Hello,

I hope you can guide me. I am working with panel data, the data is daily data for 90 days nested within individuals. I am interested in the trajectory over the 90 days period of two binary variables (yes or no answer), I will call them var1 and var2. So far, I tested the trajectories (independently) using mixed effect models (command below):
meglm var1 DAY || ID: DAY , or intmethod(lap) family(bernoulli) link(logit) dif
meglm var2 DAY || ID: DAY , or intmethod(lap) family(bernoulli) link(logit) dif

I got that the trajectory of var 1 is significant and increasing over time, while the trajectory of var2 is not significant at the 95% but it is at the 90%. I want to test if these trajectorie are asociated and of that asociation is significant. My question is: do you know how to test it? which would be the best way to do it?

I have read about people who suggest collapsing at the second level, I also saw correlation between time series, which is not appropiate in my case as time-series are not the same than panel data. Any thoughts? I also saw someone conducting the models for the trajectories (comands above) and then predicting standardized residuals and using an xtreg, re model to test the association. If, that is the correct approach, what I am looking for in the xtreg, re output?

Thank you