Various issues surrounding the use of multiple regression analysis in psychological research

Abstract

Although multiple regression analysis is a frequently used method for multivariate analysis in psychological research, it has been used inappropriately or incorrectly in most studies. To resolve these problems effectively, we investigated and summarized the issues related to the use of multiple regression analysis found in papers published in The Japanese Journal of Psychology and discussed the issues in detail. We argue that researchers should not use multiple regression analysis for simplistic reasons, such as “because there are several independent variables” or “because some relationships between independent variables or between independent and control variables are supposed.” We further argue the importance of carefully considering whether the purpose of the study is to explain or to predict and what kind of causal relationships exist between variables.