Party Identification in the Age of Obama: Evidence on the Sources of Stability and Systematic Change in Party Identification from a Long-Term Panel Survey

• Date: 01 June 2019
• Author: Jacob M. Montgomery,Patrick D. Tucker,Steven S. Smith
• Published date: 01 June 2019
• DOI: 10.1177/1065912918784215
• Subject Matter: Articles

https://doi.org/10.1177/1065912918784215

Political Research Quarterly

2019, Vol. 72(2) 309 –328

© 2018 University of Utah

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/1065912918784215

journals.sagepub.com/home/prq

Article

Introduction

The nature of party identification change is the subject of

one of the longest running, most prominent debates in the

study of American political behavior. The argument that

party identification is a stable, social psychological,

affective phenomenon is a central feature of The American

Voter and the Michigan School of political behavior

(Campbell et al. 1960). In contrast stand accounts that

treat party identification as a dynamic cognitive phenom-

enon that is influenced by short-term political forces.

In this article, we leverage data from The American

Panel Survey (TAPS), which asked the standard party

identification (PID) battery twenty times over the 2011–

2016 period, to reexamine these questions and provide a

clearer answer about the nature of party identification. A

panel interviewed on such a frequent basis over such a

long period of time—nearly the entirety of President

Obama’s second term—provides an unprecedented look

at party identification in the United States. Moreover,

TAPS included monthly questions evaluating the parties,

presidential job performance, and views on the state of

the economy, which allows us to assess the relationship

between moving political evaluations and party identifi-

cation. In all, these data give us a fresh and powerful

basis for reevaluating long-standing arguments about the

nature of party identification change.

We take advantage of the uniquely rich data from

TAPS to support three hypotheses. First, we show that

party identification changes systematically at the indi-

vidual level even after accounting for measurement error

and autocorrelation. This finding directly contradicts

several previous accounts that treat changes in party

identification as artifacts of the survey process. Second,

we find that shifts in party identification are associated

with individuals’ changing evaluations of the two major

parties and appraisals of the president. We believe that

our analysis is the most comprehensive study to date

demonstrating a direct link between evaluations of the

784215PRQXXX10.1177/1065912918784215Political Research QuarterlyTucker et al.

research-article2018

1Yale University, New Haven, CT, USA

2Washington University in St. Louis, St. Louis, MO, USA

Corresponding Author:

Patrick D. Tucker, Institution for Social and Policy Studies, Yale

University, 77 Prospect Street, New Haven, CT 06511, USA.

Email: patrick.tucker@yale.edu

Party Identification in the Age of Obama:

Evidence on the Sources of Stability and

Systematic Change in Party Identification

from a Long-Term Panel Survey

Patrick D. Tucker1, Jacob M. Montgomery2, and Steven S. Smith2

Abstract

Political scientists have long disagreed about the nature of individual-level change in party identification. While

some scholars conclude that party identification is a stable identity—attributing changes in individual responses to

measurement error—others show that aggregate party identification responds systematically to short-term forces

such as presidential approval. In this article, we use a unique long-term panel measuring party identification twenty

times in the 2011–2016 period to support a subtle compromise between these competing claims. We show that

individual-level party identification changes systematically over time even after accounting for measurement error and

that this change is related to short-term evaluations of the parties and the president. However, although such change

exists, it is modest in the medium term and more common among specific subsets of respondents. Finally, we show

that that these findings are robust to numerous alternative modeling strategies. We believe that our analysis provides

the most systematic examination to date of individual-level changes in party identification.

Keywords

party identification, panel data, presidential approval

310 Political Research Quarterly 72(2)

political environment and changes in party identification

at the individual level. Third, despite the evidence of sys-

tematic change in individual-level identification, we also

demonstrate that party identification remains stable for

certain individuals. In particular, we show that change in

party identification is less common among African

Americans and negatively associated with age and famil-

ial socialization. Finally, in “Alternative Analyses:

Possible Endogeneity” section and our online appendi-

ces, we show that these findings are robust to a number

of alternative modeling strategies.

In all, our study concludes that while party identifica-

tion is remarkably stable and persistent for many voters,

it is not the “unmoved mover” as it is sometimes charac-

terized in the literature, where changes occur only during

periods of political instability (i.e., “realignments”).

Neither, however, does party identification fluctuate flu-

idly in response to every political event. Instead, we

argue that party identification changes at the margins in

response to short-term political forces, but that this

change is gradual and restricted to individuals with less

entrenched partisan identification. That is, party identifi-

cation is a reflection of individuals’ present evaluations

of the political environment and changes meaningfully

during periods of normal politics to reflect citizens’ eval-

uations of current events and the political environment.

Yet, changes are neither rapid nor universal, occurring

gradually and primarily among individuals whose back-

ground and circumstances fail to reenforce their partisan

social identity.

Central Questions About Party

Identification

The social psychological concept of party identification

has its origins in some of the earliest work on political

behavior (e.g., Lazarsfeld, Berelson, and Gaudet 1944).

In this view, party identification is a social identity devel-

oped in early life experience, is internalized through

socialization, biases perceptions of the social world, and

has a lasting effect on the political attitudes of many, if

not most, people (Hyman 1959; Mannheim 1952). This

understanding was extended in The American Voter,

which treats party identification as an indicator of a psy-

chological attachment—a social identity—that biases the

processing of information about candidates, issues,

events, and vote choice (Campbell et al. 1960). This pro-

cess generates a “persistent adherence” to a party that,

while not immutable, stabilizes partisan preferences over

a lifetime. Further studies directly confirmed that party

identification is transmitted (if imperfectly) from genera-

tion to generation as a social identity. In particular, the

Niemi and Jennings studies (e.g., Jennings and Niemi

1981) measure party identification of both parents and

children in interviews conducted in 1965, 1973, 1982,

and 1997 and show the transmission of party identifica-

tion as a social identity through familial socialization.

In contrast stand various theoretical accounts that treat

responses to party identification as a more fluid concept:

an affective attachment that responds to the flow of politi-

cal events and reflects changing views of the two major

parties, the economy, and the president. Various theoreti-

cal accounts in this PID-as-dynamic-attitude camp, with

origins dating as far back as Downs (1957) and Key

(1966), argue that short-term political forces readily influ-

ence party identification (see also Achen 1992; Fiorina

1981). For instance, aggregate data suggest that party

identification at the national level moves systematically in

response to events such as recessions and wars (MacKuen,

Erikson, and Stimson 1989). Economic performance, con-

sumer sentiment, and crucial political events all predict

changes in aggregate-level party identification, which

suggests (but does not prove) that individual-level parti-

san identity responds to these factors as well.

The debate between these competing camps on the

nature of party identification is vast and rich, and no one

study is capable of arbitrating between all of the available

theories. (Indeed, it is not clear that there truly is a single

analysis that could put this longrunning debate to rest.)

Nonetheless, we believe that the unique granularity of the

TAPS survey, which measures party identification twenty

times over the course of four and a half years, justifies

reexamining some of the central questions in the litera-

ture: (1) Does party identification change? (2) Does party

identification change in response to short-term political

evaluations? and (3) What are the sources of individual-

level heterogeneity in party identification stability?

One of the central questions in the literature has always

been about the scope and even the existence of meaning-

ful party identification change. Broadly speaking,

between 41 and 48 percent of respondents to the American

National Election Study panels, the inevitable source of

data in this debate, stay within the same partisan catego-

ries (Clarke and McCutcheon 2009). Fiorina (1981) inter-

preted these patterns as demonstrating that party

identification is quite unstable and, therefore, that the

party identification is a moving evaluation (see also

Allsop and Weisberg 1988; Jackson 1975; Jackson and

Kollman 2011).1

In contrast stand the studies summarized in Green,

Palmquist, and Schickler (2002) that, using nearly the

same data, conclude that party identification is a stable

identity. These studies argue that survey responses

involve a random component, which, left uncorrected,

upwardly biases estimates of partisan instability. For

instance, Green and Palmquist (1990) show that uncor-

rected wave-to-wave correlations for the 1980 ANES

(American National Election Study) panel range from

.85 to .89, which implies moderate partisan instability.

However, error-corrected correlations are .98, which