Main content



Loading wiki pages...

Wiki Version:
Code and data used to perform the analysis are included in the "Files" section of this web page. The file `FakeMedicare2005_2012.csv` is provided as an example of the Medicare dataset used in the manuscript instead of the real data because the actual Medicare dataset cannot be shared. This dataset was simulated from actual Medicare data. For an example of how the analysis was performed, download all files to the same directory and run the R code in `DinD_workflow.R`. This file includes three functions: 1. `read_BIGMED` reads the simulated Medicare file and converts annual outcomes counts to rates using the appropriate denominator 2. `diff_in_diff` performs the difference in difference regressions and formats the results 3. `hypotest` performs the hypothesis tests described in the manuscript and summarizes the results. Besides the Medicare data, two other data sources are not provided on this page. These are: 1. Esri Business Analyst Census Data, which was obtained from Esri ( 2. PM2.5 concentrations from Boys et al. (2014). These were obtained from the authors, who we encourage any readers to contact if they are interested. A second resource may be the group website: