In many disciplines of social sciences, it’s widely believed that main effect is uninterpretable when moderation exists. This research starts with an observation that four popular mode of data analysis, mediation, moderation, curvilinearity, and confounding, all attempt to dissect an X-Y simple linear effect into various components. 

  Through this prism of theory, this research uncovers an important interpretation for the main effect, with or without moderation. The main effect, according to the theory of causal dissection, is a combination of two effects, namely on-group effect and in-group effect, while the in-group effect is a weighted average of the X-Y effects within each group defined by the moderator.

  The research also proposes a typology of moderation, and proposes a genre of analyses, plenty and parity analysis, which is based on univariate analysis, moderation analysis, and causal dissection analysis.