Longitudinal Regression Issue

Hi there, I am conducting a hierarchical longitudinal regression over an 18 month time span. I have main effects, quadratic terms and interactions for the same variables measured in T1 and T2 (T1 variables entered as controls). The specific issue is that the beta weights associated with the significant variables at both T1 and T2 are changing signs. For example X1*X2 = .101 at T1 and then X1*X2 = -.090 at T2.

Could this be due to multicollinearity? If so is there any way I can correct this? Any insight would be appreciated!!! :)


TS Contributor
i do not know hierarchiacl longitudinal regression but i know repeated measures anova if you have any more questions. As for this now, one way to deal with collinearity is to standardize all your variables.