Double cross-validation of a weighted application blank over time

Abstract

Reports a 2nd cross-validation of a weighted application blank for predicting tenure of clerical employees. The sample consisted of 199 female employees hired during a later time period than the initial and cross-validation samples. Application blanks were scored using the weights previously developed. A substantial loss in predictive power was found in the additional cross-validation. Changes in labor market conditions, manpower needs, and specific changes in personnel policies appear to be responsible for the loss in efficiency. Results suggest cross-validation of such blanks with samples of employees hired during a later time period and periodic checks on their continuing validity. (PsycInfo Database Record (c) 2025 APA, all rights reserved)