on management to verify that any selection device it uses to differentiate applicants related to job performance.

Managers can use a number of selection devices to reduce accept and reject errors. The best known devices include written and performance simulation tests and interviews. Let’s briefly review these devices, giving particular attention to the validity of each in predicting job performance.

How do written tests serve a useful purpose? Typical written tests include tests of intelligence, aptitude, ability, and interest. Such tests have long been used as selection devices, although their popularity has run in cycles. Written tests were widely used for 20 years after World War II, but beginning in the late 1960s, they fell into disfavor. They were frequently characterized as discriminatory and many organizations could not validate that their written tests were job related. Since the late 1980s, written tests have made a comeback; many of them are Internet based. Managers are increasingly aware that poor hiring decisions are costly and that properly designed tests could reduce the likelihood of making such decisions. In addition, the cost of developing and validating a set of written tests for a specific job has come down markedly.

A review of the evidence finds that test of intellectual ability, spatial and mechanical ability, perceptual accuracy, and motor ability are moderately valid predictors for many semi skilled and unskilled operative jobs in industrial organization. However, an enduring criticism of written tests is that intelligence and other tested characteristics can be somewhat removed from the actual performance of the job itself. For example, a high score on an intelligence test is not necessarily a good indicator that the applicant will perform well as a computer programmer. This criticism has led to an increased use of performance simulation tests.

What are Performance simulation tests?

Selection devices that are based on actual job behaviors; work sampling and assessment centers

What better way is there to find out whether an applicant for a technical writing position at Microsoft can write technical manuals than to ask him or her to do it? The logic of this question has led to the increasing interest in performance simulation tests. Undoubtedly, the enthusiasm for these tests lies in the fact that they are based on job analysis data and, therefore should more easily meet the requirement of job relatedness than do written tests. Performance simulation tests are made up of actual job behaviors rather than substitutes. The best known performance simulation tests are work sampling (a miniature replica of the job) and assessment centers (simulating real problems one may face on the job). The former is suited to persons applying for routine jobs, the latter to managerial personnel.

The advantage of performance simulation over traditional testing methods should be obvious. Because content is essentially identical to job content, performance simulation should be a better predictor of short term job performance and should minimize potential employment discrimination allegations. Additionally, because of the nature of their content and the methods used to determine content, well constructed performance simulation tests are valid predictors.