Sociological Methods & Research, Ahead of Print.
For causal effects of a binary treatment on a right-censored duration, the widely used proportional hazard contrasts are non-causal with unrealistic restrictions. This article proposes an alternative flexible causal approach, where we estimate the ...