I just started reading about difference-in-differences (DiD), and was wondering if it can be applied to my problem. I am unable to find any good references on this. I am considering the effect of introducing late-night taxis to cities and whether it resulted a reduction in drunk driving arrests. However, the treatment here (introducing late-night taxis) had many characteristics, i.e., the intensity at which it was introduced (# cars introduced), type of cars introduced (sedans, SUVs, yellow cabs, etc.), the fares at which they were introduced (different cities had different prices). I want to understand the effect of all of these treatment groups. Is it possible to model all of them in one equation itself? Can I have multiple treatment variables?