Mitigating the Turnout Effects of Bad Weather With Early Voting: 1948–2016

• Author: Martin Johnson,Robert M. Stein
• DOI: http://doi.org/10.1177/1532673X221132480
• Published date: 01 March 2023
• Date: 01 March 2023
• Subject Matter: Articles

Article

American Politics Research

2023, Vol. 51(2) 197–209

© The Author(s) 2022

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/1532673X221132480

journals.sagepub.com/home/apr

Mitigating the Turnout Effects of Bad Weather

With Early Voting: 1948–2016

Martin Johnson

1

and Robert M. Stein

2



Abstract

We identify in-person early voting and no-excuse mail voting as antidotes for the depressing effect inclement weather has on

voter turnout and the Republican dividend that accompanies rain and snow on Election Day. We offer and test an explanation

for how voters utilize early voting to anticipate and avoid the costs of voting in bad weather. Replicating and extending Gomez

et al (2007) analysis through the 2016 election, we conf‌irm the remedial effect in-person early voting and to a lesser degree no-

excuse mail voting has on turnout and the Republican advantage when bad weather coincides with Election Day. Our work

makes an important contribution to understanding how election laws effect voter participation. We discuss how taking seriously

treatment effect heterogeneity both in theoretical and empirical analyses might con tribute to our understanding of the effects of

election laws on voter participation.

Keywords

elections, early voting, political behavior

Introduction

Everyone complains about the weather, but no one ever does

anything about it. This may not be true about the weather on

Election Day. Gomez et al.’s (2007) seminal work demon-

strated that when it rains or snows on Election Day voters stay

home. Moreover, Democratic voters are more averse to

voting when bad weather coincides with Election Day than

Republican voters. The signif‌icant and negative effect in-

clement weather has on turnout and the resulting Republican

dividend that comes with bad weather has persisted for the

last 50 years.

We identify in-person early voting and to a lesser extent

no-excuse absentee voting as an antidote for the depressing

effect inclement weather has on voter turnout and the Re-

publican dividend that accompanies rain and snow on

Election Day. Replicating and extending Gomez et al.’s

analysis through the 2016 election, we conf‌irm the remedial

effect in-person early voting and no-excuse absentee voting

has on turnout and the Republican advantage when bad

weather coincides with Election Day. We offer and test an

explanation for how voters utilize early voting to anticipate

and avoid the costs of voting in bad weather.

There is only a modest consensus on how convenience

voting shapes voter participation (see Menger & Stein, 2020).

Several researchers have found in-person early voting to have

either no or a marginal effect on turnout (Stein, 1998;Stein &

Garcia-Monet, 1997;Wol f‌inger et al., 2005). More recently

others have found the opportunity to vote before Election Day

has a negative effect on voter turnout (Burden et al., 2014;

Larocca & Klemanski, 2011;Richey, 2008). Few researchers

have identif‌ied no-excuse absentee voting as having a pos-

itive effect on voter turnout (Menger & Stein, 2020). We f‌ind

the impact of early voting, both in-person and no-excuse

absentee voting, on turnout is conditional on the nature and

scope of the costs to voting. We f‌ind that the opportunity to

vote before Election Day moderates and in some instances

reverses the negative effect inclement weather has on voter

turnout and the GOP share of the vote.

Our work makes an important contribution to under-

standing how election laws effect voter participation. We

explain several anomalous f‌indings in the literature on the

turnout effects of convenience voting by conditioning the

turnout effect of early voting on the nature and scope of

voting costs. We discuss how taking seriously treatment

effect heterogeneity both in theoretical and empirical analyses

1

Manship School of Mass Communications, Louisiana State University, Baton

Rouge, LA, USA

2

Lena Gohlman Fox Professor of Political Science, Rice University, Houston,

TX, USA

Corresponding Author:

Robert M. Stein, Department of Political Science, Rice University, P.O. Box

1892, Houston, TX 77221, USA.

Email: stein@rice.edu