Employee Screening and Assessment 117 a combination of a cognitive ability test, a personality inventory, and a biodata
inventory predicted worker performance better than any one predictor alone (Konradt, Hertel, & Joder, 2003).
We have seen that standardized tests can be reliable and valid screening devices for many jobs. However, two important issues regarding this use of tests must be considered: validity generalization and test utility. The validity
generalization of a screening test refers to its validity in predicting performance
in a job or setting different from the one in which the test was validated. For example, a standardized test of managerial potential is found to be valid in selecting successful managers in a manufacturing industry. If the test is also helpful in choosing managers in a service organization, its validity has generalized from one organization to another. Similarly, validity generalization would exist if a test of clerical abilities is successful in selecting applicants for both secretarial and receptionist positions. Of course, the more similar the jobs and organizations involved in the validity studies are to the jobs and organizations that subsequently use the screening tests, the more likely it is that validity will generalize from one situation to another.
High validity generalization of a standardized test will greatly increase its usefulness—and reduce the workload of I/O psychologists—because the instrument may not need to be validated for use with each and every position and organization. Some I/O psychologists, such as Schmidt and his colleagues, argued that the validity generalization of most standardized employee screening procedures is quite high, which means that they can be used successfully in a variety of employment settings and job classifications (Schmidt & Hunter, 1977, 1981; Schmidt, Hunter, Outerbridge, & Trattner, 1986; Schmidt et al., 1993). At the other extreme is the view that the ability of tests to predict future job success is situation-specific, and validity should be established for each use of a screening instrument. Although few I/O psychologists believe that the validity of test instruments is completely situation-specific, there is some disagreement over how well their validity generalizes.
From an international perspective, some types of tests may generalize better across countries and cultures. For example, tests of cognitive abilities should be important for many jobs throughout the world, and evidence suggests they are less prone to cultural effects (Salgado, Anderson, Moscoso, Bertua, & de Fruyt, 2003), whereas personality tests, for example, may be more susceptible to cultural effects.
Test utility is the value of a screening test in helping to affect important
organizational outcomes. In other words, test utility determines the success of a test in terms of dollars gained by the company through the increased performance and productivity of workers selected based on test scores. For example, in one organization a valid screening test was used to select applicants for 600 jobs as computer programmers (Schmidt, Hunter, McKenzie, & Muldrow, 1979). The estimated money gained in one year from the increased speed and efficiency of the chosen workers was more than $97 million. The initial cost of the screening tests was only $10 per applicant, a very good return on investment.
All in all, utility analyses of standardized employee testing programs indicate that such tests are usually cost effective. Hunter and Schmidt (1982)
validity generalization the ability of a screen- ing instrument to pre- dict performance in a job or setting different from the one in which the test was validated
Stop & Review
Define five categories of employment tests.
test utility
the value of a screening test in determining important outcomes, such as dollars gained by the company through its use
went so far as to estimate that the U.S. gross national product would be increased by tens of billions of dollars per year if improved employee screening and selection procedures, including screening tests, were routinely imple- mented. Utility analyses allow the employer to determine the financial gains of a testing program and then compare them to the costs of developing and implementing the program.
Another important issue in testing is the importance of ethics in the administration and use of employment testing, including the protection of the privacy of persons being tested. I/O psychologists are very concerned about ethical issues in testing. In fact, the Society for Industrial and Organizational Psychology (SIOP) published a fourth edition of its Principles for the Validation and Use of Personnel Selection Procedures (SIOP, 2003). This publication out- lines important ethical concerns for employment testing.
A final issue concerning testing is the issue of faking. Faking is trying to “beat” the test by distorting responses to the test in an effort to present oneself in a positive, socially desirable way. Faking is a particular concern for personal- ity and integrity tests (Ryan & Sackett, 1987). Laypersons tend to believe that employment tests are easily faked, but this is not the case: First, many tests have subscales designed to determine if a test-taker is trying to fake the test. Second, it is often difficult for the test-taker to determine exactly which responses are the correct (desired) responses. Finally, there is evidence that personality and integrity tests are quite robust, still validly measuring their intended constructs even when test-takers are trying to fake (Furnham, 1997; Hough, 1998; Ones & Viswesvaren, 1998c).