3 problems

The paper considers causal risk-difference bounds under outcome-dependent sampling, with variables that may be exposure, outcome, instrumental-variable, or sampling variables; in t…

Let $(\lambda,\theta)$ denote the parameter, with nuisance parameter $\lambda$, identified or partially identified region $\Xi$, prior density $p(\lambda,\theta)$, first-order extr…

Let $X_{it}$ be a regressor taking values in a general, rather than binary, state space, and consider the identification bounds for the marginal effect in the nonseparable panel mo…