ABSTRACT
In the 1960s and 1970s individuals in the field of Industrial and Organizational Psychology grappled with a major dilemma. It was unclear whether the results of a local validation study, used to support selection decisions, were specific to the particular situation (e.g., job and organization) or whether they could be generalized to other types of jobs and organizations. Ghiselli (1966, 1970) concluded that the validity coefficients varied considerably across settings and therefore demonstrated substantial “situational specificity.” 1 If his conclusion was correct and validity results were not generalizable, then it would be necessary to engage in the highly undesirable process of conducting a new validation study every time a selection test was used in a new context. To determine whether validities were situationally specific, a new methodology was needed that could integrate validity results from numerous studies and estimate the average validity and the variance of the validities while accounting for study artifacts known to affect these quantities such as sampling error, measurement error, and range restriction.
