In many cases, certain tests are scored and interpreted by outside professionals. The results are then provided to the manager for consideration of whom to hire or promote.
There may be bias in how the test measures the trait it purports to measure. If test scores indicate that males perform better in verbal reasoning tasks than do females, when in fact they both perform equally well, the test is biased. Second, “If the test used in college admissions systematically over predicts the performance of males and underpredicts the performance of females, [then] that test functions as a biased predictor.”
Utility analysis simply answers the question, Does it pay to use a given test in a particular situation? In other words, if you use a specific test, will the quality of individuals selected be better than if you had not used the test?
Validity generalization helps determine if a test is valid in one situation, will it be so in others. Or, will the test need to be re-validated? Especially for smaller employers, properly generalizing the validity of a test can be a business lifesaver.