Replication data for the paper "Does Police Militarization Increase Repression?"
DOI: 10.1177/00220027231187301
Authors:
- Martin Stavro
- Dr. Ryan Welch
This document contains instructions on how to replicate the analyses from "Does Police Militarization Increase Repression?" Viewing the commit history also provides a record on how the analyses changed over time (model refinements, debugging, etc.).
- R (preferably up to date)
- RStudio (preferably up to date)
- Some free memory
The code provided is intended to be plug and play. All you need to do to replicate the analyses in the main paper and appendix is download "DPMIR Replication Script.R" and run all the code in the script using CTRL + SHIFT + ENTER. The R script has been set up to automatically pull the data from the repository by passing the raw content link to the read_csv() function. Alternatively, you can download all of the repository as a .zip file.
The code is divided into different sections depending on what table/figure is being replicated. In RStudio, you can use these comment dividers for navigation similar to a "table of contents" as seen in the image below.
It is highly recommended that you run the code in its entirety in its original order. This makes it easy to replicate all the data correctly. While the code is divided into sections, some sections cross-reference objects (regressions, etc.) created in earlier sections, so skipping around may produce errors.
After running the replication code in its entirely, replicated tables/figures are stored as objects. These objects can be called using this general syntax:
- For tables: outputT# (where # is the number of the table. For example, table 1 is outputT1).
- For figures: outputF#
- For tables/figures in the appendix: outputTA# or outputFA# Note that there is no object for outputF1, as this figure is simply a graphic and was not created in R.
R failed to fetch the data files from the GitHub repository on its own. This can typically be solved in one of two ways:
- Try running the script again (this tends to fix it)
- Manually change the content inside read_csv() to connect to a downloaded version of the GitHub data files on your local machine (an easy way to do this is some_data <- read_csv(file.choose())
- DPMIR Replication Script.R: The one-stop shop to replicate everything in the paper.
- DPMIR Main Data: The dataset that most of the paper relies upon.
- DPMIR US Recoded: A modification to the main data where the US is recoded to have police militarization (0 -> 1; we address in the paper why it was 0 to begin with).
- DPMIR US Removed: A modification to the main data where US cases are removed.
- Geocoding: A parsed version of the main data with a geocoded field allowing for a nice visualization with googleVis.
- License: The fine print. It's an MIT License. It lets you do with the data what you please, with some stipulations.