Machine Learning and Prediction in Psychological Assessment

Abstract. Modern prediction methods from machine learning (ML) and artificial intelligence (AI) are becoming increasingly popular, also in the field of psychological assessment. These provide unprecedented flexibility for modeling large numbers predictor variables non-linear associations between predictors responses. In this paper, we aim to look at what these may contribute assessment criterion validity their possible drawbacks. We apply a range modern statistical dataset predicting university major completed, based on subscales items scale vocational preferences. The results indicate that logistic regression combined with regularization performs strikingly well already terms predictive accuracy. More sophisticated techniques incorporating non-linearities can further accuracy validity, but often marginally.