Rosenbaum and Rubin (1983) showed in observational studies conditioning on propensity scores can lead to unbiased estimates of the exposure effect as long as certain assumptions hold:
There are no unmeasured confounders
Every subject has a nonzero probability of receiving either exposure
Change this in chapter 7:
Rosenbaum and Rubin (1983) showed in observational studies conditioning on propensity scores can lead to unbiased estimates of the exposure effect as long as certain assumptions hold:
There are no unmeasured confounders Every subject has a nonzero probability of receiving either exposure
change to our normal assumptions