I've a problem. I'm trying to find out if the changes in municipal resources have an impact on individual physical activity level? Or perhaps this question would work better are the individual level factors or municipal level factor main drivers for individual's physical activity levels ?
I have a data from year 2000 physical activity (pa_vol; 3-level ordinary outcome) and background information from 3200 individuals and also municipal (115 municipals) resources used for physical activity. Then I've this same data from same individuals from year 2011. In my opinion this is suitable for multilevel model. Individuals are nested in municipals. But how should I add time variable, year, to this model? There are off course more variables than just these but just to get an idea (varibales:5 individual level variables, municipal type and 4 municipal resources variables)
I have done xtset ID year. And then I've used

xi:meologit pa_vol age i.gender cost_per_inhabitant || municipal code : || ID:
or

xi:meologit pa_vol i.year age i.gender cost_per_inhabitant || municipal code : || ID: ||year:

Which is the right way?

And the main results behind these data are physical activity level decreased between those 2 measurement points, and € used for PA increased.

I just ordered this Multilevel and Longitudinal Modeling Using Stata, Volumes I and II, Fourth Edition (eBook) but I don't know how soon I'll get it.
Thanks beforehand,
Virpi