¿Qué son las puntuaciones de propensión?

Abstract

Medical research is constantly looking for causality. Among study designs, randomized controlled trials are the most reliable way to estimate causal effects but are not always feasible. When this is the case, observational studies must be performed but this type of design unavoidably implies bias. Propensity scores, defined as the probability to receive a treatment conditional to a set of covariables allows to overcome confusion bias when searching for causal effects.