Average Causal Effects from Non-Randomized Studies:
A Practical Guide and Simulated Example

Department of Statistics

The Methodology Center

Penn State

Here you can access data and code that accompany the manuscript "Average Causal Effects from Non-Randomized Studies: A Practical Guide and Simulated Example" by J.L. Schafer and J.D.Y. Kang.

• Initial sample of N=6,000 from our artificial population: Diet0001.dat  (381 Kb)

• Information about data and list of variables: readme.pdf

R code for computing average causal effects and standard errors by each method described in the article on the initial sample

• Mean comparisons and regression adjustments: ancova.R

• Regression estimation: regression.R

• Mahalanobis-metric matching within calipers defined by the logit-propensity score: matching.R

• Inverse-propensity weighting: ipw.R

• Propensity-based subclassification: subclass.R

• Regression estimation with weighted residual bias corrections: weightres.R

• Weighted regression estimation: weightreg.R

• Regression estimation with propensity-related covariates: propcov.R

This page is hosted by the Department of Statistics in the Eberly College of Science, and by  The Methodology Center in the College of Health and Human Development, The Pennsylvania State University.

Support for this project is provided by a grant from the National Institute on Drug Abuse, 2-P50-DA10075.