Predictive Analytics for Efficient Decision Making in Personnel Selection

This study investigates the predictive modelling in personnel selection. In particular, we focus on the prediction of interview performance using combinations of variables which assess personality and cognitive ability. Based on a dataset of 1,989 subjects, we generate 1,024 possible models with ten predictors including six personalities and four cognitive factors and apply the mixed-effect logistic regression to account for the random effect. The predictive performance of each model is evaluated by the area under receiver operating characteristic curve. The results show that the model with a combination of ambition and agreeableness as well as verbal and reasoning can predict the interview performance at 68% accuracy and this predictive power is not substantially different from the predictive performance of more complicated models. Our results suggest that personnel selection with fewer factors can be as efficient as all factors in the prediction. This study contributes to the selection literature by emphasising and justifying efficient decision making with predictive models, and it demonstrates that the personnel selection procedure can be simplified in an organisation and can save the organisation resources.