Messages matter: How voter education campaigns affect citizens’ willingness to vote for women
Governments and organizations worldwide pour money into campaigns designed to increase female political representation, including voter education campaigns. But do such campaigns promote women in politics? We argue that where single-member district contests and clientelism incentivize voters to support viable candidates, information about discrimination against women can undercut support for women in elections. Instead, messages that stress women candidates’ viability are more effective. We work with one of the longest-running voter education campaigns, Malawi’s 50:50 campaign, to combine randomized exposure to campaign videos with a conjoint experiment and text analysis of respondents’ answers to open-ended questions. We find that exposure to a campaign message makes participants more willing to vote for a woman. But, in line with our argument, a campaign message that includes information about the progress of women in politics has a stronger positive effect than one that discloses information about discrimination against women candidates.
Journal of Politics 88(1): 145–161, 2026.
Journal Data & Replication Pre-analysis plan Blog
Citation
@article{k._ofosu2026,
author = {K. Ofosu, George and Bech Seeberg, Merete and Wahman,
Michael},
title = {Messages Matter: {How} Voter Education Campaigns Affect
Citizens’ Willingness to Vote for Women},
journal = {Journal of Politics},
volume = {88},
number = {1},
pages = {145-161},
date = {2026-01-01},
url = {https://www.journals.uchicago.edu/doi/10.1086/734252},
doi = {10.1086/734252},
langid = {en},
abstract = {Governments and organizations worldwide pour money into
campaigns designed to increase female political representation,
including voter education campaigns. But do such campaigns promote
women in politics? We argue that where single-member district
contests and clientelism incentivize voters to support viable
candidates, information about discrimination against women can
undercut support for women in elections. Instead, messages that
stress women candidates’ viability are more effective. We work with
one of the longest-running voter education campaigns, Malawi’s 50:50
campaign, to combine randomized exposure to campaign videos with a
conjoint experiment and text analysis of respondents’ answers to
open-ended questions. We find that exposure to a campaign message
makes participants more willing to vote for a woman. But, in line
with our argument, a campaign message that includes information
about the progress of women in politics has a stronger positive
effect than one that discloses information about discrimination
against women candidates.}
}