Hello. I am currently doing an interrupted time-series analysis. I understand how to build the (p,d,q) model using the auto and partial correlation. However, I really have no idea how to test whether the intervention was effective. I've looked at some books but they make very little sense to me. I've seen pages on transfer functions and forecasting and am unsure about what to do. If anyone knows anything, I would greatly appreciate some guidance. Thank you.

Juan