Figure 1From: Estimating the treatment effect from non-randomized studies: The example of reduced intensity conditioning allogeneic stem cell transplantation in hematological diseases Illustration of the different distributions of a covariate (X) in two non-randomized samples (A & B). The propensity score method (PS) aims at re-creating the conditions of a pseudorandomization, while the inverse probability weighting (IPW) approach aims at re-creating a pseudopopulation where patients A and B are exchangeable. Both methods aim at obtaining a similar distribution of the covariate X in the two groups.Back to article page