MIT Logo
Home ] Research ] [ Statistical Software ] Videos ] Courses ] CV ]

Software Packages

Synth: Synthetic Control Method

Synth is a statistical software package for R, Stata, and MATLAB that implements synthetic control methods for causal inference in comparative case studies as described in Abadie and Gardeazabal (2003)Abadie, Diamond, and Hainmueller(2010), Abadie, Diamond, and Hainmueller (2011), and Abadie, Diamond, and Hainmueller (2013).

See a short video about the method by clicking on the picture below.

ebal: Entropy Balancing

ebal is a statistical software package for R and Stata that implements entropy balancing for causal effects as described in Hainmueller (2012).

KRLS: Kernel Regularized Least Squares

KRLS is a statistical software package for R that implements Kernel Regularized Least Squares (KRLS) as described in Hainmueller and Hazlett (2013). KRLS is a machine learning method that is designed to flexibly fit response surfaces of the form y=f(x) as they arise in regression and classification problems without relying on linearity, additivity, or other assumptions that use the columns of the predictor matrix X directly as basis functions.


Conjoint Survey Design Tool

The Conjoint Survey Design Tool assists researchers in creating multi-dimensional choice experiments that can be readily incorporated into any pre-existing web survey software (such as Qualtrics). Conjoint analysis is a type of survey experiment often used by market researchers to measure consumer preferences over a variety of product attributes. Hainmueller, Hopkins and Yamamoto (2013) demonstrate the value of this design for political science applications. Conjoint experiments present respondents with a choice among set of profiles composed of multiple randomly assigned attributes. This approach allows researchers to estimate the effect of each individual component on the probability that the respondent will choose a profile. This software tool is designed as a companion to Hainmueller, Hopkins and Yamamoto (2013), providing a graphical user interface for generating conjoint experiments.