The difference-in-differences analysis is used to evaluate the effect of the intervention while controlling for baseline and endline differences in the study arms. It stems from a causal inference perspective to understand the true effect of the intervention. Originally, it is more suitable for linear outcomes however, it can equally be used for binary outcomes (Yes/No), which is the format for our data. On the other hand doing a normal chi-square between the intervention & control will not really tease out the baseline differences. It would be helpful if you have come across such and would like to offer suggestions. Thank you.

More Udochisom Anaba's questions See All
Similar questions and discussions