Estimation of Causal Effects through Longitudinal Study: An Example of Internet Use and Aggression

Abstract

In this article, we reviewed how a longitudinal study could be used to estimate causal effects in the field of personality psychology. First, characteristics and limitations of experiment, cross-sectional survey, and longitudinal survey were illustrated. Then, focusing on panel study, which is a form of longitudinal study that is able to examine reciprocal causal relationships, we described the method for analyzing causal relationships using cross-lagged effect model. As an example of panel study that was conducted to estimate causal effects, a study by the current authors was discussed, which investigated the relationship between Internet use and aggression. In the study, reciprocal causal effects were examined: the effect of Internet use on aggression, and that of aggression on Internet use. At the conclusion of the article, advices for researchers who would attempt to estimate causal effects with high accuracy were summarized.