ABSTRACT
Generalizability is a concern for many scientific studies, including those in HIV/AIDS research. The internal validity of causal effect estimates in randomized controlled trials or observational studies, including the application of methods to minimize the potential for bias due to measurement error, confounding, selection, and model misspecification, is well described. There are several methods that provide a quantitative approach for generalizing results from a randomized trial to a specified target population. Some of these methods utilize a model of the probability of trial participation conditional on covariates. This chapter considers estimators based on sampling scores to generalize trial effects to a specified target population. Estimators that employ sampling scores and their corresponding confidence intervals provide inferences about the treatment effect in the target population, that is, a contrast in the survival outcome had everyone in the target population received treatment compared to the outcome if everyone in the target population did not receive treatment.
