and .
## : There is no causal effect of on given and . The NCO is independent of the treatment given and .
# U-Comparability: The unmeasured confounders of the association between and are the same for the association between and .

Given assumption 1 - 4, a non-null association between and , can be explained by , and not by another mechanism. A possible violation of Latent Exchangeability will be when only the people that are influenced by a medicine will take it, even if both and are the same. For example, we would