Non-Disruptive Protest Can Work: Evidence From the Women's March

Abstract

Do protests impact politics? Many scholars argue they do, but disagree about their mechanisms of impact, with some arguing protests must disrupt to have impact while other focus on protest’s processes of political activation. Findings on protest’s political impact from prior literature also often suffer from endogeneity
problems that make it difficult to tease out causal relationships. In this article, I provide a rigorous test of the political impact of protest in a case that allows for further insight into the mechanisms for protest’s impact: the 2017 Women’s March in the United States. I test the impact of the Women’s March using detailed geo-coded data on local marches’ size and location. To isolate the Marches’ effects I employ an instrumental variables analysis, instrumenting march size with precipitation and temperature data. I measure the effects of instrumented march size on three variables: the creation of “Indivisible” groups, donations to Democratic politicians, and 2018 election Democratic vote share. I find the Women’s March had a significant positive effect on all three dependent variables. These findings provide strong evidence that protests can have significant political impacts, even when they involve no disruption.

Full text

Non-Disruptive Protest Can Work: Evidence From the Women’s
March
Jonathan Pinckney∗
Word Count: 9,853
Keywords: protest, social movements, Womens March, elections, mobilization
Acknowledgments
Many thanks to Thea Johansen for her excellent research assistance, and to Charles Butcher, Erica
Chenoweth, Bob Edwards, Banks Miller, Martin Smidt, Ingrid Vik Bakken, Marius Wishman and all the
participants at my panel at the Mobilization Conference on Nonviolence and Social Change for helpful
comments and feedback on earlier versions of this paper. The early stages of this research were supported
by generous internal funding from the Norwegian University of Science and Technology during my time as
a post-doctoral researcher there.
∗Corresponding Author: School of Economic, Political, and Policy Sciences, The University of Texas at Dallas. 800 W
Campbell Rd, Richardson, TX, 75080. Phone: 202-971-6469. e-mail: jonathan.pinckney@utdallas.edu

Abstract
Do protests impact politics? Many scholars argue they do, but disagree about their mechanisms of impact,
with some arguing protests must disrupt to have impact while other focus on protest’s processes of political
activation. Findings on protest’s political impact from prior literature also often suffer from endogeneity
problems that make it difficult to tease out causal relationships. In this article, I provide a rigorous test of the
political impact of protest in a case that allows for further insight into the mechanisms for protest’s impact:
the 2017 Women’s March in the United States. I test the impact of the Women’s March using detailed
geo-coded data on local marches’ size and location. To isolate the Marches’ effects I employ an instrumental
variables analysis, instrumenting march size with precipitation and temperature data. I measure the effects
of instrumented march size on three variables: the creation of “Indivisible” groups, donations to Democratic
politicians, and 2018 election Democratic vote share. I find the Women’s March had a significant positive
effect on all three dependent variables. These findings provide strong evidence that protests can have
significant political impacts, even when they involve no disruption.
1

1 Introduction
The Women’s March of January 21st, 2017 is one of the most prominent examples of public protest in
American history, and was followed by a massive outpouring of social dissent that became known as “The
Resistance” (Andrews, Caren, and Lu 2020; Gose and Skocpol 2019; David S. Meyer and Tarrow 2018). In
the 2018 mid-term elections, the first chance for the national electorate to weigh in on the Trump presidency,
the Democratic party won the popular vote by more than seven percent and captured control of the House of
Representatives in a “blue wave” (Fisher 2019). Many took the Democratic shift as indicative of the power
of the movement sparked by the Women’s March.
Yet other voices leveled harsh criticisms. Some of these centered on matters of racial or gender justice, for
instance criticizing the predominately white leadership of the Women’s March organization for appropriating
the terminology of the 1963 “March on Washington,” for the transwoman-excluding nature of some of the
protest imagery, or around polarized opinion on the Israeli-Palestinian conflict.1
Yet a set of more fundamental critiques called into question the very logic of an event like the Women’s March.
Columnist David Brooks stated it bluntly: “These marches can never be an effective opposition. . . marching
is a seductive substitute for action” (Brooks 2017). The center-right Brooks was joined by a chorus of
criticism from the Left, with critiques centering on the march’s focus on vaguely positive big-tent messaging
rather than substantive policy demands, and on the lack of disruptive or confrontational behaviors. Some
activists who had attended the march, reflected: “It felt more like a charity 5k than a protest,”2or, more
pointedly “Why did the Women’s March feel so ineffective?”3
Which of these stories is correct? Answering that question is critical because it goes to deeper debates about
efficacy in social movements. Do protests lead to political change primarily because they disrupt current
politics, imposing costs on elites (Piven and Cloward 1977)? Or do they lead to political change primarily
through their activating and mobilizing influence on participants (Madestam et al. 2013; Wasow 2020;
Gledhill, Duursma, and Shay 2022)? Should social movement leaders seek, like the leaders of the Women’s
March, to focus on broad, positive messaging that maximizes participation? Or should they focus on tactics
that confront and disrupt?
Yet answering these critical questions is inherently empirically challenging. At their core, major moments
of protest reflect deeper changes in public opinion, and may only be symptomatic of underlying trends
rather having their own direct effects. Individual protests also typically take place during sequences of social
1For an excellent summary of some of the early critiques, see Gantt-Shafer, Wallis, and Miles (2019).
2https://www.thecrimson.com/column/femme-fatale/article/2017/2/16/hu-womens-march/
3https://medium.com/when-women-speak-back/why-did-the-womens-march-feel-so- ineffective-b4c1d510e731
2

movement tactics, making it difficult to disentangle to what degree a particular protest like the Women’s
March impacted outcomes independent of other events.
These challenges in part explain why, while there is a robust literature on the outcomes of protest (Amenta
et al. 2010), this literature often comes to conflicting answers (e.g. Giugni 2007; Weldon 2022), both on the
general impact of social movements and on the specific impact of particular tactics like public demonstrations.
In this paper I build on this robust literature. I leverage detailed geocoded data on local Women’s Marches
in 2017 to examine the Marches’ local-level effects on three areas: movement building, as measured by the
creation and size of “Indivisible” groups (Brooker 2018; Corrigall-Brown 2021); political donations to the
Democratic party and Democratic candidates, and electoral outcomes, as measured Democratic party vote
share in the 2018 midterm elections. Regression models show a strong and highly significant effect on all
three dependent variables.
Yet how can we be confident these effects are due to the Women’s Marches? I isolate the effects of the
Women’s Marches through an instrumental variable analysis. Drawing on an extensive literature on the
impact of good weather on public protest (e.g. Madestam et al. 2013; Wasow 2020; Caren, Andrews, and
Nelson 2023) I instrument the level of participation in the 2017 Women’s March with measures of county-
level precipitation and temperature. When instrumented, not only do the effects of the Women’s March
on Democratic vote share, political donations, and movement building remain statistically significant, they
double or triple in size. I thus conclude that the Women’s Marches played a key role in the growth of the
“Resistance” and the 2018 Democratic “Blue Wave.” This further provides evidence not just that public
protest “works,” but that its mechanisms of impact do not rely on disruption, but rather through the
activation of large new constituencies socialized into greater political participation.
The article proceeds as follows. In the following section I discuss what we know about the political effec-
tiveness of protest. Then I discuss the origin and dynamics of the 2017 Women’s March. Then I introduce
my research design, including the use of good weather as an instrument for protest size. I then present and
discuss my findings, before concluding with implications of the research for the study of protest.
2 Literature Review
Most social movements scholarship strikes generally optimistic tones about the power of protest to create
political change (Baumgartner and Mahoney 2005; Chenoweth and Stephan 2011; Han, McKenna, and Oy-
akawa 2021; Weldon 2022; Wouters and Walgrave 2017). Other scholars tend to come down more negatively
3

(Burstein and Linton 2002; Giugni 2007; Skocpol 2003; S. Soule et al. 1999), and some point out that
the generally positive tone of the literature on social movement efficacy is likely driven in part by selection
bias (Amenta, Andrews, and Caren 2018; Bosi and Uba 2021). Yet most would agree that the most honest
answer to the question of whether protest “works” is, of course, “it depends.”4It depends first on context
(Uba 2009). Protest, like any social movement tactic, exists in a political opportunity structure that shapes
protest’s emergence, the resources that movements can mobilize, and any possible outcomes (Kitschelt 1986;
McAdam 1982; D. S. Meyer and Minkoff 2004; S. A. Soule and Olzak 2004).
It depends second on what we mean by working. Scholars have examined a vast array of political outcomes
that protest might impact. A non-exclusive list of these include changes in public opinion (Branton et al.
2015; Collingwood, Lajevardi, and Oskooii 2018; Mazumder 2018; Reny and Newman 2021; Wouters 2019),
agenda-setting (King and Soule 2007; Johnson 2008; Walgrave and Vliegenthart 2012), and policy change
(Htun and Weldon 2012; Olzak 2021; Sato and Haselswerdt 2022). For instance, Gause (2020) finds that
legislators are more likely to support the preferences of disadvantaged protesting groups, and Gillion (2012)
finds that minority protest in congressional districts changes roll-call vote behavior.
When it comes to influencing elections, the literature is a bit sparser. Andrews (1997) finds a range of
electoral effects from civil rights mobilization in Mississippi. McAdam and Tarrow (2010) called for greater
attention to the interconnection between movements and elections, and highlighted six mechanisms linking
movements to electoral campaigns. Gillion and Soule (2018) finds that protest in US congressional districts
significantly increased vote share for the party most closely aligned with the protesters’ issues. And Bremer,
Hutter, and Kriesi (2020) find that economic protest across 30 European countries from 2000 to 2015 was
associated with greater electoral punishment of incumbents. I discuss several other recent studies on protest
and elections in the section on causal inference below.
It also depends on what we mean by protest. Mass mobilization takes a variety of forms, with differing
attendant mechanisms for potential impact. Scholars differ on which of these mechanisms are efficacious.
One influential set of arguments focuses on how protest functions as a costly signal to political elites
(Gause 2020; McAdam and Su 2002). In a classic treatment of the subject, Lohmann (1993) argues that
protest signals elites on underlying shifts in public opinion. Size is a clear indicator of the strength of this
signal (DeNardo 1985).5Protest frequency is another important indicator of the strength of elite signaling
(Chenoweth and Belgioioso 2019; Olzak and Soule 2009). Magali Fassiotto and Sarah Soule show that, for
4For excellent summaries of major research efforts on the effectiveness and long-term outcomes of social movements, see
Amenta et al. (2010) and Amenta, Andrews, and Caren (2018).
5Though see Butcher and Pinckney (2022) for cross-national research indicating some complicating factors in the signalling
power of protest size.
4

Women’s protests in the United States, not just the strength but the clarity of the signal also matter, and
caution movements against “muddying their signal by including too many other groups or issues” (Fassiotto
and Soule 2017, 30). Similar insights come from Mueller (2024) and Wouters and Walgrave (2017).
Another focuses on the disruption of public contention, and the ability of protest to impose direct economic
or political costs on its opponents (Gamson 1975; Piven and Cloward 1977). Labor strikes and other methods
of noncooperation are the most prominent form of action that rely on this mechanism. Yet other work has
examined the impacts of rioting, property destruction, or other forms of violent protest (Enos, Kaufman,
and Sands 2019; Kadivar and Ketchley 2018)
Finally, a set of arguments focus on the ways in which participation in protest activates protesters. Partici-
pation in protest tends to lead to greater political participation in the future, as well as socialization into the
attitudes of the protest group (Opp and Kittel 2010). Mass demonstrations operate in this framework not as
a direct form of pressure or signalling to elites, but as a recruitment strategy. This recruitment strategy may
be particularly effective if it is focused on providing positive, enjoyable experiences (Jasper 2018; Gledhill,
Duursma, and Shay 2022). As Rules for Radicals author and organizer Saul Alinsky (1989, 132) put it: “A
good tactic is one that your people enjoy.”
While these mechanisms overlap, which of these we think the political effects of protest primarily operate
through leads to significantly different implications. If protest functions primarily as signal, then organizers
should select tactics that boost the strength and clarity of the signal. If protest works primarily through
disruption, then organizers should seek to disrupt. And if protest works primarily through its indirect,
activating effects then organizers should select tactics that most powerfully activate the largest number of
people as possible.
The characteristics of the Women’s March, while they do not provide a full range of comparison cases for
evaluating the efficacy of these mechanisms, at least provide interesting detail on which of these are necessary
or sufficient to achieve political impact. I return to this below after first touching on some of the empirical
challenges in evaluating the effects of protest.
2.1 Challenges to Inference
The complex set of context-dependencies and varying mechanisms described above help explain why opinion
on protest’s impact is conflicted. In addition, most studies of protest also face a fundamental endogeneity
problem. The occurrence and intensity of protest are not randomly assigned, exogenous shocks to the
political system (Bosi and Uba 2021; Burstein and Sausner 2005). Many of the observable correlates of
5

protest and political impact overlap. Insofar as these correlates are measurable, this issue can be remedied
in statistical analysis through including appropriate control variables. However, many of the factors that
plausibly influence both protest and political outcomes are not so easily observed. Thus, correlations between
protest and political outcomes are empirically suspect.
Some work has begun to address these challenges, but leaves many questions unanswered. Biggs and An-
drews (2015) find that lunch counter sit-ins made desegregation significantly more likely in cities across the
American South in 1960, controlling for several determinants of protest. Yet their measure of protest is a
simple binary variable that does not allow for disaggregation in terms of size or intensity of protest.
Madestam et al. (2013) show that protest size on Tax Day, 2009, which they instrument using rainfall,
significantly predicts several indicators of movement success for the Tea Party. Their strategy addresses
many of the methodological challenges of determining the impact of individual protest events, but the limited
scale of the Tea Party tells us little about protest more broadly. Furthermore, their rainfall instrument is
a simple binary indicator of whether or not a county experienced more than 0.1 inches of rain, leaving out
more complex weather dynamics.
Omar Wasow (2020) examines the attributes of effective protest in more detail with an examination of
the civil rights movement in the 1960s. Wasow shows that peaceful protests significantly increased public
attention to the issue of civil rights, and that counties within 100 miles of a peaceful protest had major shifts
towards the Democratic party in the 1968 presidential election, while counties within 100 miles of violent
riots in the aftermath of the assassination of Dr. Martin Luther King, Jr. had significant shifts towards
the Republicans. Yet Wasow does not include any measures of protest intensity to distinguish between the
effects of small protests and large protests.
Caren, Andrews, and Nelson (2023) show that intensity of peaceful protest during the month-long peak of
the 2020 Black Lives Matter movement significantly increased Democratic vote share at the county level
in the 2020 presidential election, while violent protest decreased Democratic vote share. They address the
endogeneity problem through an instrumental variables strategy using average good weather over the course
of June 2020 as a predictor of protest intensity to address endogeneity concerns.
In this article, I build on these scholars’ insights. The Women’s March has several characteristics which make
it ideal for this testing. First, the march was an outlier in American protest, both in size and dispersion.
Thus, it provides a strong “best-case scenario” test of the power of protest. If protest can be effective, the
Women’s March protests should be effective. If the Women’s March failed to have a significant political
impact, it provides strong evidence for skeptics of the power of protest. Second, the large number of events
6

taking place on the same day provides an ideal environment for a natural experiment. Some localities
experienced the simultaneous treatment of a Women’s March on January 21, 2017, and others did not.
Once I have addressed the endogenous aspects of generating the protest itself (which I discuss below), this
simultaneous shock gives us strong grounds for causal inference. In contrast, when looking at longer protest
waves such as the civil rights movement or the 2020 Black Lives Matter wave, diffusion from earlier points
in the wave may impact later protest dynamics.
So, having framed the discussion on the general effects of protest and the challenges to inference in studies of
protest in general, if the Women’s March “worked,” what should we expect that “working” to look like? And
in particular, which of the mechanisms described above should we expect to operate in the Women’s March?
I answer this by examining the march’s precipitating causes, characteristics, and its immediate aftermath.
3 The 2017 Women’s March
Planning for a “Million Woman March” on Washington to protect Women’s rights began immediately fol-
lowing the 2016 election of President Donald Trump. Teresa Shook, a retired lawyer from Hawaii, created
a Facebook group planning an event after discussions on the pro-Hillary Clinton Facebook Group “Pantsuit
Nation.” The event, publicized in the midst of the beginning of a wave of protest against President Trump’s
election, grew rapidly, with over 10,000 people saying they would participate in the first 24 hours and over
100,000 soon afterwards (Stein 2017).
The march was not explicitly partisan, and the organizers went to some lengths to ensure their organization
was not seen as an arm of the Democratic party. However, the causes around which march participants
organized were overwhelmingly on the left of the American political spectrum (Fisher, Dow, and Ray 2017).
The marches were geographically dispersed. In addition to the primary event in Washington, DC, organizers
planned over 600 “sister marches.” The largest events took place in major left-leaning cities, but marches
took place across the country, in small relatively rural areas and in purple and red states as well as blue
states. A comprehensive tally of events from the day of the Women’s March includes everything from the
estimated 1,600,000 who attended the three largest marches in DC, New York, and Los Angeles, to ten brave
souls in the tiny town of Adak, Alaska, a village on a remote Aleutian island that has the distinction of
being the most remote municipality in the United States (Chenoweth and Pressman 2017). Figure 1 shows
the location and size of marches in the contiguous United States.
The marches had a carnivalesque atmosphere. While many participants expressed concerns over the potential
7

Figure 1: Women’s Marches Size and Location
dangers they perceived as coming from the new presidential administration, they described the marches
themselves as overwhelmingly positive and joyful (Gantt-Shafer, Wallis, and Miles 2019). Most protest signs
focused on themes of unity and many involved jokes and humorous illustrations (McClelland-Cohen and
Endacott 2020; Weber, Dejmanee, and Rhode 2018). While a majority of participants expressed a concern
for women’s rights as their primary motivation, almost half expressed other motivations (Fisher, Dow, and
Ray 2017). Disruption was also at a minimum. There were no significant reports of violence, and far from
imposing economic costs, the protests led to a surge in revenue for local businesses, particularly restaurants
(Filloon 2017). The march was also a major mobilizing moment for people who had never before participated
in protest. In Fisher, Dow, and Ray (2017)’s data collected on the day of the march, 34.7% of participants
reported having never attended a protest before, while 24.8% reported it was their first protest in five years.
As described in the introduction, anecdotally, the Women’s Marches played a significant role in jump-starting
the so-called anti-Trump “resistance.” They were the first in a series of semi-regular major protest marches
around various themes that took place during the first two years of the Trump presidency, including major
protests that took place at American airports in support of Muslim refugees after the issuing of the first
“Muslim ban,” the “March for Science,” the “People’s Climate March” and widespread marches focused on
healthcare during the summer of 2017 as the Republican-led congress attempted to repeal the Affordable Care
Act. The Women’s March itself became an annual tradition for several years, with widespread demonstrations
on the anniversary in 2018, 2019, and 2020.
8

The impact of the Women’s March on the subsequent anti-Trump Resistance is evidenced by the demograph-
ics of the Resistance. As Putnam and Skocpol (2018) identify, the resistance was a movement dominated
not by the traditional “activist class” of the young, highly-educated, and urban. Instead, the movement was
spearheaded by local groups of predominately middle-aged educated white women in the suburbs, similar to
the dominant demographics of the Women’s March identified by Fisher, Dow, and Ray (2017).
Unlike prior progressive movements such as “Occupy Wall Street,” the Resistance focused on electoral change,
particularly in the lead-up to the 2018 election. As Skocpol, Putnam, and Tervo (2020), 289 put it: “By
early 2018, the electoral turn was virtually universal.” Resistance groups pursued a range of strategies as
part of this electoral turn, from some continuing to hold themselves apart from partisan politics to others
who simply folded their organizing infrastructure into local Democratic parties. Yet nearly all shared a focus
on turning the organizing momentum of 2017 and 2018 into concrete political change.
Research on the Women’s March has both helped parse the motivations and identities of participants (Fisher,
Dow, and Ray 2017; Fisher, Jasny, and Dow 2018; Martin and Smith 2020), and identified several politically
consequential outcomes of the marches. Gordon (2021) describes how the “Emily’s List” organization used
the marches as a key framing device in their recruitment of progressive women to run for political office.
Einwohner and Rochford (2019) identifies how the momentum of the march was maintained through sustained
engagement on social media, and, in the closest parallel to this study, Larreboure and Gonzalez (2021) show
that Women’s Marches led to increased turnout in the 2018 election and increased vote shares by marginalized
groups, particularly women.
Given these characteristics, what would political impacts from the Women’s March tell us about the general
mechanisms of protest’s impact on politics? While the size of the marches certainly would represent a
powerful signal to elites, the lack of a specific issue focus and the presence of a complex set of diverse
motivations would undermine the clarity of that signal (Fassiotto and Soule 2017). Similarly, it is hard to
argue that any effects from the Women’s March arose through disruption, since the Marches had little to
no disruption. Instead, we should expect to see any effects operating through an activation mechanism,
due to the high levels of participation and the high proportion of participants who were new to engaging in
protest.
In the following section I link this observation to specific dependent variables.
9

4 Hypotheses
Building on the insights from prior research into the Women’s March and the Resistance, the first major
outcome I examine is the electoral consequences of Women’s Marches. As described above, while the Women’s
March itself was non-partisan, its progressive agenda and focus on opposition to the policies of the Trump
administration quickly led the movement coming out of the March to associate itself with the Democratic
party. Resistance groups were directly involved in recruiting Democratic candidates to run for office and
interfaced with local Democratic party infrastructures. Thus, I expect that the marches should lead to
increased turnout for Democratic candidates. Given the high level of geographic dispersion in the Women’s
March, and the leverage to be gained from exploring variation in march occurrence and size I also focus on
the local level, specifically the county. Formally stated, my first hypothesis is thus:
H1: County-level Women’s March size in 2017 will significantly increase county-level 2018 Democratic vote
share.
In addition to examining this direct political consequence of the Women’s March, I also seek to trace the
activation mechanism through which such an effect obtains. Effective protests should not simply remain
on the streets but instead turn into long-term organizing for the future. As described above, the Women’s
Marches in 2017 were followed by an outpouring of local-level grassroots organizing. I expect that such orga-
nizing is at least in part a direct consequence of the activation mechanism described above - socialization into
progressive ideas and connection to progressive social networks built through participation in the Women’s
March. The more people who participated in the local Women’s March, the larger the pool of potential
organizers activated to engage in local-level activism. Formally stated, my second hypothesis is thus:
H2: County-level Women’s March size in 2017 will significantly increase county-level Resistance organizing.
Another mechanism for potential political impact is through political donations. Fundraising is a critical part
of the US political process, and a major focus for all political campaigns. A massive Democratic fundraising
advantage was frequently pointed to as a key source of the Democratic 2018 “Blue Wave.” Democrats
received over $300 million more in campaign donations than Republicans in the 2018 election. While media
coverage of these donations frequently focuses on large donors, small donor contributions surged to over a
billion dollars in the 2018 election.6
6These numbers come from OpenSecrets. See “Most expensive midterm ever: Cost of 2018 election surpasses $5.7 billion”
https://www.opensecrets.org/news/2019/02/cost-of-2018-election- 5pnt7bil/
10

As with participation in local-level Resistance organizing, I expect donating to Democratic candidates to be
a natural consequence of participation in local Women’s Marches in 2017. As participants were socialized
into more progressive attitudes through their participation in the march, and joined more progressive social
networks through connecting at the march, I expect them to be more likely to donate to Democratic candi-
dates. Many local Resistance groups also focused on fundraising, in particular for Democratic candidates in
races previously considered to be non-competitive by the Democratic party. Formally stated, thus, I expect
the following:
H3: County-level Women’s March size in 2017 will significantly increase Democratic political donation share
in the 2018 election.
5 Research Design
5.1 Independent Variables
My primary independent variables are the occurrence and size of a Women’s march on January 21, 2017. My
data source for the occurrence of a march and the number of marchers is the Crowd Counting Consortium
(CCC) data (crowdcounting.org), which records 656 distinct marches with between one and 725,000 total
participants (best estimates), including eight marches with over 100,000 participants. The CCC data is
based on aggregating multiple sources, including media reports (which in turn primarily rely on police or
government estimates), social media, and activist self-reporting.7Sobolev et al. (2020) later verified the
accuracy of the CCC data by comparing its best estimate counts to cell-phone location data on the day
of the Women’s March, increasing confidence that the CCC estimates represent close to the ground truth
numbers of participants.
CCC records named locations for each Women’s March included in their dataset (typically the city or town).
I geo-coded these locations using the Google Maps API, then conducted a spatial merge of the geo-coded
locations with county boundaries from the US Census. If multiple marches took place in the same county
I combine participation in these marches to a single county-level participation sum. I then normalize the
variable with a natural log transformation, and also create a transformation of number of marchers per
capita. My population estimates (and thus my calculation of the per capita number of marchers) comes
from the US census.
7See description of the CCC data collection procedure in Fisher et al. (2019).
11

5.2 Dependent Variables
To test Hypothesis 1 on the electoral consequences of the Women’s March, I draw on county-level election
results from Dave Leip’s election atlas of the United States, a well-respected source of US elections data
employed by a many studies of US elections.8My dependent variable is the county-level Democratic vote
share. Since my dependent variable is continuous, I use linear regression as my modeling strategy.
To test Hypothesis 2 on future Resistance activity, my dependent variable is the creation and size of “In-
divisible” groups (Brooker 2018). Indivisible was started in December 2016 by a group of congressional
staffers interested in spreading effective strategies of political engagement for people opposed to the “Trump
agenda.” Their book Indivisible: A Practical Guide for Resisting the Trump Agenda encouraged concerned
citizens to create local groups that would pressure elected officials to resist the Trump agenda (Bethea 2016).
Indivisible groups often became one of the most prominent parts of the activist space directly devoted to
anti-Trump resistance (Gose and Skocpol 2019).
I selected Indivisible groups as my measure of Resistance activity for several reasons. First, I was interested
in the origins of new social movement activity, rather than mobilization through existing social movement
organizations, to more directly examine new activation into activism. Second, Indivisible is one of the few
national networks of grassroots organizations focused on the “Resistance agenda.” Third, summary data on
the geographic dispersion and size of groups was easily available for analysis. Skocpol, Putnam, and Tervo
(2020) use a similar approach, drawing on the distribution of Indivisible groups to map Resistance activity
in Pennsylvania.
My data on Indivisible comes from the listing of groups on the main Indivisible website.9Most Indivisible
groups were small, grassroots organizations founded in 2017 after the Women’s March. However, an im-
portant subset of groups on the Indivisible website do not fall into this category. Many existing left-wing
organizations sought to associate themselves with the Resistance through joining the Indivisible site, in-
cluding several wings of the Democratic party. The publication of the Indivisible guide also pre-dated the
Women’s March, and some of the first Indivisible groups were created before the March. Thus, to address
potential reverse causality, for each Indivisible group I investigated the public social media accounts listed on
the Indivisible website as associated with the group (typically a Facebook page) and recorded the founding
date. I then removed all groups with no information on founding date or a founding date before January
21, 2017. I also removed any groups that were explicitly listed as local branches of the Democratic party,
even if the founding date of their social media group was subsequent to the 2017 Women’s March. I then
8See, e.g. Ferrer, Geyn, and Thompson (2023), Goetz et al. (2019)
9www.indivisible.org
12

geographically matched Indivisible groups to counties and created two county-level variables: a binary indi-
cator of whether or not a county saw the creation of an Indivisible group between the Women’s March at
the 2018 election, and a count variable of the number of Indivisible groups created in the county between
the Women’s March and the 2018 election. My primary models are linear probability models of the binary
indicator of the creation of at least one Indivisible group in the county. I conduct tests on the variable
counting the number of groups in the appendix.
To test Hypothesis 3 on donations I look at the county-level Democratic share of political donations in
the 2018 election cycle. My source for donations data is the Database on Ideology, Money in Politics, and
Elections (DIME) dataset, version 3.1 (Bonica 2023). The full DIME dataset contains over 500 million
itemized political contributions across local, state, and federal elections from 1979 to 2022. I selected all
donations from the 2018 electoral cycle, grouped donations by county, and then took the percentage of
donations that went to Democratic candidates as my dependent variable. I use linear regression as my
modeling strategy.
5.3 Control Variables
I control for several plausible alternative explanations. In the Democratic vote share and Indivisible models I
control for the Democratic vote share in the prior mid-term elections in 2014, with data also coming from the
Dave Leip Election Atlas. In the donations models I control for the Democratic donation share in the 2014
mid-terms. Across all models I control for five demographic characteristics: the total population (logged),
a binary indicator of whether the county was urban, and the percentages of the population that identify
as white, Black, and Hispanic, based on census data. I also control for three economic indicators from the
census: median income (logged), the unemployment rate, and the poverty rate.
Finally, many of the larger Women’s Marches - particularly the central event in Washington DC - were not
purely local events but drew in participants from the surrounding areas. Thus it is crucial to control for the
spatial effects of being in a county close to one of these major Women’s Marches. Thus, for counties within
50 miles of one of the 30 largest marches (476 counties) I control for the logged number of participants in
the large nearby march. I set this variable to zero for counties more than 50 miles from one of the largest
marches.
13

5.4 Instrumental Variables Analysis
As described above, one of the key drawbacks in much of the existing work on protest effectiveness is that,
insofar as studies examining the impact of discrete protest events have been done, they are not able to fully
account for the endogeneity of protest to prior political conditions. Thus, any findings on effectiveness are
potentially spurious. Studies have typically attempted to address this through the use of control variables to
close off other observable explanations. However, this does not address potential unobservable factors that
influence both protest size and political outcomes.
To address this issue, in my analysis I use a two-stage least-squares model with instrumental variables.
Instrumental variables are the typical econometric technique for addressing the problem of endogenous
independent variables. An instrumental variable affects an endogenous independent variable, but is itself
exogenously assigned and only affects the dependent variable through its effect on the independent variable
(the “exclusion restriction”). Instrumental variables are common in economics, and increasingly common in
political science, but are infrequently employed in studies of protest or social movements.10
I use two instruments for local Women’s March size, both related to the weather on the day of the protest.
The first is average precipitation (in inches). The use of precipitation as an exogenous predictor of public
dissent, from protests and riots to organized violence, is well-established (Madestam et al. 2013; Ritter
and Conrad 2016; Wasow 2020). Heavy rain increases the personal discomfort and cost of protest, reducing
public participation. For example, Hong Kong’s pro-democracy “Umbrella Revolution” sit-in in 2014 was
significantly demobilized when its titular accessories failed to protect activists from sustained torrential
downpours (Wan 2014).11 My second instrument is the deviation from average temperature recorded on the
day of the Women’s Marches, similar to the strategy employed by Larreboure and Gonzalez (2021).
Some recent scholarship objects to using weather variables as exogenous instruments. In an extensive review
of the literature Mellon (2023) points out that weather consistently affects many variables that have wide-
ranging social and political impacts on factors including mood, pollution, and migration. More directly
relevant to this study, precipitation on an election day may significantly affect voter turnout and swing
elections (Gomez, Hansford, and Krause 2007; Shachar and Nalebuff 1999). Insofar as these factors may
subsequently affect a study’s dependent variable of interest, the exclusion restriction is violated and the
analysis may be invalid.
10A search on Google Scholar of the terms “protest” and “instrumental variable” returned only twelve articles from the
American Sociological Review, six articles from Social Forces, and two articles from Mobilization.
11Larreboure and Gonzalez (2021) find that a binary indicator of rainy weather (more than 0.1 inches of precipitation) fails
to predict Women’s March participation well, and thus they do not use precipitation as an instrument in their study. My
instrument is more precise in several ways: I use the absolute value of precipitation, rather than a binary cutoff; I control
for historical average levels of precipitation; and I use a more precise method of weighting weather station data using Voronoi
tessellation.
14

The potential exclusion restriction violations considered by Mellon (2023) pertain to either the effects of
average long-term weather patterns (such as weather’s effects on migration) or the short-term effects of
weather at one point in time on attitudes or behavior in its immediate aftermath. Neither of these categories
of violation are relevant for my instrument, since I use weather on a single day rather than an average trend,
and since my dependent variables of interest are measured months or years subsequent to the measurement
of the instrumental variable. A single day of cold, rainy or snowy weather does not plausibly affect voting
behavior two years later, or donations and activist group formation over several months, except insofar as it
affects critical events on the day of the Women’s March.
However, since levels of precipitation and temperature deviations on the day of the 2017 Women’s March may
proxy for general weather trends, I control for the average temperature and level of precipitation measured
in the week prior to and following January 21 for ten years leading up to 2017 in my main models and thus
seek to close off any potential indirect violations of the exclusion restriction.
My weather data comes from the National Oceanic and Atmospheric Administration (NOAA). NOAA collects
data on several weather-related indicators from its more than 19,000 weather stations across the United
States. I use NOAA’s data on weather stations’ location to create a grid of weather-station area polygons
using Voronoi tessellation (Voronoi 1908). The Voronoi tessellation algorithm draws polygons around each
weather station such that no point within a weather station’s polygon is closer to any other weather station.
To generate my county-level measures of precipitation and temperature deviation I sum the precipitation and
temperature reported from each weather station whose Voronoi polygon intersects the county in question,
and then average them, weighting the average by the percentage of the county’s area accounted for by each
Voronoi polygon.12
This provides a more accurate weather estimate than simply giving equal weight to the average precipitation
estimates from all weather stations inside a county or congressional district, as Madestam et al. (2013) do, or
simply matching each county with the individual weather station closest to it, as Larreboure and Gonzalez
(2021) do.
Table 1 shows summary statistics for the main variables.
12Thanks to [Colleague - blinded for peer review] for suggesting the use of the Voronoi tessellation algorithm to calculate
the geographic coverage of weather stations.
15

Table 1: Summary Statistics
Statistic N Mean St. Dev. Min Max
Marchers (log) 3,142 0.982 2.458 0.000 13.494
Marchers (per capita) 3,142 0.003 0.024 0.000 1.052
Precipitation (in) 3,133 0.147 0.339 0.000 3.424
Temp. Deviation 3,086 11.277 9.295 −26.012 33.344
Dem. Vote Share 2018 3,110 0.361 0.178 0.000 1.000
Dem. Donations Share 2018 3,139 0.265 0.187 0.000 1.000
Indivisible (binary) 3,142 0.208 0.406 0 1
Indivisible (Num groups) 3,142 0.452 1.868 0 65
Dem. Vote Share 2014 3,105 0.330 0.187 0.000 1.000
Dem Donations Share 2014 3,138 0.226 0.180 0.000 1.000
Percent White 3,142 0.847 0.164 0.039 0.993
Percent Black 3,142 0.101 0.146 0.001 0.864
Percent Hispanic 3,142 0.094 0.137 0.005 0.963
Total Population (log) 3,142 10.272 1.490 4.477 16.134
Unemployment Rate 3,141 4.619 1.676 1.600 20.100
Median Income (log) 3,141 10.810 0.244 10.029 11.822
Poverty Rate 3,141 15.378 6.277 3.000 56.700
Nearby Large Marchers (log) 3,142 1.667 3.972 0.000 13.494
Av. Historic Prcp. (in) 3,133 0.093 0.072 0.002 0.657
Av. Historic Temp. 3,091 42.986 12.952 −4.451 78.244
6 Findings
Table 2 presents results from the Democratic vote share models. Models 1 and 2 are naive models showing
the direct effect of logged marchers and marchers per capita on the 2018 Democratic vote share, with no
attempt to account for the endogenous elements of Women’s March size. There is a positive and significant
impact, however, this result is questionable because of the likely endogeneity problem.
Models 3 and 4 present the instrumental variable analysis, replicating Models 1 and 2 respectively but
instrumenting either the logged number of marchers or the marchers per capita with the variables for level
of precipitation and temperature deviations (First stage models available in Appendix Table 1). In both
cases not only does the Women’s March effect remain statistically significant, the coefficient increases in size,
nearly doubling for logged marchers and more than tripling for marchers per capita.
What is the magnitude of these effects in real-world terms? Figure 2 shows the marginal effects of instru-
mented Women’s March participation on 2018 Democratic vote share, with all continuous variables held
at the mean, and the urban-rural binary set to “Urban.” Increasing instrumented march size from 1,000
participants to 10,0000 participants changes the predicted Democratic vote share from 42.5% to 44.7%.
What might the 2018 election have looked like absent the Women’s March? To examine this question I
16

Table 2: 2018 Democratic Vote Share Models
Model 1 Model 2 Model 3 Model 4
Marchers (log) 0.006∗∗∗ 0.010∗∗
(0.001) (0.003)
Marchers PC 0.730∗∗∗ 3.018∗∗
(0.117) (1.136)
Dem. Vote Share 2014 0.660∗∗∗ 0.666∗∗∗ 0.578∗∗∗ 0.554∗∗∗
(0.010) (0.010) (0.014) (0.022)
Rural/Urban 0.004 0.005 0.013∗∗ 0.018∗∗∗
(0.004) (0.004) (0.004) (0.005)
Percent White −0.192∗∗∗
−0.202∗∗∗
−0.226∗∗∗
−0.239∗∗∗
(0.023) (0.023) (0.023) (0.024)
Percent Hispanic 0.055∗∗∗ 0.061∗∗∗ 0.188∗∗∗ 0.202∗∗∗
(0.012) (0.012) (0.015) (0.015)
Percent Black −0.039 −0.054∗0.054∗0.053
(0.023) (0.023) (0.026) (0.028)
Unemployment Rate 0.003∗∗ 0.003∗0.002 0.003
(0.001) (0.001) (0.001) (0.002)
Poverty Rate 0.000 0.000 0.000 −0.000
(0.001) (0.001) (0.001) (0.001)
Median Income (log) 0.069∗∗∗ 0.066∗∗∗ 0.036∗0.006
(0.016) (0.016) (0.015) (0.022)
Population (log) 0.016∗∗∗ 0.020∗∗∗ 0.014∗∗∗ 0.017∗∗∗
(0.002) (0.001) (0.003) (0.002)
Nearby Marchers (log) 0.001∗∗ 0.001∗∗ 0.001∗∗ 0.002∗∗
(0.000) (0.000) (0.001) (0.001)
Historic Av. Prcp. 0.130∗∗∗ 0.111∗∗
(0.031) (0.037)
Historic Av. Temp. −0.003∗∗∗
−0.003∗∗∗
(0.000) (0.000)
Constant −0.639∗∗∗
−0.629∗∗∗
−0.104 0.218
(0.181) (0.182) (0.182) (0.256)
Num. obs. 3105 3105 3050 3050
∗∗∗p < 0.001;∗∗ p < 0.01;∗p < 0.05
17

35.0%
37.5%
40.0%
42.5%
45.0%
47.5%
0 100 1000 10000 1e+05
Number of Women's March Participants
Predicted Democratic Vote Share
Predicted Democratic vote share from instrumented model.
Error bars are a 95% confidence interval.
Figure 2: Marginal Effects of Instrumented Participation on Democratic Vote Share
simulated the 2018 mid-term election, keeping all control variables at their actual values but setting the
Women’s March numbers in every county to zero and then generating predicted Democratic vote shares for
counties in which Women’s Marches took place based on Model 3 from 2. The reduction in Democratic
vote share nationwide in the absence of the Women’s March translates to a total reduction of nearly three
million Democratic votes. These absolute numbers should be interpreted with caution, as they rely on the
assumption that reduced Democratic vote share occurs with the same overall local-level turnout. However,
they provide an indicative sense of the magnitude of the Women’s March’s impact.
What explains the increased Democratic vote share? To examine this question, I turn now to my hypotheses
on movement-building and political donations. Table 3 presents four models on Indivisible group creation,
with a similar model set-up as in 2. I present linear probability models of the binary variable creation of any
Indivisible group. Results are substantively identical in logistic regression models (See Appendix Table 2).
As in Table 2, Models 1 and 2 are naive models without instrumenting Women’s March size, while Models 3
and 4 instrument Women’s March size with precipitation and temperature deviations. The size of Women’s
Marches in 2017 has a similar effect on Indivisible Group formation as it does on the 2018 Democratic vote
share. The effect is significant in both naive and instrumented models, but becomes significantly larger and
more significant in the models where march size is instrumented with precipitation and temperature.
18

Table 3: Indivisible Group Formation Models
Model 1 Model 2 Model 3 Model 4
Marchers (log) 0.032∗∗∗ 0.078∗∗∗
(0.003) (0.014)
Marchers PC 2.025∗∗∗ 26.322∗∗∗
(0.457) (6.130)
Dem. Vote Share 2014 0.197∗∗∗ 0.254∗∗∗ 0.063 −0.160
(0.037) (0.038) (0.059) (0.120)
Rural/Urban −0.018 −0.017 −0.010 0.033
(0.016) (0.017) (0.017) (0.026)
Percent White −0.225∗
−0.274∗∗
−0.164 −0.267∗
(0.089) (0.090) (0.098) (0.128)
Percent Hispanic 0.040 0.071 0.018 0.125
(0.046) (0.047) (0.065) (0.084)
Percent Black −0.350∗∗∗
−0.437∗∗∗
−0.213 −0.199
(0.089) (0.090) (0.110) (0.150)
Unemployment Rate −0.005 −0.008 −0.001 0.010
(0.005) (0.005) (0.006) (0.009)
Poverty Rate −0.004 −0.003 −0.005 −0.011∗∗
(0.002) (0.002) (0.003) (0.004)
Median Income (log) 0.062 0.067 0.010 −0.257∗
(0.061) (0.062) (0.065) (0.117)
Population (log) 0.103∗∗∗ 0.124∗∗∗ 0.067∗∗∗ 0.091∗∗∗
(0.006) (0.006) (0.012) (0.011)
Nearby Marchers (log) 0.004∗0.003 0.008∗∗∗ 0.016∗∗∗
(0.002) (0.002) (0.002) (0.004)
Historic Av. Prcp. 0.084 −0.102
(0.133) (0.199)
Historic Av. Temp. −0.000 −0.001
(0.001) (0.001)
Constant −1.320 −1.530∗
−0.447 2.435
(0.703) (0.715) (0.778) (1.381)
Num. obs. 3105 3105 3050 3050
∗∗∗p < 0.001;∗∗ p < 0.01;∗p < 0.05
19

Table 4: Democratic Donation Share Models
Model 1 Model 2 Model 3 Model 4
Marchers (log) 0.012∗∗∗ 0.034∗∗∗
(0.001) (0.006)
Marchers PC 0.367∗∗ 8.010∗∗∗
(0.121) (2.217)
Dem. Donat. Share 2014 0.495∗∗∗ 0.517∗∗∗ 0.428∗∗∗ 0.406∗∗∗
(0.016) (0.016) (0.020) (0.034)
Rural/Urban −0.007 −0.008 0.004 0.022
(0.007) (0.008) (0.008) (0.013)
Percent White −0.129∗∗∗
−0.167∗∗∗
−0.063 −0.146∗∗
(0.036) (0.036) (0.042) (0.056)
Percent Hispanic −0.062∗∗
−0.044∗
−0.016 0.018
(0.021) (0.022) (0.032) (0.043)
Percent Black −0.161∗∗∗
−0.202∗∗∗
−0.018 −0.166∗∗
(0.038) (0.038) (0.045) (0.062)
Unemployment Rate 0.005∗0.004 0.005∗0.003
(0.002) (0.002) (0.002) (0.003)
Poverty Rate −0.002 −0.001 −0.001 −0.005∗
(0.001) (0.001) (0.001) (0.002)
Median Income (log) 0.016 0.025 −0.042 −0.210∗∗
(0.028) (0.028) (0.030) (0.071)
Population (log) −0.003 0.006∗
−0.019∗∗∗
−0.005
(0.003) (0.003) (0.005) (0.005)
Nearby Marchers (log) 0.004∗∗∗ 0.003∗∗∗ 0.006∗∗∗ 0.009∗∗∗
(0.001) (0.001) (0.001) (0.002)
Historic Av. Prcp. −0.041 −0.002
(0.058) (0.083)
Historic Av. Temp. −0.002∗∗∗
−0.002∗∗∗
(0.000) (0.001)
Constant 0.129 −0.017 0.925∗∗ 2.741∗∗∗
(0.321) (0.326) (0.353) (0.821)
Num. obs. 3137 3137 3077 3077
∗∗∗p < 0.001;∗∗ p < 0.01;∗p < 0.05
Table 4 turns from movement-building to political donations, looking at the country-level Democratic dona-
tions share as an outcome of Women’s March size. Here too we see a similar pattern: highly significant but
relatively small effects from naive models but much larger highly significant effects in instrumental variable
models.
7 Discussion and Conclusion
In this article, I have examined whether and how the 2017 Women’s March, possibly the largest single day
of protest in US history, affected local-level politics in the lead-up to the 2018 mid-term election. I have
exploited the single temporal “shock” of the marches and their significant geographic dispersion to get at
20

local-level effects. Further, I have isolated the effects of the marches themselves, addressing endogeneity
of protest to current political conditions through an instrumental variable analysis using precipitation and
temperature deviations as exogenous predictors of protest participation.
The analysis provides strong and consistent evidence that the Women’s March significantly increased county-
level Democratic vote share in the 2018 election, to the tune of potentially millions of additional votes.
Analysis looking at Indivisible group formation and donations provide evidence of how this electoral effect
obtained. Larger Women’s Marches were much more likely to result in the formation of local Indivisible
groups. These groups in turn, as is well-established in by scholars such as Gose and Skocpol (2019) and
Fisher (2019), played a central role in the local-level organizing around the 2018 election. Larger marches
also led directly to increased donations for Democrats, helping to fuel the electoral machine. While such
a relationship has been qualitatively examined by scholars such as Gose and Skocpol (2019) or Fisher et
al. (2019), this study provides robust and rigorous evidence that it was at least in part the size of the
Women’s Marches themselves that led to follow-up action and ultimately to improved electoral prospects for
Democratic candidates.
In addition to providing insight into recent American political history, these results also provide insights
into the mechanisms through which protest affects political outcomes more generally. In particular they
highlight the sufficiency of an activation mechanism, whereby large demonstrations with little disruption
or confrontation function as socializing and recruiting opportunities through which new constituencies are
created and activated to advance a movement’s goals. Even if a protest’s signal to elites is clouded by a
broad, big-tent agenda, and its level of physical disruption is minimal, protest can still “work.”
However, the research also comes with many limitations. First, just as the Women’s March was an outlier in
size, it is possible it is also an outlier in its effects. Similar outcomes may or may not obtain in protests that
were smaller, or not part of a coordinated national moment of protest. Second, measuring protest size is
always a tricky empirical question, and comes with many issues of selection and reporting bias. The Crowd
Counting Consortium data, on which my results are based, helps to alleviate some of these issues through
triangulating multiple media and online sources, and its protest sizes are supported by cell phone location
data from the day of the march (Sobolev et al. 2020). However, even cell phone data does not truly get us
to the absolute numbers, and may be subject to biases. Third, the lack of disruption at the Women’s March
means that I am unable to say anything definitive about how more disruptive protest may or may not have
impacted the effects observed here.
Both the findings and their limitations open several avenues for further research. First, the time horizons
of this study are relatively brief, focusing on outcomes over a single electoral cycle. To what degree do such
21

outcomes endure? Mazumder (2018) shows that civil rights protests in the United States continued to impact
attitudes on race many decades after the end of the civil rights movement. Might we observe similar changes
in attitudes or other enduring political outcomes from the Women’s March and the subsequent “Resistance”?
Second, this research focuses on organizing and electoral outcomes, but has not examined policy change, as
is the case in much of the broader literature on the effects of protest. Third, the sole source of variation I
have examined in this paper is size, since this is the primary way in which local Women’s Marches varied.
Yet future research could examine the degree to which other characteristics of the protests, such as whether
they had local organizational support, and whether specific follow-up actions took place after the March
itself, affected their impact.
For activists and social movement leaders, this research speaks to the importance of public demonstrations,
and in particular demonstrations such as the Women’s March, as a key mobilizing moment for initiating
future activism. The 2017 Women’s Marches were dominated by positive, activating messages, very low levels
of physical risk (including no reports of violence that I was able to uncover), and a welcoming environment
for almost all participants. Cross-national research indicates that events like the Women’s March that create
an enjoyable, welcoming environment, are associated with greater movement success (Gledhill, Duursma,
and Shay 2022). Yet many organizers and some scholars dismiss the impact of these kinds of events because
they do not directly disrupt opponents. This research, while it cannot speak to the efficacy of more disruptive
protest, does at least indicate that big, fun, public demonstrations do have an important mobilizing function
to play in the larger repertoire of social movement tactics.
The research certainly does not imply that large peaceful public demonstrations are a panacea, and that
movements should invest as much as possible in big days of protest at the expense of other forms of organizing.
Indeed, Chenoweth (2020) identifies such an over-emphasis on public demonstrations as one factor that may
be behind a decline in the effectiveness of movements around the world. One key impact of the Women’s
March was that its participants went on to do the hard work of local political organizing. But for movements
that are seeking to jumpstart their organizing in a difficult environment, and spark a larger movement for
change, a protest that feels more like a celebration may be the best way forward.
8 References
Alinsky, Saul. 1989. Rules for Radicals: A Pragmatic Primer for Realistic Radicals. New York, NY: Vintage.
Amenta, Edwin, Kenneth T. Andrews, and Neal Caren. 2018. “The Political Institutions, Processes, and
Outcomes Movements Seek to Influence.” In The Wiley Blackwell Companion to Social Movements,
22

edited by David A. Snow, Sarah A. Soule, Hanspeter Kriesi, and Holly J. McCammon, 1st ed., 447–65.
Wiley. https://doi.org/10.1002/9781119168577.ch25.
Amenta, Edwin, Neal Caren, Elizabeth Chiarello, and Yang Su. 2010. “The Political Consequences of Social
Movements.” Annual Review of Sociology 36 (1): 287–307. https://doi.org/10.1146/annurev-soc-070308-
120029.
Andrews, Kenneth T. 1997. “The Impacts of Social Movements on the Political Process: The Civil Rights
Movement and Black Electoral Politics in Mississippi.” American Sociological Review, 800–819. https:
//www.jstor.org/stable/2657361.
Andrews, Kenneth T., Neal Caren, and Todd Lu. 2020. “Racial, Ethnic, and Immigration Protest During
Year One of the Trump Presidency.” In Racialized Protest and the State: Resistance and Repression in
a Divided America, edited by Hank Johnston and Pamela Oliver, 208–26. London, UK: Routledge.
Baumgartner, Frank R., and Christine Mahoney. 2005. “Social Movements, the Rise of New Issues, and the
Public Agenda.” In Routing the Opposition: Social Movements, Public Policy, and Democracy, edited by
David S. Meyer, Valerie Jenness, and Helen M. Ingram, 65–86. Minneapolis, MN: University of Minnesota
Press.
Bethea, Charles. 2016. “The Crowdsourced Guide to Fighting Trump’s Agenda,” December.
Biggs, Michael, and Kenneth T. Andrews. 2015. “Protest Campaigns and Movement Success: Desegregating
the U.S. South in the Early 1960s.” American Sociological Review 80 (2): 416–43. https://doi.org/10.
1177/0003122415574328.
Bonica, Adam. 2023. “Database on Ideology, Money in Politics, and Elections: Public Version 3.1.” Stanford,
CA: Stanford University Libraries.
Bosi, Lorenzo, and Katrin Uba. 2021. “Collective Action Outcomes: Ways Forward for the Subfield.”
PArticipation and COnflict 14 (3): 987–97.
Branton, Regina, Valerie Martinez-Ebers, Tony E. Carey, and Tetsuya Matsubayashi. 2015. “Social Protest
and Policy Attitudes: The Case of the 2006 Immigrant Rallies.” American Journal of Political Science
59 (2): 390–402. https://doi.org/10.1111/ajps.12159.
Bremer, Björn, Swen Hutter, and Hanspeter Kriesi. 2020. “Dynamics of Protest and Electoral Politics in the
Great Recession.” European Journal of Political Research 59 (4): 842–66. https://doi.org/10.1111/1475-
6765.12375.
Brooker, Megan E. 2018. “Indivisible: Invigorating and Redirecting the Grassroots.” In The Resistance: The
Dawn of the Anti-Trump Opposition Movement, edited by David S. Meyer and Sidney Tarrow, 162–84.
New York, NY: Oxford University Press.
Brooks, David. 2017. “Opinion | After the Women’s March.” The New York Times, January.
23

Burstein, Paul, and April Linton. 2002. “The Impact of Political Parties, Interest Groups, and Social
Movement Organizations on Public Policy: Some Recent Evidence and Theoretical Concerns.” Social
Forces 81 (2): 380–408. https://doi.org/10.1353/sof.2003.0004.
Burstein, Paul, and Sarah Sausner. 2005. “The Incidence and Impact of Policy-Oriented Collective Action:
Competing Views.” Sociological Forum 20 (3): 403–19. https://doi.org/10.1007/s11206-005-6595-4.
Butcher, Charles, and Jonathan Pinckney. 2022. “Friday on My Mind: Re-Assessing the Impact of Protest
Size on Government Concessions.” Journal of Conflict Resolution 66 (7-8): 1320–55. https://doi.org/10.
1177/00220027221099887.
Caren, Neal, Kenneth T. Andrews, and Micah H. Nelson. 2023. “Black Lives Matter Protests and the 2020
Presidential Election.” Social Movement Studies OnelineFirst.
Chenoweth, Erica. 2020. “The Future of Nonviolent Resistance.” Journal of Democracy 31 (3): 69–84.
https://doi.org/10.1353/jod.2020.0046.
Chenoweth, Erica, and Margherita Belgioioso. 2019. “The Physics of Dissent and the Effects of Movement
Momentum.” Nature Human Behaviour 3 (10): 1088–95.
Chenoweth, Erica, and Jeremy Pressman. 2017. “This Is What We Learned by Counting the Women’s
Marches.” The Washington Post, February.
Chenoweth, Erica, and Maria J. Stephan. 2011. Why Civil Resistance Works: The Strategic Logic of
Nonviolent Conflict. New York, NY: Columbia University Press.
Collingwood, Loren, Nazita Lajevardi, and Kassra A. R. Oskooii. 2018. “A Change of Heart? Why
Individual-Level Public Opinion Shifted Against Trump’s ‘Muslim Ban’.” Political Behavior 40 (4):
1035–72. https://doi.org/10.1007/s11109-017-9439-z.
Corrigall-Brown, Catherine. 2021. “Indivisible Against Trump: Coalition Strategies and Movement Success
Across City Contexts.” Mobilization: An International Quarterly 26 (2): 157–78. https://doi.org/10.
17813/1086-671X-26-2-157.
DeNardo, James. 1985. Power in Numbers: The Political Strategy of Protest and Rebellion. Princeton, NJ:
Princeton University Press.
Einwohner, Rachel L., and Elle Rochford. 2019. “After the March: Using Instagram to Perform and Sustain
the Women’s March.” Sociological Forum 34 (S1): 1090–1111. https://doi.org/10.1111/socf.12542.
Enos, Ryan D., Aaron R. Kaufman, and Melissa L. Sands. 2019. “Can Violent Protest Change Local Policy
Support? Evidence from the Aftermath of the 1992 Los Angeles Riot.” American Political Science Review
113 (4): 1012–28. https://doi.org/10.1017/S0003055419000340.
Fassiotto, Magali, and Sarah A. Soule. 2017. “Loud and Clear: The Effect of Protest Signals on Congressional
Attention.” Mobilization 22 (1): 17–38.
24

Ferrer, Joshua, Igor Geyn, and Daniel M. Thompson. 2023. “How Partisan Is Local Election Administra-
tion?” American Political Science Review, July, 1–16. https://doi.org/10.1017/S0003055423000631.
Filloon, Whitney. 2017. “The Women’s March Drove Business to Restaurants Across the Country.” Eater.
https://www.eater.com/2017/1/26/14397086/womens-march-restaurants-business.
Fisher, Dana R. 2019. American Resistance: From the Women’s March to the Blue Wave. New York, NY:
Columbia University Press.
Fisher, Dana R., Kenneth T. Andrews, Neal Caren, Erica Chenoweth, Michael T. Heaney, Tommy Leung,
L. Nathan Perkins, and Jeremy Pressman. 2019. “The Science of Contemporary Street Protest: New
Efforts in the United States.” Science Advances 5 (10).
Fisher, Dana R., Dawn M. Dow, and Rashawn Ray. 2017. “Intersectionality Takes It to the Streets:
Mobilizing Across Diverse Interests for the Women’s March.” Science Advances 3 (9): eaao1390. https:
//doi.org/10.1126/sciadv.aao1390.
Fisher, Dana R., Lorien Jasny, and Dawn M. Dow. 2018. “Why Are We Here? Patterns of Intersectional
Motivations Across the Resistance.” Mobilization: An International Quarterly 23 (4): 451–68.
Gamson, William A. 1975. The Strategy of Social Protest. Belmont, CA: Wadsworth.
Gantt-Shafer, Jessica, Cara Wallis, and Caitlin Miles. 2019. “Intersectionality, (Dis)unity, and Processes
of Becoming at the 2017 Women’s March.” Women’s Studies in Communication 42 (2): 221–40. https:
//doi.org/10.1080/07491409.2019.1616021.
Gause, LaGina. 2020. “Revealing Issue Salience via Costly Protest: How Legislative Behavior Following
Protest Advantages Low-Resource Groups.” British Journal of Political Science, 1–21. https://doi.org/
10.1017/S0007123420000423.
Gillion, Daniel Q. 2012. “Protest and Congressional Behavior: Assessing Racial and Ethnic Minority Protests
in the District.” The Journal of Politics 74 (4): 950–62. https://doi.org/10.1017/S0022381612000539.
Gillion, Daniel Q., and Sarah A. Soule. 2018. “The Impact of Protest on Elections in the United States.”
Social Science Quarterly 99 (5): 1649–64.
Giugni, Marco. 2007. “Useless Protest? A Time-Series Analysis of the Policy Outcomes of Ecology, Antinu-
clear, and Peace Movements in the United States, 1977-1995.” Mobilization: An International Quarterly
12 (1): 53–77.
Gledhill, John, Allard Duursma, and Christopher Shay. 2022. “Glee and Grievance: Emotive Events and
Campaign Size in Nonviolent Resistance.” Journal of Global Security Studies 7 (4): 1–18.
Goetz, Stephan J., Meri Davlasheridze, Yicheol Han, and David A. Fleming-Muñoz. 2019. “Explaining the
2016 Vote for President Trump Across U.S. Counties.” Applied Economic Perspectives and Policy 41 (4):
703–22. https://doi.org/10.1093/aepp/ppy026.
25

Gomez, Brad T., Thomas G. Hansford, and George A. Krause. 2007. “The Republicans Should Pray for
Rain: Weather, Turnout, and Voting in US Presidential Elections.” The Journal of Politics 69 (3):
649–63.
Gordon, Katie M. 2021. “Candidate Emergence as Movement Mobilization: An Analysis of Women’s Post-
2016 Electoral Engagement.” Sociology Compass 15 (6): e12878. https://doi.org/10.1111/soc4.12878.
Gose, Leah E., and Theda Skocpol. 2019. “Resist, Persist, and Transform: The Emergence and Impact of
Grassroots Resistance Groups Opposing the Trump Presidency.” Mobilization: An International Quar-
terly 24 (3): 293–317. https://doi.org/10.17813/1086-671X-24-3-293.
Han, Hahrie, Elizabeth McKenna, and Michelle Oyakawa. 2021. Prisms of the People: Power and Organizing
in Twenty-First-Century America. Chicago, IL: University of Chicago Press.
Htun, Mala, and S. Laurel Weldon. 2012. “The Civic Origins of Progressive Policy Change: Combating
Violence Against Women in Global Perspective, 1975–2005.” American Political Science Review 106 (3):
548–69. https://doi.org/10.1017/S0003055412000226.
Jasper, James M. 2018. The Emotions of Protest. Chicago, IL: The University of Chicago Press.
Johnson, Erik W. 2008. “Social Movement Size, Organizational Diversity and the Making of Federal Law.”
Social Forces 86 (3): 967–93.
Kadivar, Mohammad Ali, and Neil Ketchley. 2018. “Sticks, Stones, and Molotov Cocktails: Unarmed
Collective Violence and Democratization.” Socius 4: 2378023118773614.
King, Brayden G, and Sarah A. Soule. 2007. “Social Movements as Extra-Institutional Entrepreneurs: The
Effect of Protests on Stock Price Returns.” Administrative Science Quarterly 52 (3): 413–42. https:
//doi.org/10.2189/asqu.52.3.413.
Kitschelt, Herbert P. 1986. “Political Opportunity Structures and Political Protest: Anti-Nuclear Movements
in Four Democracies.” British Journal of Political Science 16 (1): 57–85. https://www.jstor.org/stable/
193981.
Larreboure, Magdalena, and Felipe Gonzalez. 2021. “The Impact of the Women’s March on the U.S. House
Election.” Working Paper.
Lohmann, Susanne. 1993. “A Signaling Model of Informative and Manipulative Political Action.” The
American Political Science Review 87 (2): 319–33. https://doi.org/10.2307/2939043.
Madestam, Andreas, Daniel Shoag, Stan Veuger, and David Yanagizawa-Drott. 2013. “Do Political Protests
Matter? Evidence from the Tea Party Movement.” Quarterly Journal of Economics 128 (4): 1633–85.
Martin, Jennifer L., and Julia Smith. 2020. “Why We March! Feminist Activism in Critical Times: Lessons
from the Women’s March on Washington.” Women’s Studies International Forum 81 (July): 102375.
https://doi.org/10.1016/j.wsif.2020.102375.
26

Mazumder, Soumyajit. 2018. “The Persistent Effect of U.S. Civil Rights Protests on Political Attitudes.”
American Journal of Political Science 62 (4): 922–35.
McAdam, Doug. 1982. Political Process and the Development of Black Insurgency, 1930-1970. Chicago, IL:
University of Chicago Press.
McAdam, Doug, and Yang Su. 2002. “The War at Home: Antiwar Protests and Congressional Voting, 1965
to 1973.” American Sociological Review 67 (5): 696–721.
McAdam, Doug, and Sidney Tarrow. 2010. “Ballots and Barricades: On the Reciprocal Relationship
Between Elections and Social Movements.” Perspectives on Politics 8 (2): 529–42.
McClelland-Cohen, Avigail, and Camille G. Endacott. 2020. “The Signs of Our Discontent: Framing
Collective Identity at the Women’s March on Washington.” Communication Studies 71 (5): 842–56.
https://doi.org/10.1080/10510974.2020.1784246.
Mellon, Jonathan. 2023. “Rain, Rain, Go Away: 195 Potential Exclusion-Restriction Violations for Studies
Using Weather as an Instrumental Variable.” Available at SSRN 3715610.
Meyer, D. S., and D. C. Minkoff. 2004. “Conceptualizing Political Opportunity.” Social Forces 82 (4):
1457–92. https://doi.org/10.1353/sof.2004.0082.
Meyer, David S., and Sidney Tarrow. 2018. “Introduction.” In The Resistance: The Dawn of the Anti-
Trump Opposition Movement, edited by David S. Meyer and Sidney Tarrow, 1–23. New York, NY:
Oxford University Press.
Mueller, Lisa. 2024. “Crowd Cohesion and Protest Outcomes.” American Journal of Political Science 68
(1): 42–57. https://doi.org/10.1111/ajps.12725.
Olzak, Susan. 2021. “Does Protest Against Police Violence Matter? Evidence from U.S. Cities, 1990 Through
2019.” American Sociological Review 86 (6): 1066–99. https://doi.org/10.1177/00031224211056966.
Olzak, Susan, and Sarah A. Soule. 2009. “Cross-Cutting Influences of Environmental Protest and Legisla-
tion.” Social Forces 88 (1): 201–25.
Opp, Karl-Dieter, and Bernhard Kittel. 2010. “The Dynamics of Political Protest: Feedback Effects and
Interdependence in the Explanation of Protest Participation.” European Sociological Review 26 (1): 97–
109. https://doi.org/10.1093/esr/jcp008.
Piven, Frances Fox, and Richard A. Cloward. 1977. Poor People’s Movements: Why They Succeed, How
They Fail. New York: Pantheon Books.
Putnam, Lara, and Theda Skocpol. 2018. “Middle America Reboots Democracy.” Democracy Journal.
https://democracyjournal.org/arguments/middle-america-reboots-democracy/.
Reny, Tyler T., and Benjamin J. Newman. 2021. “The Opinion-Mobilizing Effect of Social Protest Against
Police Violence: Evidence from the 2020 George Floyd Protests.” American Political Science Review 115
27

(4): 1499–1507.
Ritter, Emily Hencken, and Courtenay R. Conrad. 2016. “Preventing and Responding to Dissent: The
Observational Challenges of Explaining Strategic Repression.” American Political Science Review 110
(1): 85–99. https://doi.org/10.1017/S0003055415000623.
Sato, Yuko, and Jake Haselswerdt. 2022. “Protest and State Policy Agendas: Marches and Gun Policy After
Parkland.” Policy Studies Journal 50 (4): 877–95. https://doi.org/10.1111/psj.12463.
Shachar, Ron, and Barry Nalebuff. 1999. “Follow the Leader: Theory and Evidence on Political Participa-
tion.” American Economic Review 89 (3): 525–47.
Skocpol, Theda. 2003. Diminished Democracy: From Membership to Management in American Civic Life.
Norman, OK: University of Oklahoma press.
Skocpol, Theda, Lara Putnam, and Caroline Tervo. 2020. “Citizen Activism and the Democratic Party.”
In Upending American Politics: Polarizing Parties, Ideological Elites, and Citizen Activists from the Tea
Party to the Anti-Trump Resistance, edited by Theda Skocpol and Caroline Tervo, 0. Oxford University
Press. https://doi.org/10.1093/oso/9780190083526.003.0013.
Sobolev, Anton, M. Keith Chen, Jungseock Joo, and Zachary C. Steinert-Threlkeld. 2020. “News and
Geolocated Social Media Accurately Measure Protest Size Variation.” American Political Science Review
114 (4): 1343–51.
Soule, Sarah A., and Susan Olzak. 2004. “When Do Movements Matter? The Politics of Contingency and
the Equal Rights Amendment.” American Sociological Review 69 (4): 473–97. https://doi.org/10.1177/
000312240406900401.
Soule, Sarah, Doug McAdam, John McCarthy, and Yang Su. 1999. “Protest Events: Cause or Conse-
quence of State Action? The US Women’s Movement and Federal Congressional Activities, 1956-1979.”
Mobilization: An International Quarterly 4 (2): 239–56.
Stein, Elizabeth A. 2017. “Are ICTs Democratizing Dictatorships? New Media and Mass Mobilization.”
Social Science Quarterly 98 (3): 914–41.
Uba, Katrin. 2009. “The Contextual Dependence of Movement Outcomes: A Simplified Meta-Analysis.”
Mobilization: An International Quarterly 14 (4): 433–48. https://doi.org/10.17813/maiq.14.4.
m8477j873p47p546.
Voronoi, Georges. 1908. “Nouvelles Applications Des Paramètres Continus à La Théorie Des Formes Quadra-
tiques. Deuxième mémoire. Recherches Sur Les Parallélloèdres Primitifs.” Journal für Die Reine Und
Angewandte Mathematik 134: 198–287.
Walgrave, Stefaan, and Rens Vliegenthart. 2012. “The Complex Agenda-Setting Power of Protest: Demon-
strations, Media, Parliament, Government, and Legislation in Belgium, 1993-2000.” Mobilization: An
28

International Quarterly 17 (2): 129–56.
Wan, William. 2014. “Exhaustion Saps Fervor of Hong Kong Protests.” The Washington Post, October,
A13.
Wasow, Omar. 2020. “Agenda Seeding: How 1960s Black Protests Moved Elites, Public Opinion, and
Voting.” American Political Science Review 114 (3): 638–59.
Weber, Kirsten M., Tisha Dejmanee, and Flemming Rhode. 2018. “The 2017 Women’s March on Washing-
ton: An Analysis of Protest-Sign Messages.” International Journal of Communication 12: 2289–2313.
Weldon, Laurel. 2022. When Protest Makes Policy: How Social Movements Represent Disadvantaged
Groups. University of Michigan Press. https://doi.org/10.3998/mpub.1285595.
Wouters, Ruud. 2019. “The Persuasive Power of Protest. How Protest Wins Public Support.” Social Forces
98 (1): 403–26. https://doi.org/10.1093/sf/soy110.
Wouters, Ruud, and Stefaan Walgrave. 2017. “Demonstrating Power: How Protest Persuades Political
Representatives.” American Sociological Review 82 (2): 361–83.
29