interrupted time series analysis with non-equivalent dependent variables

[Tom Scott]

Hello,

I'm familiar with Stata's arima and itsa commands for estimating interrupted time series analysis and comparative interrupted time series analysis, respectively, but I'm not sure of how to incorporate a non-equivalent dependent variable into these analyses. It seems like you have to run two separate analyses, one for each dependent variable, but does anyone know the correct way to test whether the difference in coefficients across dependent variables is statistically significant? Ariel Linden

Thank you for your help!