Examining Trends in Ideological Identification: 1972–2016
| Published date | 01 May 2021 |
| Date | 01 May 2021 |
| DOI | http://doi.org/10.1177/1532673X20961314 |
| Subject Matter | Articles |
research-article2020
Article
American Politics Research
2021, Vol. 49(3) 259 –268
Examining Trends in Ideological
© The Author(s) 2020
Article reuse guidelines:
sagepub.com/journals-permissions
Identification: 1972–2016
https://doi.org/10.1177/1532673X20961314
DOI: 10.1177/1532673X20961314
journals.sagepub.com/home/apr
Adrien A. Halliez1 and Judd R. Thornton1
Abstract
Has polarization influenced how members of the public identify with ideological labels? In our analysis of patterns of ideological
identification since the 1970s, we demonstrate that there has been an increase in the proportion of the electorate willing
to locate themselves on the standard seven-point ideological scale as the parties have polarized. Moreover, consistent with
existing evidence of partisan-ideological sorting, our results indicates that most of the increase in identifying with a label is
associated with an increase in partisans selecting the ideological label that matches their partisanship. Finally, we show that
attitudes toward moral traditionalism are increasingly related to ideological identifications. Our evidence indicates that the
broader political system influences how members of the public identify with ideological labels.
Keywords
ideology, polarization, moral traditionalism
Why do individuals choose to identify with an ideological
answer the self-identification questions compared to more
label? Ideological self-identification remains an object of
than a third of respondents in 1980. Moving to an examina-
inquiry even though such identifications do not represent a
tion of partisans, we find that the increase of respondents
belief system for many in the electorate. The “ideological
answering the question has coincided with an increase in
innocence” conclusion reached by Converse (1964) extends
partisan-ideological sorting, while the percentage of parti-
to more recent studies (e.g., Kinder & Kalmoe, 2017;
sans who identify as moderate has remained relatively stable
Lupton et al., 2015). While the evidence for this claim is
over the last four decades. Finally, the relationship between
strong, examinations of the public have also long made
traditional moral values and self-identification has increased
clear that “the ‘nonideological’ use of ideological labels is
markedly since the 1980s. Taken together, our findings indi-
an important feature of mass politics in America” (Levitin
cate that as political conflict becomes more unidimensional
& Miller, 1979, p. 751), a finding corroborated by contem-
and polarized among political elites, those same areas that
porary evidence (e.g., Mason, 2018). In other words, large
were absorbed by the first dimension of elite conflict now
portions of the public identify with ideological labels for
play a role in how the public identifies with ideological
non-ideological reasons and doing so has important conse-
labels. While the public, on average, remains more moderate
quences. Thus, understanding the conditions under which
than elected officials, elite polarization has nevertheless had
people choose to identify with a label remains vital.
a noticeable impact on how citizens evaluate and identify
We argue that individuals should be more willing to locate
with political symbols, including ideological labels.
themselves on the standard seven-point ideology scale as a
result of clearer signals from polarized elites. Further, consis-
Background
tent with evidence of partisan-ideological sorting, we expect
that this increase has primarily resulted in partisans picking
As the political parties have polarized and the primary
the “correct” ideological label. Finally, we contend that as
dimension of partisan conflict has absorbed previously sub-
issues that were once cleavages within the parties come to rep-
sidiary concerns (e.g., McCarty et al., 2006), scholars have
resent areas of competition between the parties, the relevance
attempted to measure the response of the mass public. Some
of these issues will be increasingly associated with ideological
argue that individuals are responding to elite polarization by
identifications. In particular, we expect that attitudes about
moral traditionalism will increase in importance.
1Georiga State University, Atlanta, GA, USA
Using the American National Election Studies, we find
support for all three expectations. The results indicate that
Corresponding Author:
Judd R. Thornton, Georiga State University, 38 Peachtree Center Ave.,
the public has become more willing to pick an ideological
Suite 1005, Atlanta, GA 30350, USA.
label—in 2016, fewer than 20% of respondents refused to
Email: jrthornton@gsu.edu
260
American Politics Research 49(3)
polarizing themselves (Abramowitz & Saunders, 2008) and
identification is relatively stable over time (Kinder &
that issue attitudes, ideology, and partisanship are now more
Kalmoe, 2017, p. 75). Further, one’s ideological identifica-
closely aligned (Bafumi & Shapiro, 2009). Others view the
tion has its own independent effect on attitudes and behavior:
electorate’s response as more modest: for example, while
for example, “policy judgments are shaped by both symbolic
partisans are more likely to support their party, they have
and principled considerations” (Popp & Rudolph, 2011,
remained mostly moderate (Ansolabehere et al., 2006;
p. 818); and, identification can influence how one labels
Fiorina et al., 2011) and much of the public nowadays does
specific policies (Claassen et al., 2015), along with ratings of
not exhibit appreciably greater issue constraint (Baldassarri
candidates (Malka & Lelkes, 2010) and one’s vote choice
& Gelman, 2008; Lupton et al., 2015). Besides, a non-trivial
(Devine, 2015). Moreover, Mason (2018, p. 885) demon-
portion do not cleanly map on to the left-right dimension
strates that ideological identification “can drive affective
(e.g., Carmines et al., 2012b).1 In spite of these disagree-
ideological polarization even when individuals are naïve
ments, a strong consensus has emerged regarding the
about policy.” In summary, it is fair to say that one’s ideo-
increased importance and salience of partisanship (e.g.,
logical identity has important influence on attitudes and
Bafumi & Shapiro, 2009; Bartels, 2000) and the trend of
behavior even when it does not represent a comprehensive
partisan-ideological sorting (Abramowitz, 2010; Levendusky,
belief system.
2009).2 Such sorting occurs when an individual’s ideological
An important way in which ideological identification dif-
identification is in line with his or her partisan identification
fers from that of partisanship is that a large proportion of
and has important consequences, including on voting behav-
respondents decline to answer the question—Kinder and
ior (Davis & Mason, 2016).
Kalmoe (2017, p. 58) note that 27.5% of respondents in a
Although much has been learned about the causes and
pooled sample from 1972 to 2012 do not place themselves on
consequences of partisan-ideological sorting, relatively less
the ideological scale. However, looking at a pooled sample
recent attention has been paid to the process of whether citi-
may miss important dynamics in how the electorate has iden-
zens identify with an ideological label at all. This lack of
tified with ideological labels over the last several decades
attention may be due to the extensive evidence that a large
(e.g., Carmines et al., 2012a; McCarty et al., 2006; Norrander
portion of the electorate falls short of ideological thinking.
& Wilcox, 2008; Theiss-Morse et al., 2018, pp. 191–193;
Campbell et al. (1960, p. 192) define ideology as a “close-
Thornton, 2013). It is precisely such a phenomenon that we
woven, and far-ranging structure of attitudes.” Converse’s
seek to examine.
(1964, p. 207) characterization of belief systems being
“bound together by some form of constraint” speaks to that
Empirical Expectations
logic. Of course, these early studies concluded that much of
the electorate failed to engage in ideological thinking. Still,
The argument presented here builds off of previous evidence
while only a fraction of respondents at any given time would
indicating that as the parties sort and polarize, the signals
be classified as an “ideologue” (Campbell et al., 1960;
sent to the electorate become clearer (Hetherington, 2001), a
Lewis-Beck et al., 2008), many non-ideologues are perfectly
trend that has continued (Smidt, 2017). Building upon such
willing to identify with an ideological label (Levitin &
findings, we argue that if there is some cognitive “cost” that
Miller, 1979). And, while the precise influence of one’s ideo-
exists to identify with an ideological label, a portion of that
logical identification varies in the electorate (e.g., Holm &
cost can be “subsidized” by polarized parties and non-elected
Robinson, 1978; Jacoby, 1991; Knight, 1985), it appears to
political actors, such as the media, interest groups, and pub-
be important regardless of one’s penchant for engaging in
lic intellectuals. When members of the public are more
abstract thinking about politics.
frequently exposed to the terms “liberal” and “conservative,”
Self-identification’s influence arises in part because it can
it makes ideological labels more salient and relevant to mem-
act as a “predisposition” (e.g., Sears et al.,...
To continue reading
Request your trial