Does anyone have any idea of how the importance weights (iw) change the nearest neighbor (md) matching estimator defined by the user-written program md? I have looked at the help file & the power point presentation, but could not find an explanation.