My R package WeightIt has a huge new update, making it one of the biggest updates since I started the project. Version 1.0.0 introduces a few breaking changes, including the possibility that old results will not align with results from newer version of the package. In most cases, though, this only means improvements (i.e., better balance). For those that don’t know, WeightIt is an R package designed to provide access to propensity score weighting (also known as inverse probability weighting...| Posts | Noah Greifer
As I’ve been studying M-estimation and the covariate balancing propensity score (CBPS) (Imai and Ratkovic 2014), I’ve been noticing some interesting connections between these methods and want to share them with you. Logistic Regression and M-estimation First, what is logistic regression? I’ll discuss that in more detail in another post, but briefly it’s a way of modeling the relationship between \(K\) predictors \(\mathbf{X}\) (which include an intercept) and an outcome \(A\) (yes, I...| Posts | Noah Greifer
I’m oftenasked how the matching weights produced by MatchIt are computed. The weights are necessary for estimating the treatment effect in the matched sample; indeed, the weights determine the matched sample. While the weights for simple methods like 1:1 matching are straightforward (i.e., 1 if matched and 0 if unmatched), for more complicated scenarios, like full matching, matching with replacement, and variable ratio matching, the weights take on variable values and are critical to includ...| Posts | Noah Greifer
Multiply imputed data always makes things a little harder. Essentially, you have to perform each step of the analysis in each imputed dataset and then combine the results together in a special way.| Noah Greifer
A patient comes into urgent care presenting with an unpleasant rash on his arm. “Fortunately, we have a treatment,” says the doctor, “but it is expensive, has unpleasant side effects, and isn’t guaranteed to work.| Noah Greifer
Genetic matching sounds cool and science-y, something we social scientists love because nobody thinks what we do is “real” science. And genetic matching is cool and science-y, but not because it has anything to do with genes or DNA.| Noah Greifer
Today I’m going to demonstrate performing a subgroup analysis after propensity score matching using R. Subgroup analysis, also known as moderation analysis or the analysis of effect modification, concerns the estimation of treatment effects within subgroups of a pre-treatment covariate.| Noah Greifer
