Robustness of Causal Claims
David Liu comments in regards to "Robustness of Causal Claims" :
It was said in the paper that: "if the two estimates of b happen to disagree in a specific study, we can conclude that the disagreement must originates with violation of those extra assumptions that are needed for the second, and we can safely discard the second in favor of the first." But, if the first (minimal assumptions set) is false, then the two estimates of b may disagree too. So we can only conclude that the first or the second is false.