Psychological research has mostly been focused on finding an explanation for behavior, rather than on finding a model that accurately predicts behavior. This approach often results in the use of...Show morePsychological research has mostly been focused on finding an explanation for behavior, rather than on finding a model that accurately predicts behavior. This approach often results in the use of models that fit very well to the sample used for testing but are difficult to generalize to new samples. Similarly, models are often too complex and take into consideration too many variables. Classically, replications would be performed to account for these issues. This is however an expensive, time-consuming, and laborious process, leading to the perpetuation of nonreplicated studies in psychology. This thesis presents a possible solution stemming from the field of machine learning, namely using cross-validation. With cross-validation, the predictive performance of a model can be assessed using only one dataset. This thesis examined the use of cross-validation by applying it over an existing dataset and comparing its output to the output of conventional null-hypothesis testing. The results show that using cross-validation reduces the likelihood of making overly optimistic claims, by reducing the chances of using excessively complex models unable to generalize to new samples. Furthermore, cross-validation gives the opportunity to examine the predictability of models while preserving the explanatory power. It therefore proves to be a useful tool in the field of psychological research.Show less
