IRS Non Profit Financial Data and Political Partisanship

Published: 9 March 2021| Version 3 | DOI: 10.17632/nfgb92wh72.3
Contributors:
Benjamin Ho,
,

Description

We classified the political ideology of all non-profits in the public IRS 990 Form dataset by using a Bayesian classification of their self-reported mission statements, using a classifier trained on the political speeches of members of Congress.

Files

Steps to reproduce

Downloaded IRS 990 Data from IRS AWS servers and open990. Downloaded Congressional Speech Data from Genztkow and Shapiro's Stanford repository. Ran survey data on Amazon mturk.

Institutions

Vassar College, University of Chicago, KDI School of Public Policy and Management

Categories

Economics, Political Science, Non-Profit Entity

Licence