The second method is nearest-neighbor matching. This matches each immunized child with a control that is the un-immunized child with the nearest propensity score. Although reasonable, on occasion this leads to matches that are quite far apart in terms of propensity score