2020 U.S. Election Emails

As of Sep 7 2021, the database contains ~948,034 emails from 16082 senders. Request access to full dataset

Search the Database of State and Federal 2020 Election Emails

As of Sep 7 2021, the database contains ~948,034 emails from 16082 senders. Request access to full dataset

Summary of main findings

  • • We found six tactics that senders use to manipulate recipients into opening emails. The typical sender used at least one manipulative tactic in about 42% of their emails. Most senders — 99% — use them at least occasionally.
  • • We found 322 entities that shared our email address with other entities but the majority (133/322) had no privacy policy and only about a quarter (77/200) disclosed their email sharing in the privacy policy.

Less Emails

More emails

* These graphs are computed using keywords from our topic model.

Data Request

A preliminary release of the data is now available for download below. We will release additional data in due course.

Data release v1.0 documentation

Cite the paper

@article {doi:10.1177/20539517221145371,
title = {Manipulative tactics are the norm in political emails: Evidence from 300K emails from the 2020 US election cycle},
author = {Arunesh Mathur and Angelina Wang and Carsten Schwemmer and Maia Hamin and Brandon M Stewart and Arvind Narayanan},
year = 2023,
journal = {Big Data \& Society},
volume = 10,
number = 1,
pages = 20539517221145371,
doi = {10.1177/20539517221145371},
url = {https:/doi.org/10.1177/20539517221145371},
eprint = {https:/doi.org/10.1177/20539517221145371}
}
Princeton University
Data Driven Initiative

This project is led by researchers at Princeton University, affiliated with its Center for Information Technology Policy (CITP).