I was wondering if there was any rule for choosing moderating variable. I am aware that moderating variable needs to have strong theoretical support in relation with the independent variable. However, there could multiple moderating variables having strong theoretical support relation with the independent variable (e.g. sex, gender, demographics, income, popularity etc.). Is it important to have motivation to choose a moderating effect among a group of variables considering all have theoretical support? Kindly, type down an article that you believe would be helpful.
Thanks!
Including moderation effects generally entails the idea that the main causal relationship between the central independent and dependent variable fundamentally changes. So yes, the selection of a moderating variable should be carefully underpinned thereotically. As far as I know, including moderation variables as control variables is not done too often. Normally, if your paper is theoretically driven, the selection of a moderation effect will feel quite 'natural'. For example, I am currently writing a paper on the relationship between grade retention and tolerance for people with migration background. I, however, believe that being retained increases frustration levels, which in turn decreases intergroup tolerance. In this case, the moderation effect strenghtens the understanding of the causal relationship in reality.
In multi-level modeling, if a variable predicts the effect size, that variable is called the moderator variable. Alternative: In a non-linear relationship, a significant value of a moderator variable does not prove the true moderator effect.
It has to go according to a theoretical framework that suggests that it modulates the relationship between other variables.
Moderator variables are to so chosen based on their capacity to deplete or embolden the relationship between the respective independent and dependent variables, drawing upon the literature. the most insightful document in this regard is linked here.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of personality and social psychology, 51(6), 1173