Chapter 8: Range Restriction in Employment Interviews: An Influence Too Big to Ignore

The emergence of meta-analysis as a formal research technique in the early 1980s raised awareness of the need to consider the influence of various statistical “artifacts” in research (e.g., Hunter, Schmidt, & Jackson, 1982). Sampling error, for instance, artificially increases variability across coefficients, which could result in the conclusion that validity is highly specific to individual selection situations and not generalizable. Measurement error (particularly in performance criteria) and restriction in range (hereafter range restriction) reduce the magnitude of validity coefficients, thereby making selection approaches appear less effective than they really are in predicting job performance. Construct validity can also be affected by artifacts. For example, range restriction can artificially reduce correlations in a multitrait-multimethod (MTMM) analysis, lowering confidence that similar measures are assessing a common construct.