Do Abortion Attitudes Lead to Party Switching?

• Date: 01 December 2008
• Author: Clyde Wilcox,Mitchell Killian
• DOI: 10.1177/1065912907312981
• Published date: 01 December 2008
• Subject Matter: Articles

Political Research Quarterly
Volume 61 Number 4
December 2008 561-573
© 2008 University of Utah
Do Abortion Attitudes Lead to
10.1177/1065912907312981
http://prq.sagepub.com
Party Switching?
hosted at
http://online.sagepub.com
Mitchell Killian
George Washington University, Washington, D.C.
Clyde Wilcox
Georgetown University, Washington, D.C.
The notion that issues and ideology can move partisanship remains controversial. The authors explore the stronger
claim that issues can lead people to switch political parties and whether the effect of abortion attitudes is asymmetri-
cal (i.e., abortion attitudes may influence party switching in only one direction). They show that in several short-term
National Election Studies panels, pro-life Democrats were significantly more likely than other Democrats to become
Republicans, but pro-choice Republicans were not likely to become Democrats. However, using panel data over a
long time frame, 1982 to 1997, the authors also demonstrate that the cumulative effect of abortion attitudes led pro-
life Democrats and pro-choice Republicans to switch parties.
Keywords:
party identification; abortion; party switching; changes in party identification
During football season, discussions of upcoming 2002; Green and Palmquist 1990, 1994) argued that
games are frequently pitched as a clash between
party identification is a deeply rooted social identity,
the “irresistible force” and the “unmovable object.”
similar to ethnic and religious identities, that is rela-
Academic discussions of the relationship between
tively impervious to outside forces. They charge that
partisanship and deeply held political views use dif-
previous studies that showed that partisanship
ferent words but paint a similar picture. Partisanship
responded to issues and candidates were the result of
is the unmovable object, a stable orientation inherited
measurement error and that once measurement error
through socialization and reinforced through social
is controlled for, party identification is largely
identity networks. Yet some issues are also deeply
unchanged by issue positions.
rooted in central moral and religious values and rein-
Even the most ardent proponents of partisanship as
forced by social identities and are used by political
an unmovable object acknowledge that issues can at
elites to mobilize support. What happens when the
times influence partisanship. Campbell et al. (1960,
irresistible force of a deeply held political issue
135) suggested that issues that are inconsistent with
clashes with the unmovable force of partisanship?
partisanship can “exert some pressure on the individ-
The relationship between partisanship and strongly
ual’s basic partisan commitment. If this pressure is
held political attitudes has been the subject of debate
intense enough, a stable partisan identification may
for nearly fifty years. The American Voter argued that
actually be changed.” And Green and Palmquist
partisanship was a deeply held, enduring psychologi-
(1994, 456) acknowledge that “some dataset yet
cal attachment that influenced attitudes toward candi-
unanalyzed may turn up evidence of micropartisan
dates and issues (Campbell et al. 1960). But a
adjustment in response to issues, candidates, perfor-
revisionist literature in the 1980s and 1990s suggested
mance, or voting behavior.”
that partisanship was more malleable and that individ-
uals might alter their partisanship based on campaign
events (Allsop and Weisberg 1988), economic condi-
Mitchell Killian, PhD Candidate in Political Science, George
tions and evaluations of the president (MacKuen,
Washington University; e-mail: mkillian@gwu.edu.
Erikson, and Stimson 1989), ideology, parents’ politi-
Clyde Wilcox, Professor of Government, Georgetown University;
cal activity and age (Clark et al. 1991), and retrospec-
e-mail: wilcoxc@georgetown.edu.
tive evaluations of the economy (Fiorina 1981).
Authors’ Note: John Bruce, Kim Gross, Bill Jacoby, Eric Lawrence,
More recently, a series of studies by Donald Green
Barbara Norrander, John Sides, and Lee Sigelman provided extremely
and colleagues (Green, Palmquist, and Schickler
useful comments and suggestions on earlier versions of this article.
561

---

562
Political Research Quarterly
But other research continues to suggest that posi-
The issue evolution of abortion and partisanship
tions on issues can over time influence partisanship.
has been dramatic: over the 1970s, the correlation
Several studies have argued that ideological identities
between abortion attitudes and partisanship in the
can influence partisan ones. Putz (2002) shows that
General Social Survey was .07; in 2004 it was –.24.1
individuals adjusted their partisanship in the early to
A growing convergence of partisanship and abortion
mid-1990s to bring it more in line with their ideology.
attitudes could come about in several ways. First, new
More broadly, Abramowitz and Saunders (1998) argue
cohorts of voters with consistent attitudes may have
that the increasing correlation between ideology and
replaced older voters who were conflicted. Second, as
partisanship over time is almost entirely because of cit-
parties sent increasingly clear signals, partisans may
izens adjusting their partisanship to match their ideo-
have been persuaded and adjusted their abortion atti-
logical positions. But can specific issue positions cause
tudes to meet their partisanship (Zaller 1992; Layman
individuals to change their partisanship?
and Carsey 1998; Layman and Carsey 2002).
But it is also possible that some citizens changed
their partisanship to fit their abortion attitudes.
Party Identification and
Adams (1997) suggests but does not demonstrate that
Abortion Attitudes
individuals adjusted their partisanship to fit their
abortion attitudes. He argues that long-term panel
If any issue attitude could play the role of “irre-
data would be needed to determine the causal order-
sistible force,” it is abortion. The issue has transformed
ing. Using shorter-term panels, Carsey and Layman
ordinary Americans into extraordinary activists
(2006) show that both processes took place during
(Maxwell 2002; Verba, Schlozman, and Brady 1995).
the 1992–94–96 panel survey. Those respondents
It has produced collective action of unusual intensity
whose abortion positions are the most salient are
and duration, inspiring mass demonstrations and vio-
likely to adjust their partisanship to fit with their
lence in isolated incidents. Abortion attitudes influence
views on abortion, and those with less salient abor-
vote choice in presidential, Senate, and gubernatorial
tion attitudes are likely to be persuaded by their
elections, even when the conventional wisdom sug-
party’s position on the issue.
gests that “it’s the economy stupid” (Abramowitz
Carsey and Layman (2006) show that abortion atti-
1995; Cook, Jelen, and Wilcox 1994a, 1994b; Cook,
tudes lead to changes in partisanship along the tradi-
Wilcox, and Hartwig 1993).
tional 7-point National Election Studies (NES) scale.
Yet in the aggregate, the public holds an ambivalent
The effects of abortion on partisanship overall in this
and collectively moderate position on abortion (Cook,
brief period are relatively small. Indeed, they con-
Jelen, and Wilcox 1992). This ambivalence comes from
clude that those who believe that abortion should
conflicting values; many Americans value both an emer-
never be allowed increase their Republican identifi-
gent fetal life and women’s moral autonomy (Alvarez
cation just 0.3 points more on the 7-point partisan-
and Brehm 1995). To the question of whether abortion
ship scale than individuals who believe abortion
should be allowed by law, the median voter replies “it
should always be allowed.
depends”—on the reasons the woman is seeking the
But political professionals are convinced that abor-
abortion, on the timing of the abortion, and even on the
tion attitudes not only influence the strength of parti-
procedures used (Norrander and Wilcox 2002).
san attachments but also lead to party switching. Thus,
Adams (1997) chronicles the “issue evolution” of
we are not interested in whether abortion moves citi-
the abortion issue, from one pitting well-educated and
zens from weak to strong Republicans or from weak
more-secular citizens against those with less education
Republicans to independents. We seek to determine
and more religiosity to one that divided partisans. Pro-
whether the “irresistible force” of the abortion issue
life and pro-choice activists used party nominations to
moves some citizens from being Republicans to
replace abortion moderates with candidates who held
Democrats, and vice versa. In this article, we explore
more extreme positions. Over time, the voting behav-
the role of abortion attitudes on party switching dur-
ior of members of Congress diverged on abortion
ing the past two decades at the individual level.
along party lines, and subsequently the correlation
We examine this question differently for those
between partisanship and abortion attitudes began to
who begin as Republicans and those who begin as
rise in the general public. Carmines and Woods (2002)
Democrats, for there is no reason to believe that the
suggest that the polarization occurred among other sets
impact of abortion on partisanship is symmetrical.
of activists, including convention delegates and those
Indeed, there are reasons to believe that the impact of
who are intermittently active in campaigns.
abortion on partisan attachments might be...