I am a co-author of RItools, a R package for randomization inference, with Ben Hansen and Jake Bowers. RItools currently provides tools for checking balance on control-treatment matchings and performing tests of sharp null hypotheses using a difference of means statistic. Current work in the
randomization-distribution branch seeks to provide a useful interface for the techniques developed in several of my published and unpublished papers.
I am also a co-author of Optmatch, a R package for optimal matching of observational studies based on a variety of criteria. Matching is a useful technique to make observational studies exhibit a key component of a randomized control trial: balance of covariates. Optmatch helps users created balanced subsets of control and treatment units so that on observed covariates, control and treatment groups appear similar. If unobserved variables correlate with observed covariates, they are also balanced under the match.
While optmatch is a separate project, users can benefit from the balance checking routines in RItools to evaluate the quality of proposed matchings.
For my fellow UIUC students I offer some help. Do you find yourself surfing the web and coming across some material that you need to access through the library’s proxy service? Drag Proxify! to your browser’s tool bar. When you find yourself on a page you need to acccess (e.g. Jstor, ACM Portal, Cambridge Journals), simply click the link and you’ll be sent through the proxy service.