This repository contains code to replicate the primary analysis in the paper "Showing High-Achieving College Applicants Past Admissions Outcomes Increases Undermatching", by Sabina Tomkins, Joshua Grossman, Lindsay Page, and Sharad Goel.

Please use the following citation when referencing this work or data: CITATION TO BE UPDATED .

We have provided anonymized and aggregated data to replicate the primary analysis shown in Table One. This analysis shows that students exposed to the tool described in the paper undermatched at a greater rate.

Code Description

The script , contains the code for the primary analysis, and reads in a file which is in this repository "table_one_column_one.csv".

Requirements:

• tidyverse 1.3.1
• stargazer 5.2.3

To run the code navigate to this directory and execute Rscript run_analysis.R

Data Description

In the paper each observation is a student. Here, to preserve anonymity, the level of observation is a high-school year. In this file each row is an observation at a given school in a given year. The value stored in the undermatched column is the proportion of high-achieving students at this school who have undermatched in this year. The value in the column named treatment is True if we estimate that students had access to the Scattergram tool, and False otherwise.

If you have any questions please contact Sabina Tomkins at stomkins@umich.edu.