| Previous: matchit: Create matched data | Up: matchit: Create matched data | Next: Syntax |
MatchIt implements the suggestions of () for improving parametric statistical models by preprocessing data with semi-parametric matching methods. It uses a sophisticated array of matching methods to select well-matched treated and control units from the original data set, thus reducing the dependence of causal inferences on functional form and other parametric assumptions. After pre-processing, MatchIt output can be used just like any other dataset in Zelig to estimate causal effects. In this way, MatchIt improves rather than replaces existing parametric models, reducing sensitivity to modeling assumptions. The matching methods available in MatchIt include exact matching on all covariates, nearest neighbor matching, subclassification, optimal matching, genetic matching, and full matching. An outline of all options are provided below; see the full documentation (available at http://gking.harvard.edu/matchit/) for more details.