Straying from the Flock? A Look at How Americans’ Gender and Religious Identities Cross-Pressure Partisanship

• Published date: 01 March 2020
• Date: 01 March 2020
• DOI: 10.1177/1065912919889681
• Author: Erin C. Cassese
• Subject Matter: Articles

889681PRQXXX10.1177/1065912919889681Political Research QuarterlyCassese
research-article2019
Article
Political Research Quarterly
2020, Vol. 73(1) 169 –183
Straying from the Flock? A Look at
© 2019 University of Utah
Article reuse guidelines:
sagepub.com/journals-permissions
How Americans’ Gender and Religious
https://doi.org/10.1177/1065912919889681
DOI: 10.1177/1065912919889681
journals.sagepub.com/home/prq
Identities Cross-Pressure Partisanship
Erin C. Cassese1
Abstract
White evangelicals–both men and women–are a mainstay of the Republican Party. What accounts for their ongoing
loyalty, particularly when Republican candidates and leaders fail to embody closely held moral standards around
sexual monogamy and propriety, as Donald Trump did in 2016? To answer this question, I draw on research about
social sorting and polarization, as well as gender and religion gaps in public opinion, to theorize about the nature of
the cross-pressures partisans may experience as a result of the religious and gender identifications they hold. Using
data from the 2016 American National Election Study, I evaluate whether cross-cutting identities have a moderating
effect on partisans’ thinking about gender issues, their evaluations of the presidential candidates, and their relationship
to the parties. I find only modest evidence that gender and evangelical identification impact political thinking among
white Republicans, including their reactions to the Access Hollywood tape. Other groups, however, experienced
more significant cross-pressures in 2016. Both evangelical Democrats and secular Republicans reported less polarized
affective reactions to the presidential candidates and the parties. The results highlight the contingent role that gender
and religious identities play in the United States’ highly polarized political climate.
Keywords
evangelical protestants, partisanship, sorting, polarization, hostile sexism, 2016 U.S. presidential election, gender gap,
white women voters
Introduction
It turned out that white women proved to be important
base of support for Donald Trump. This was particularly
During the 2016 presidential race, Donald Trump was
true for white evangelical women, 79 percent of whom
plagued by speculation about his deficit with women vot-
voted for Trump, compared with 80 percent of white
ers. Polling data collected during the primaries showed
evangelical men.1 Misapprehensions about women voters
low favorability ratings for Trump among Republican
stemmed from the assumption that all women would
women compared with Republican men (Bump 2016),
respond to gender issues in campaigns and elections in a
with approximately 70 percent of all women holding an
similar fashion—that an underlying gender identity can
unfavorable view of his candidacy (Newport and Saad
unite women voters or induce them to respond similarly
2016). Donald Trump regained some ground during the
to particular kinds of campaign events. Research shows
general election, but the release of the Access Hollywood
that women are not a unified voting block (Junn 2017)—
tape in October of 2016, and the spate of sexual assault
and that women do not typically cross party lines to vote
allegations that followed, escalated the crisis narrative
for female candidates (Dolan 2014)—but these relation-
around his inability to appeal to women voters. Trump’s
ships had never been tested against a campaign like this
poor performance with women in the polls, coupled with
one. It seemed plausible that aspects of Trump’s candi-
campaign dynamics that were widely interpreted as sexist
dacy pushed against the limits of party loyalty, but the
(e.g., Keith 2016), contributed to the expectation that
outcome suggests otherwise. This election highlights the
women voters would respond to all of this in essentially
the same way—by not voting for him. This same expecta-
1University of Delaware, Newark, USA
tion applied to white evangelical women, who could gen-
erally be counted on to vote Republican, because these
Corresponding Author:
campaign events highlighted Trump’s deviation from firm
Erin C. Cassese, Department of Political Science & International
Relations, University of Delaware, 465 Smith Hall, 18 Amstel Ave.,
evangelical standards concerning sexual morality and the
Newark, DE 19716, USA.
sanctity of marriage (e.g., Nancy 2016; Turner 2017).
Email: ecassese@udel.edu

---

170
Political Research Quarterly 73(1)
need to better understand the conditions under which
things being equal, conservatives and Republicans tend
gender and religious identifications might cross-pressure
to attend church because they see this practice as confir-
partisan identity.
mation of ideology and partisanship and as demarcation
In this paper, I use the 2016 American National
from the out-group (liberals and Democrats).” This point
Election Study (ANES) to evaluate how gender, religious
highlights how the social sorting process aligns and rein-
identity, and partisanship jointly shape white Americans’
forces connections between social and political identi-
political thinking.2 I look for evidence of social cross-
ties, contributing to polarization.
pressures in gender attitudes—in terms of levels of hos-
Social sorting is incomplete, however, and significant
tile sexism, expressions of politicized gender identity,
portions of the electorate do not fit neatly into these sort-
and reactions to the Access Hollywood tape. I also con-
ing profiles. For example, about one in five women iden-
sider the implications of these cross-pressures for party
tify with the Republican Party. A similar proportion of
polarization and evaluate group differences in percep-
white evangelicals identifies with the Democratic Party.
tions of the presidential candidates, the parties them-
Generally, social identities that do not align with parti-
selves, and strength of partisan attachment. The results
sanship, sometimes referred to as cross-cutting identi-
suggest the effects of cross-cutting identities are actually
ties, have moderating effects on political thinking and
quite contingent, and the relative strength of cross-cutting
behavior (Mason 2015; 2017). People who are not
influences is variable. I find little evidence that evangeli-
socially sorted tend to have more contact with members
cal Republican women are cross-pressured by their gen-
of the opposing party, and thus face conflicting social
der identification. They are no less committed to the
pressures, or “cross-pressures.” These cross-pressures
Republican Party and no more critical of Donald Trump
typically produce weaker partisan attachments, muted
than other Republican voters in 2016. Instead, secular
affective responses to the parties, and moderate political
Republican women and evangelical Democrats experi-
positions. For example, Barnes and Cassese (2017)
ence some cross-pressures that coincide with elevated
found that Republican women hold more moderate pol-
rates of cross-party voting. The results contribute to a
icy attitudes than Republican men across a number of
growing body of literature on heterogeneity among
issues areas, though their positions are still quite distant
women voters, particularly the distinctive electoral from Democratic women.
behavior of white women voters, and also further our
The effects of cross-cutting identities collectively
understanding of social sorting and party polarization.
function to reduce polarization and to facilitate compro-
mise (Mason 2018) and even cross-party voting (Davis
Social Sorting and Cross-Cutting
and Mason 2016). The more cross-cutting identities a
person has, the more pronounced the moderating effect.
Identities
Expectations about women voters in 2016 reflected these
Research on social sorting suggests that the way identi-
ideas about how cross-cutting identities work—that
ties align or fail to align can have significant implica-
Republican women would be cross-pressured by their
tions for public opinion and political behavior. Sorting is
gender and move to the left politically, given the salience
a process by which the American political parties are
of gender issues in the campaign.
growing increasingly homogeneous (Abramowitz 2010;
Levendusky 2009; Mason 2018). Social identities linked
Gender as a Cross-Cutting Identity
to factors like race, ethnicity, religious tradition, gender,
and class are becoming aligned with partisanship, mean-
Most of the gender gap literature has focused on average
ing that the overlap in membership between these social
differences between men and women, which has contrib-
and political categories is on the rise. For example,
uted to the assumption that women are uniformly more
long-term processes of social sorting have resulted in a
liberal than men (Barnes and Cassese 2017). But scholar-
gender gap—whereby women are more likely to be
ship on gender identity shows that women vary a great
aligned with the Democratic Party—and a God gap—
deal in terms of whether they think about gender in a polit-
whereby religious Americans are more likely to be
icized fashion. Women who identify with the Democratic
aligned with the Republican Party and secular Americans
Party are also more likely to identify as feminists, though
with the Democratic Party (Claasen 2015; Gillion, Ladd,
feminist identification is relatively uncommon overall,
and Meredith 2018; Kaufmann and Petrocik 1999;
and many...