Accounting for the apparently limited validity of preemployment drug and alcohol tests: States versus traits as predictors of future performance and suitability

Preemployment drug and alcohol tests can provide valid information about current impairment and/or recent use of specific substances, but their relevance for making decisions about job applicants may be limited. We show how the testing method, the narrow window of time in which specific drugs or their metabolytes can be detected, and the circumstances of testing combine to substantially affect the inferences that can be drawn about applicants who pass or fail these tests. We suggest a state-trait framework for analyzing preemployment drug and alcohol tests, and note that the very weak and indirect links between the states being measured (e.g., recent drug use) and the trait inferences being drawn from the test (e.g., long-term suitability for employment) may limit the validity of these tests.