Statistical procedure to reduce the bias introduced by confounding factors. One of these techniques is propensity score matching which models the probability of belonging to the treatment group of interest by accounting for covariate factors that may predict outcome.