Identifiability of causal effects
the causal effect of intervention is identifiable if all backdoor paths between intervention and outcome variables can be blocked in the causal diagram; stratifying on the sufficient set removes all confounding bias in estimates of the causal effect of 
on 