TY - RPRT TI - Estimating causal effects from large data sets using propensity scores. AU - D B Rubin PY - 1997 DO - 10.7326/0003-4819-127-8_part_2-199710151-00064 UR - https://pubmed.ncbi.nlm.nih.gov/9382394/ ID - 9382394 ER -