q131/random_k4/random_4.md

DOI: 10.1111/polp.12517

O R I G I N A L   A R T I C L E

The fluid voter: Exploring independent voting patterns
over time

Thom Reilly1

|  Dan Hunting2

1School of Public Affairs, Arizona State
University, Phoenix, Arizona, USA

2Lodestar Center for Philanthropy and
Nonprofit Innovation, Arizona State
University, Phoenix, Arizona, USA

Correspondence

Thom Reilly, School of Public Affairs, Arizona
State University, 411 N Central Ave. Office 422K,
Mail Code 3720, Phoenix, AZ 85005, USA.
Email: thom.reilly@asu.edu

Abstract
Independents remain hard to categorize because they are, by
their choice of self-identification, resisting the standard cate-
gories of political classification. Despite the growth in inde-
pendent  voter  identity,  many  political  strategists  still  view
independents as partisans. In this article, we contribute to the
academic literature on independent voting behavior by explor-
ing  whether  those  who  identify  as  politically  independent
function as true independents by accounting for their voting
patterns over time. We do this by analyzing data produced by
the American National Election Studies (ANES) on political
identification and voting choices from 1972 to 2020 on each
of the three ANES measures of party affiliation. Our findings
show when tracking independent voting behavior over more
than  one  election,  there  is  a  significant  volatility  in  voting
loyalty and independents as a group are distinct from parti-
sans. This volatility was observed in all three measures of party
affiliation used by the ANES survey data. The research also
finds evidence that a sizeable number of independents move
in and out of independent status from one election to another.

K E Y W O R D S
ANES, elections, fluid voter, independent voter, over time, partisanship,
political behavior, political parties, United States, volatility, voter identifi-
cation, voting behavior, voting loyalty

Related Articles
Grossmann, Matt. 2014. “The Varied Effects of Policy Cues
on  Partisan  Opinions.”  Politics  &  Policy  42(6):  881–904.
https://doi.org/10.1111/polp.12102.
Reilly, Thom, and E. C. Hedberg. 2022. “Social Networks
of Independents and Partisans: Are Independents a Moder-
ating Forcer?” Politics & Policy 50(2): 225–43. https://doi.
org/10.1111/polp.12460.
Saeki,  Manabu.  2019.  “Anatomy  of  Party  Sorting:  Parti-
san Polarization of Voters and Party Switching.” Politics &
Policy 47(4): 699–747. https://doi.org/10.1111/polp.12318.

Politics & Policy. 2023;00:1–22.

wileyonlinelibrary.com/journal/polp

1

© 2023 Policy Studies Organization.
2

THE FLUID VOTER

Americans are increasingly declaring independence from the political parties. The rise in political
independence is likely an outgrowth of Americans' record or near-record negative views of the
U.S. two-party system (Ingraham, 2021) and their low level of trust in government (PEW, 2022).
Self-defined  independent  voters  now  number  between  40%  and  46%  of  the  U.S.  electorate
(Gallup, 2022) and currently constitute either the largest or second-largest group of registered
voters in half the states (Gruber & Opdycke, 2020). Despite the historical increase in independ-
ent voter identification, many political strategists still view independents as partisans (Magleby
et al., 2011; Petrocik, 2009) and contend that the overwhelming majority of Americans who say
they are “independent” really lean toward one party or the other. However, other scholars have
disputed the findings that most independents are leaners and suggest that there is no conclusive
evidence for this position (Abrams & Fiorina, 2011).

Our study seeks to contribute to the academic literature by exploring whether those who are
identified as politically independent function as true independents by accounting for their voting
patterns over time. We are interested in determining whether independents move in and out of
independent  status.  We  do  this  by  reviewing  the  voting  behavior  of  Democrats,  Republicans,
and independents over multiple election cycles. Our research seeks to address the following three
questions:

1.  Does political identification change across time?
2.  How do respondents allocate their votes across parties?
3.  Do voting choices change over time?

LITERATURE REVIEW

The classification of voters as independent dates back to the seminal work of Angus Campbell
and  his  colleagues,  who  first  published  The  American  Voter  in  1960  (Campbell  et  al.,  1960).
Analyzing data collected under the University of Michigan Survey Research Center (and later
aggregated by the American National Election Studies; ANES Data Center, 2021), the authors
describe  the  identity  of  party  affiliation  as  a  central  characteristic  explaining  voting  behavior
and other political attitudes and behaviors. The surveys that The American Voter analyzed have
been considered by many to be the gold standard in the field. Though officially founded in 1978,
the American National Election Studies  (ANES) program has  continuous  survey  data on the
electorate since 1948. The survey is usually administered every other year, but occasionally every
fourth year. ANES is a comprehensive survey which provides much information on respondents'
background and political attitude.

The  American  Voter  authors  acknowledged  that  some  kind  of  “independent”  existed  but
characterized the independent as having little interest in campaigns and outcomes and suggested
their choice between competing candidates is uninformed. Most of what we know about inde-
pendents comes from survey data, and most surveys predispose the majority of independents as
leaners toward either of the two political parties. Since 1952, when individuals identified them-
selves as an independent, researchers and pollsters have asked a follow-up question on whether
respondents prefer one party over the other if they had to vote then and there.

In addition to asking respondents to identify themselves from a three-point scale: Democrat,
Republican, and independent, respondents were asked to self-identify on the ANES seven-point
political spectrum (ANES Data Center, 2015):

1.  Strong Democrat
2.  Democrat
3.  Independent, leans Democrat
4.  Independent

REILLY anD HUnTInG

3

5.  Independent, leans Republican
6.  Republican
7.  Strong Republican

Since the seven-point measure was introduced in the 1952 survey, researchers accessing the ANES
data  were  able  to  use  several  measures.  They  could  use  the  seven-point  measure,  a  five-point
measure by collapsing the three independent categories into one (as the authors of the Amer-
ican  Voter  did),  a  three-point  measure  with  leaners  classed  as  independents,  or  a  three-point
measure with leaners classified as partisans. Researchers used any or all of these measures often
depending on which coding decision gave them big enough cell sizes for analysis by re-coding the
data or not (DeBell, 2010).

Viewing the majority of independents as partisans originates from the formative research
popularized in The Myth of the Independent Voter (Keith et al., 1992), which claimed that the
ANES'  “Seven-Point  Scale”  should  only  include  three  actual  categories  (Democrat,  Repub-
lican,  and  Independent).  After  the  Petrocik  (2009)  and  Keith  and  others'  (1992)  articles,  it
became  more  common  to  use  a  five-point  or  three-point  measure  with  leaners  classified  as
independents.

Most  independents  indicated  a  lean  toward  one  of  the  two  major  political  parties'  candi-
dates. Political scientists have labeled these individuals as “independent leaners” and have argued
that the number of pure independents is actually quite small—below 10%. This percentage has
remained constant since the 1950s (Mayer, 2008; PEW, 2019; Sides, 2013), and many political
scientists assert that the overwhelming majority of Americans who say there are “independent”
lean toward one party or the other (Teixiera, 2012).

Klar and Krupnikov (2016) have recently added some important research on the independent
voter. They explored the social significance of the growth in people refusing to identify them-
selves with a political party and suggested that independents and partisans differ psychologically
(Klar  &  Krupnikov,  2016).  They  do  not  dispute  the  notion  that  independents  may  be  “closet
partisans” (they call them “undercover partisans”); but they do dispute the bias that independ-
ents are not politically engaged, stating that “engagement levels are comparable across independ-
ents  and  partisans”  (Klar,  2014).  They  assert  that  many  Americans  are  embarrassed  by  their
political party and do not wish to be associated with either side. Instead, they intentionally mask
their party preference, especially in social situations (Klar & Krupnikov, 2016). Nonetheless, they
contend that the refusal to publicly identify with a party must be revealing something important.
And they believe the predictors of independent political engagement differ substantially from
partisans.

However,  there  are  some  researchers  that  disagree  with  the  assertion  that  independents
are  leaners  and  suggest  there  is  more  volatility  in  their  voter  patterns,  and  that  a  sizeable
number  of   independents  move  in  and  out  of   independent  status  in  ways  that  impact  inde-
pendent  voting  over  time  (Abrams  &  Fiorina,  2011;  Fiorina,  1977,  2016;  Jackson,  1975;
Page & Jones, 1979). Their identification may depend on specific candidates or issues on the
ballot (Reilly et al., 2023) or may derive from short-term interest rather than a long-standing
loyalty (Miller, 1991). Fiorina (2017), professor of  political science at Stanford University and
former  chairman  of   the  board  of   the  ANES,  contends  that  following  independent  leaners
over several elections is key to understanding their voting patterns. Along with his colleague
Samuel  J.  Abrams,  they  conducted  such  an  analysis  and  found  that,  following  independent
leaners  across  multiple  elections,  their  partisan  stability  is  closer  to  pure  independents  than
weak partisans (Fiorina, 2017). They also noted that “classifying all leaners as weak partisans
mis-characterizes the partisanship of  Americans and overestimates the rate of  party voting”
(Abrams & Fiorina, 2011). Other researchers have argued that responses to survey question
probes  asking  independents  if   they  lean  toward  the  Democratic  or  Republican  Party  are
significantly contaminated by short-term electoral elements operating in the campaign, such

4

THE FLUID VOTER

as  the  candidates  and  specific  issues  (Abrams  &  Fiorina,  2011;  Brody,  1978,  1991;  Brody  &
Rothenberg, 1988; Miller, 1991).

Finally,  given  the  lack  of  data  on  voting  patterns  of  independents  in  state  races  and
down-ballot  (other  than  for  president,  governor,  and  Congress),  there  is  growing  interest  in
examining  the  characteristics  and  attitudes  of  unaffiliated  or  independent  voters  as  they
compare to voters from the two major parties. Bitzer and others (2022) researched down-ballot
voters in North Carolina and found unaffiliated voters were not simply shadow partisans but
varied from Democrats and Republicans in terms of demographics, political behavior, and polit-
ical attitudes.

OUR EXPECTATIONS

Our  study  seeks  to  contribute  to  the  academic  literature  by  exploring  whether  those  who  are
identified as politically independent function as true independents by accounting for their voting
patterns over time. The study also explores whether independents move in and out of independ-
ent status.

We begin with a description of ANES data and the three measures of party affiliation used
by  the  survey.  Analysis  then  begins  with  a  look  at  how  the  political  identification  of  voters
changes over multiple survey waves using each of the three ANES scales to be described in the
next section. This analysis includes a parallel review of respondents who voted in both waves and
those who voted in neither wave. We next investigate how frequently respondents to the ANES
vote  “straight  tickets”—always  choosing  candidates  from  the  same  party—or  “mixed  tickets”
where  some  Republican  and  some  Democrat  candidates  are  chosen.  It  is  expected  that  those
identifying  as  Democrat  or  Republican  will  mostly  choose  candidates  from  their  own  party,
while independents will show more variety in their choices. These results will also be reported on
each of the three political identification scales discussed below. Finally, we explore the degree to
which individuals change their voting choices over time.

METHODOLOGY

The ANES Cumulative Data File (CDF) is used to examine political identification and voting
choices from 1972 to 2020 (ANES Data Center, n.d.). Although the CDF contains data dating
back to 1948, restricting the analysis to data from 1972 onward provided the best balance of: (a)
providing a large enough sample to be useful and (b) capturing attitudes and trends that are rele-
vant in the current social and political climate. There have been substantial demographic changes
in the United States over the 72 years of ANES data. Additionally, prior to the passage of the
Voting  Rights  Act  of  1965,  large  portions  of  the  population  were  effectively  disenfranchised.
These  changes  become  evident  when  pre-1972  ANES  data  are  compared  to  1972–2020  data.
Prior to 1972, 6.4% of ANES respondents who reported voting were non-White, but the figure
jumps to 22.4% when looking at voters between 1972 and 2020. This percentage is much more in
line with current voter profiles. Smaller, but still important, changes are evident in the age distri-
bution  and  gender  of  respondents  across  the  two  time  periods.  Eighteen-  to  twenty-year-olds
made  up  1.5%  of  the  pre-1972  respondents  and  5.9%  after  that.  Females  made  up  52.4%  of
the pre-1972 respondents and 53.8% from 1972 to 2020. The 1972–2020 data are more closely
aligned with current voter demographics, making this analysis more applicable to today's voters.
This dataset includes political identification information for respondents and self-reported
voting  choices  for  president,  Senate,  Congress,  and  governor  races.  Since  the  ANES  does  not
show 2020 election preference data in the CDF, these data were linked  to  the  CDF using the
Respondent ID from the ANES 2020 timeseries file. The resulting file was then formatted so that

REILLY anD HUnTInG

5

each record represented a unique respondent, capturing party identification on three scales and
reported voting choices for all survey waves that each respondent answered.

We examine three questions, looking at each on three different political identification scales:

1.  Does political identification change across time?
2.  How do respondents allocate their votes across parties?
3.  Do voting choices change over time?

To probe these areas, we look at political identification with three different scales commonly used
by  the  ANES.  First,  with  the  initial  party  identification  queried  in  question  VCF0302,  which
asks, “Generally speaking, do you usually think of yourself as a Democrat, a Republican, an
Independent, or what?” This “Initial Party ID Response” gives a three-point scale with Demo-
crats  and  Republicans,  and  all  minor-party  and  independent  respondents  grouped  under  the
independent  umbrella.  Second,  we  use  the  ANES  Seven-Point  Scale  (VCF0301,  “Seven  Point
Scale”) that divides Democrats and Republicans into “strong” and “weak” supporters of their
parties, and divides independents into Democrat-leaners, Republican-leaners, and true independ-
ents. Finally, we use the modified three-point scale from VCF0303 “Summary 3-Category.” The
Summary  3-Category  measure  collapses  the  Seven-Point  Scale  by  counting  Democrat-leaning
independents as Democrats and Republican-leaning independents as Republicans. This leaves
only a small fraction of the respondents as independents.

American National Election Studies Survey

The CDF was downloaded in SPSS format and filtered to include responses from 1972 to 2020.
Since  2020  post-election  voting  information  is  not  currently  in  the  CDF,  data  from  the  2020
time-series file were joined to the CDF to provide complete information on the 2016–2020 panel.
The resulting file contains responses from 43,423 individuals, of whom 27,832 voted in at least
one election (Table 1).

The  ANES  contains  data  on  how  respondents  reported  voting  on  four  different  contests,
giving the party choice for president, Congress, Senate, and governor. Respondents do not neces-
sarily vote in each of these races due to the timing of elections. Choices for a total of 77,729
contests are recorded for the 27,832 voters in the data. For each election cycle, the total number
of votes for Democratic, Republican, and third-party candidates were totaled for each respond-
ent, along with their party identification at that time. The survey has also included time-series
panel data from time to time, where respondents are contacted multiple times over the years. For
respondents that appeared in multiple waves of the survey, their votes and party identification
were tallied at each survey point.

Since 1970, there have been seven panels as shown in Table 2, each covering a single presi-
dential election. Of the 13,399 respondents to the survey in these panels, 4770 voted in all waves
available to them. These voters reported their voting choice on a total of 25,024 races for pres-
ident, Congress, Senate, and governor. Due to the timing of election cycles, not all respondents
reported voting in each of these races in each wave of the survey.

Party identification on three scales

ANES CDF classifies the political identification of respondents according to their answers to
questions VCF0302, VCF0301, and VCF0303.

VCF0302 is the initial party identification response and asks:

6

THE FLUID VOTER

T A B L E   1  ANES respondents by year

Total respondents

Respondents who voted

Total votes tallied

Year

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2008

2012

2016

2020

2705

475

1323

2304

1614

1418

2257

2176

2040

1980

1126

1036

398

1281

1807

324

1212

2322

5914

4270

5441

1718

237

691

1167

989

798

1427

1087

1226

1236

807

693

239

648

1240

160

829

1580

4355

3124

3581

7215

760

1731

2463

2572

1860

3314

2549

3104

4294

3888

2575

562

1509

3838

247

2058

3624

11,322

12,981

5263

77,729

Total

43,423

27,832

T A B L E   2  Multiple-wave voters in ANES data, 1972–2000

Year of 1st wave

Year of 2nd wave

Total respondents

Voted in both waves

Total votes in
Wave 1

Total votes
in Wave 2

1972

1974

1990

1992

1994

2000

2016

Grand total

1974

1976

1992

1994

1996

2002

2020

2705

475

1980

1126

1036

1807

4270

13,399

658

83

582

444

367

604

2032

4770

1644

189

902

763

560

1171

2858

8087

701

131

916

463

500

673

2298

5682

Generally speaking, do you usually think of yourself as a Democrat, a Republican,
an Independent, or what?

VCF0301  is  party  identification  on  a  Seven-Point  Scale  and  is  constructed  by  combining  the
VCF0302  question  with  one  of  two  follow-up  questions.  Respondents  who  identify  as  either
Republicans or Democrats in their initial response are asked a follow-up question:

REILLY anD HUnTInG

7

Would  you  call  yourself  a  strong  [Democrat/Republican]  or  a  not  very  strong
[Democrat/Republican]?

These responses form the two ends of the Seven-Point Scale, with Strong Democrats coded as 1,
Weak Democrats coded as 2, Weak Republicans coded as 6, and Strong Republicans as 7. A small
number  (92)  of  those  who  expressed  partisan  affiliation  in  VCF0302  were  coded  as  DK,  NA,
Other at this point in the survey.

Those  who  did  not  identify  as  either  Democrats  or  Republicans  are  given  this  follow-up

question:

Do you think of yourself as closer to the Republican Party or to the Democratic
Party?

These responses are used to construct the middle three categories of the Seven-Point Scale. Those
answering “Democratic” are assigned to Independent-Democrats (3), with “Republican” coded as
Independent-Republican (5). Respondents who choose “Neither” are Independent-Independents at
scale point 4. Note that although the above question is asked in the ANES time series surveys,
it  does  not  appear  in  the  CDF  data.  The  results  of  this  question  are  captured  in  CDF  item
VCF0301. Also note that a small number of respondents answered “do not know” or refused
to answer VCF0302 but indicated a party preference in the follow-up. These were moved to the
Independent-Republican and Independent-Democrat categories for VCF0301.

Finally,  the  Seven-Point  Scale  of  VCF0301  is  collapsed  to  three  categories  for  VCF0303.
This is done by combining the strong and weak Democrats with Independent-Democrats under
Democrats  (including  leaners)  and  ANES  political  identification  data  on  43,423  respondents
from 1972 through 2020 is shown on these three scales in Table 3. Independents make up 36%
of the total respondents over these surveys, with 13% identified as having no party leaning or
Independent-Independents.

RESULTS

Change in political identification over time—voters

We now use ANES data from multiple waves to look at how political identification changes for
voters over time. Between 1972 and 2020, ANES has data on 4770 respondents who voted in
two consecutive waves of the survey (Table 2). The following section looks at how these voters
changed  their  party  identification  from  one  survey  cycle  to  the  next.  A  small  portion  of  this
number did not report party identification on one of the three scales analyzed below.

Initial Party ID Response (VCF0302)

Of the 4745 respondents shown in Figure 1, 22% changed their initial-response party ID from
one wave to another. Independents were more fluid on this metric than party-affiliated respond-
ents: 16% of both Democrats and Republicans were changed at the second survey wave, while
36% of Independents changed.

In each major party, 13% of the wave-one respondents changed their position and identi-
fied as independents at wave two, while 3% went to the opposite party. Independents saw 17%
of  the  wave-one  respondents  move  to  Democrat  and  19%  move  to  Republican.  The  absolute
number of voters switching from party-affiliated to independent (426) and from independent to
party-affiliated (520) are similar, but the percentage of independents becoming party-affiliated

8

THE FLUID VOTER

T A B L E   3  Three political ID scales: 1972–2020 ANES data

Democrat

Republican

Independent

Other

DK/NA Total

VCF0302 (initial
response)

Percent

16,263

37%

10,976

25%

VCF0301 (7-Point

Scale)

Strong
Dem

Weak

Weak

Dem

Rep

Strong
Rep

12,772

29%

3085

7%

327

1%

43,423

Ind-Dem Ind-Ind Ind-Rep DK/NA Total

Percent

VCF0303

(Summary
3-Category)

Percent

8523

20%

7677

18%

5354

12%

5592

13%

5515

13%

5677

13%

4726

11%

359

1%

43,423

100%

Democrat (incl.
leaners)

21,715

50%

Republican (incl.

Independent

DK/NA Total

leaners)

15,672

36%

5677

13%

359

1%

43,423

100%

F I G U R E   1  Change in initial response (VCF0302) voters, 1972–2020

is nearly three times that of party-affiliates becoming independent. Respondents coded as Inde-
pendent, No Preference, and Other are included in the independent category. This flow is illus-
trated in Figure 1.

Seven-Point Scale (VCF0301)

When  the  Seven-Point  Scale  of  variable  VCF0301  is  analyzed  across  two  survey  waves,  more
fluidity in political identification becomes apparent. Of the 4735 respondents who voted in two
waves of the survey and recorded scores on this scale in both waves, 43% changed their identifi-
cation at the second wave. This is nearly double the 22% rate seen in the initial response variable
above.

Partisan respondents at the extremes of the scale were more consistent in their identification,
with 21% of Strong Democrats and 22% of Strong Republicans changing at the second wave.
Fifty-seven percent of Weak Democrats and 55% of Weak Republicans changed their identifica-
tion at wave two. Overall, 36% of Democrats and 36% of Republicans changed their party iden-
tification by at least one point on this scale between wave one and wave two; this is substantially
higher than the 36% of party-identified respondents who changed their position.

REILLY anD HUnTInG

9

F I G U R E   2  Change in 7-point scale (VCF0301) voters, 1972–2020

The  complex  flow  of  the  three  independent  classifications  is  shown  in  Figure 2.  As  noted
above, 36% of those reporting to be independent at the first wave interview were identifying with
one of the two parties at the second wave. In addition, another 22% of independents remained
independent at wave two, but shifted their position within the three independent categories on
this scale. Overall, 57% of those in one of the three independent categories of the Seven-Point
Scale changed their identification at wave two. This flow is illustrated in Figure 2.

Summary 3-Category Scale (VCF0303)

The collapsed categories of the Summary 3-Category Scale presented in VCF0303 necessarily
suppress  much  of  the  change  in  party  identification  seen  in  the  previous  two  scales.  By  this
measure,  10%  of  Democrats  (including  leaners)  and  12%  of  Republicans  (including  leaners)
changed their identification between the two survey waves. The independents represent just 7%
of respondents on this scale, as those who identify as independent but leaning toward a party are
grouped with their respective parties. The 328 remaining “true” independents fractured nearly
in thirds when queried at the second wave: 36% remained Independent while 30% switched to
Democrat and 34% to Republican. A total of 64% of the Independents from wave one were clas-
sified as partisans at wave two. Among those classified as either Democrats or Republicans at
wave one, just 11% changed their identification on this scale at wave two (Figure 3).

10

THE FLUID VOTER

F I G U R E   3  Change in Summary 3-Category Scale (VCF0303) voters, 1972–2020

F I G U R E   4  Change in initial response (VCF0302) non-voters, 1972–2020

Change in political identification over time—Non-voters

ANES  data  1972–2020  also  contain  information  on  political  identification  for  respondents
who  did  not  vote.  In  this  section  we  examine  changes  in  political  identification  in  the  15,292
respondents who answered two consecutive waves of the survey but voted in neither election.
For the purposes of this article, we use the term non-voters to identify those respondents who
did not vote in either wave of this analysis. It is unknown what portion of this population votes
on occasion.

Initial party response (VCF0302)

Of the 15,592 respondents shown in Figure 4, 60% changed their initial response party identifi-
cation from one wave to the next. This is nearly three times the 22% rate seen among those who
voted in both waves. Fifty-seven percent of the non-voting Democrat-identified respondents and

REILLY anD HUnTInG

11

69%  of  Republicans  changed  their  identification.  Taken  together,  62%  of  the  party-identified
non-voters changed their identification at wave two. This compares with 59% of the independ-
ents who changed.1

Of the 5201 who identified with Democrats at wave one, just 43% identified as such at wave
two, with 34% now seeing themselves as independents, and 23% crossing over to be Republicans.
This contrasts with just three percent of voting Democrats crossing over to the other party at
wave  two.  Non-voting  Republicans  showed  even  more  fluidity  between  the  two  survey  waves.
Just  31%  of  those  who  said  they  were  Republican  at  wave  one  maintained  that  identification
at wave two, with a third of the respondents switching to Democratic identification and 36%
now calling themselves Independent. Non-voting independents switched to Democrat in 36% of
the cases, and to Republican 23%, with 41% remaining independent. This flow is illustrated in
Figure 4.

Seven-Point Scale (VCF0301)

Non-voting respondents to ANES changed their political identification on the Seven-Point Scale
between wave one and wave two in 81% of the cases. This fluidity was relatively consistent across
the scale, ranging from 76% of Strong Democrats changing identification by at least one scale
point to 87% of Independent-Republicans.

Changes in party identification for the three independent categories are shown in Figure 5.
A total of 5923 (83%) of those who identified as one of the three independent categories at wave
one had moved by at least one scale point at wave two. Of this amount, 1722 (29%) remained
within the independent domain while 4201 (81%) moved to one of the parties. Changes in inde-
pendent non-voter identification are shown in Figure 5.

Summary 3-Category Scale (VCF0303)

When  non-voting  ANES  respondents  are  examined  on  the  summary  3-category  scale  of
VCF0303, 58% are seen to change categories. Democrats (including leaners) moved to another
point  on  the  scale  46%  of  the  time,  with  32%  identifying  as  Republicans  at  wave  2  and  13%
as independents. Well over half (60%) of the non-voters identified as Republicans at wave one
changed identification at wave two, with 47% later identifying as Democrat and 13% as inde-
pendent. Independent non-voters changed identification 81% of the time, with 48% moving to
Democrat at wave two and 33% to Republican. These flows are illustrated in Figure 6.

Voting patterns: Straight and split-ticket voters

ANES asks respondents to state how they voted in four contests: president, Congress, Senate7,
and governor. Respondents may not have the opportunity to vote in each of these races, depend-
ing on the timing of the election cycle. There are 27,832 respondents in the ANES data that said
they voted in at least one election, and these respondents give us information on a total of 77,729
races.

Of the ANES respondents who reported voting from 1972 to 2010, Figure 7 shows that 20,521
(73.7%) always voted a straight ticket for either Democrats (11,638 respondents) or Republicans
(8883). Conversely, 9316 respondents never voted for a Democrat and 12,201 never voted Repub-

1 Z-test for proportions <.01.

12

THE FLUID VOTER

F I G U R E   5  Change in 7-Point Scale (VCF0301) non-voters, 1972–2020

lican (Figure 7) in the races surveyed. Note that the 0% and 100% columns do not exactly mirror
each other due to the number of votes for minor parties.

The great majority of voters surveyed by ANES exclusively vote for one of the major parties,
with  only  a  small  percentage  splitting  their  votes  between  Republicans  and  Democrats.  The
40%–59% bracket in Figure 7 shows that 4655 people (16.7%) divided their votes evenly between
the two parties. Considering both the small number of election contests available for analysis
and the polarized nature of voting noted above, further analysis divides voters into three groups:
those  who  voted  for  Democrats  in  100%  of  the  contests,  those  who  voted  for  Republicans  in
100% of the contests, and those who voted for some mix of Democrats and Republicans. With
this information, we can see what portion of the sample consistently vote for one party and what
portion switch their votes between parties (Figure 7).

Initial Party ID Response (VCF0302)

As expected, Democrats generally vote a straight ticket for Democrats, at a rate of 74% while 71%
of Republicans, 5678 of the total 7990 Republican voters, always vote Republican (Figure 8).
Incongruously, 5% of Democrats and 5% of Republicans report that they always vote for the
opposite party.

Independents are much more evenly divided in their vote choices. A significant portion still
vote  straight  tickets  for  one  party  or  the  other,  with  34.8%  always  voting  for  Democrats  and
30%  always  voting  Republican.  A  plurality  of  independents  (35.2%)  split  their  votes  between
Democrats and Republicans at least occasionally. This compares with the combined figures for

REILLY anD HUnTInG

13

F I G U R E   6  Change in Summary 3-Category Scale (VCF0303) non-voters, 1972–2020

 12,000

12,101

 10,000

9,316

11,638

8,883

s
t
n
e
d
n
o
p
s
e
R
S
E
N
A

 8,000

 6,000

 4,000

 2,000

 -

1,882

2,086

2,327

2,328

2,061

1,869

165

191

443

374

0%

1%-19% 20%-39% 40%-59% 60%-79% 80%-99%

100%

Dem Votes

Rep Votes

F I G U R E   7  Percent of each respondent's votes by party, 1972–2020

Democrats and Republicans showing that 22% of the party-identified voters voted a mixed ticket
at least once.2 Respondents coded as Independent, No Preference, and Other are included in the
Independent category (Figure 8).

Seven-Point Scale (VCF0301)

With  the  Independent-Democrats  and  Independent-Republicans  broken  out  on  the
7-Point  Scale  of  VCF0301,  we  see  a  steady  progression  from  left-to-right,  with  decreas-
ing  Democratic  straight-ticket  voting  and  increasing  Republican  support  (Figure  9).  The
Independent-Independents at the middle of the scale have truly mixed voting choices: 31% only
voting for Democrats, 27% on voting for Republicans, and 43% choosing a mixture of Democrat
and Republican candidates. The votes reported by Independent-Independents represent 8% of
the 27,704 reported to ANES by respondents (Figure 9).

2 Z-test for proportions <.01.

14

THE FLUID VOTER

8,160

s
t
n
e
d
n
o
p
s
e
R
S
E
N
A

 9,000

 8,000

 7,000

 6,000

 5,000

 4,000

 3,000

 2,000

 1,000

 -

5,678

3,026

3,056

2,610

2,324

1,906

560

406

Democrat

Republican

Independent

Always Votes DEM

Mixed Votes

Always Votes REP

F I G U R E   8

Straight-ticket and mixed voting by Initial Party ID Response, 1972–2020

s
t
n
e
d
n
o
p
s
e
R
S
E
N
A

 6,000

 5,000

 4,000

 3,000

 2,000

 1,000

 -

Strong Dem Weak Dem Ind-Dem

Ind-Ind

Ind-Rep Weak Rep Strong Rep

Always vote DEM

Mixed Votes

Always vote REP

F I G U R E   9

Straight-ticket and mixed voting by 7-point scale, 1972–2010

Interestingly, Independent-Democrats show a higher percentage of straight-ticket support of
Democratic candidates (63%) than the Weak Democrats who explicitly declare support for the
party (60%).3

Summary 3-Category Scale (VCF0303)

With the Independent-Democrats and Independent-Republicans of VCF0301 included as “lean-
ers” in their respective partisan buckets for VCF0303, the Independent-Independents are high-
lighted (Figure 10). Combining the straight-ticket Democrat and Republican votes shows that

3 Z-test for proportions p < .05.

REILLY anD HUnTInG

15

s
t
n
e
d
n
o
p
s
e
R
S
E
N
A

 12,000

 10,000

 8,000

 6,000

 4,000

 2,000

 -

10,207

7,412

3,284

3,019

817

665

705

984

611

Democrat (incl. leaners)

Republican  (incl. leaners)

Independent

Always Votes DEM

Mixed Votes

Always Votes REP

F I G U R E   1 0

Straight-ticket and mixed voting by Summary 3-Category scale, 1972–2010

77% of the Democrat (including leaners) category always votes a straight ticket, 73% of Repub-
licans (including leaners), and 57% of independents (Figure 10).

Voting change over time

Changes  in  voting  behavior  for  ANES  two-wave  voters  are  summarized  in  Table  4.  The  1464
respondents reported voting a straight Democrat ticket both the first and second time they were
interviewed, with 1239 voting straight Republican both times, and 434 casting mixed votes at
each wave. This total of 3137 represents 66% of the total 4770 who voted in two waves of the
survey. This leaves 34% who altered their behavior across the election cycles.

The 4754 respondents who voted in two waves of the ANES survey and stated their political
identification are summarized in Table 5.4 Respondents who identified as Republicans or Demo-
crats as their Initial Party ID Response for VCF0302 are grouped together as Party Affiliated in
this table, for comparison to independent voters. Respondents coded as Independent, No Prefer-
ence, and Other are included in the independent category.

Party-affiliated  voters  exhibited  the  same  voting  behavior,  either  straight-ticket  or  mixed
voting  in  each  wave,  in  70%  of  the  cases.  For  Independents,  this  percentage  drops  to  57%.
Conversely, 30% of party affiliates changed their voting patterns across two elections, while 43%
of Independents changed.

Party-affiliated voters voted a straight ticket for the same party at a rate of 62%. Independ-
ents, while still voting consistently for a single party at a significant rate (44%), were still more
likely to change their voting behavior at the second wave of the survey. A small number of voters
voted a straight ticket for one party at Wave 1 and then switched to the other party at Wave 2.
Here we see that independents were twice as likely to make this large shift than party affiliates.
There is a small difference between the two groups on percentages that went from straight ticket
at Wave 1 to mixed at Wave 2. Independents were more likely to go from mixed to straight voting
and much more likely to vote for a mix of the two parties in both waves.

4 Sixteen of the 4770 ANES respondents who reported voting in two waves of the survey are coded as DK or NA, refused in VCF0302,
leaving 4754 who could have their political identification classified.

16

THE FLUID VOTER

T A B L E   4  Changes in voting behavior, all respondents

Wave 2

Wave 1

Straight REP

Mixed vote

Straight DEM

Total

Straight REP

Mixed Vote

Straight DEM

1239

399

55

1693

285

434

264

983

123

507

1464

2094

T A B L E   5  Voting changes over two waves of ANES data

Party affiliated

Independent

Initial Party ID Response (VCF0302)

Straight ticket Wave 1 & Wave 2: same party both waves*

Mixed ticket, both Wave 1 & Wave 2*

Subtotal: Same voting behavior in Wave 1 & 2*

Straight ticket Wave 1 & Wave 2: switched parties*

Straight ticket in Wave 1, mixed ticket in Wave 2**

Mixed ticket in Wave 1, straight ticket in Wave 2*

Subtotal: Different voting behavior in Wave 1 & 2*

n

2056

237

2293

96

355

547

998

%

62%

7%

70%

3%

11%

17%

30%

n

640

196

836

81

191

355

627

Total

1647

1340

1783

4770

%

44%

13%

57%

6%

13%

24%

43%

Total

*Z-test for proportions p < .01.
**Z-test for proportions p < .05.

Initial Party ID Response (VCF0302)

3291

100%

1463

100%

When the party-affiliated category shown in Table 5 is broken out to show the two parties, we see
the results are largely the same. Democrats continued their pattern from Wave 1, either voting
a straight ticket for a particular party or voting a mixed ticket, in 71% of the cases. This left
29% of self-identified Democrats changing their voting behavior in some fashion, with 32% of
Republicans exhibiting changed behavior. The p-value for a comparison of these percentages is
.176, suggesting that there is little meaningful difference between Democrats and Republicans on
this measure.

As noted above, 43% of independents showed changed behavior between Wave 1 and Wave 2,

with a p-value of <.01 compared to party-affiliated respondents (Table 6).

Seven-Point Scale (VCF0301)

When  this  same  measure  of  voting  change  over  time  is  applied  to  the  Seven-Point  Scale  of
VCF0301, we see that, unsurprisingly, the Strong Democrats and Strong Republicans are most
consistent in their behavior, with less than a quarter of these groups changing their voting profile
from one wave to the next (Table 7). Independent-Independents were the most likely to change their
voting  profile  at  48%,  followed  by  Independent-Republicans  at  45%.  Independent-Democrats
again seem to be as committed to Democratic candidates as the Weak Democrats with the party
in the question posed in VCF0301. Weak Democrats voted for a straight Democratic ticked at a

REILLY anD HUnTInG

T A B L E   6  Changes in voting behavior by Initial Party ID Response

VCF0302

Democrat

Republican

Subtotal: Party-identified

Independent, Other, No Pref.

Total

Voted in two waves

Percent that changed voting behavior

1803

1488

3291

1463

4754

29%

32%

30%

43%

34%

17

Percent that
did not change

71%

68%

70%

57%

66%

T A B L E   7  Changes in voting behavior by 7-Point Scale

VCF0301

Voted in two waves

Percent that changed voting behavior

Percent that did not change

Strong Democrat

1045

Weak Democrat

Independent-Democrat

Independent-Independent

Independent-Republican

Weak Republican

Strong Republican

Total

757

547

332

588

637

850

4756

22%

40%

38%

48%

45%

42%

24%

34%

78%

60%

62%

52%

55%

58%

76%

66%

rate of 45%, while Independent-Democrats did so at a rate of 47% with a p-value of .472 between
the two rates, indicating little actual difference between the two groups.

Summary 3-Category Scale (VCF0303)

The  Summary  3-Category  Scale  once  again  shrinks  the  size  of  the  independent  category  but
maximizes  its  difference  from  the  enlarged  party-identified  categories.  Democrats  (including
leaners) were slightly less likely to change their behavior (31%) than Republicans (including lean-
ers) at 35%.5 True independents, as above, changed their voting profile 48% of the time between
the first and second survey waves (Table 8).

DISCUSSION

Analyzing each of the three ANES measures of party affiliation (Initial Party ID, Seven-Point
Scale, and Summary 3 Category) over multiple elections provides some important findings on
the voting patterns of independents. We find evidence that, when tracking independent voting
behavior  over  more  than  one  election,  there  is  a  significant  volatility  in  voting  loyalty  and  as
a  group,  independents  are  distinct  from  partisans.  The  research  also  confirms  that  a  sizeable
number of independents move in and out of independent status from one election to another.

In  the  first  analysis  on  how  political  identification  of  voters  changes  over  multiple  survey
waves, we explored how these respondents changed their party identification from one survey
cycle to the next. The Initial Party ID scale found that independents were more fluid on this metric
than party-affiliated respondents with the percent of independents changing at the second wave

5 Z-test for proportions <.01.

18

THE FLUID VOTER

T A B L E   8  Changes in voting behavior by Summary 3-Category Scale

VCF0303

Voted in two waves

Percent that changed voting behavior

Democrat (including leaners)

Republican (including leaners)

Subtotal: Party-Identified

Independent

Total

2349

2075

4424

332

4756

31%

35%

33%

48%

34%

Percent that
did not change

69%

65%

67%

52%

66%

by more than double that of partisans. Additionally, the percentage of independents becoming
party affiliated was nearly three times that of party affiliates becoming independent. Moving to
the analysis of the Seven-Point Scale, even more fluidity was found with an overall 57% of those
in one of the three independent categories of the Seven-Point Scale changing their identification
at wave two. The last analysis collapsed categories of the Summary 3-Category Scale in party
identification seen in the previous two scales and resulted in a significantly reduced (7%) number
of independent respondents. Despite this small percentage of independent respondents, almost
two-thirds or 64% of the independents changed classification to partisans at Wave 2.

ANES  participants  who  responded  to  two  consecutive  waves  of  the  survey  but  voted  in
neither of the corresponding elections showed much greater fluidity in their political identifica-
tion between waves. Overall, 60% of respondents changed their Initial Party Identification. In
contrast with the voting respondents, non-voting Independents were substantially the same as
party affiliates, 59% making a switch to party affiliation. Non-voting independents were more
likely to move by at least one point on the Seven-Point Scale (64%) than either Democrats (50%)
or Republicans (49%). Thirty-nine percent of non-voters who identified as Independent in the
first  survey  wave  chose  party  affiliation  at  wave  two.  When  the  Summary  3-Category  Scale  is
analyzed,  we  see  that  81%  of  the  non-voting  independents  identified  with  one  of  the  major
parties at wave two, compared with 58% of all respondents.

Our analysis on how independent voters and non-voters changed their party identification
from one cycle to the next showed a significant amount of fluidity with non-voters being espe-
cially unpredictable. On all three political identification scales, independent respondents changed
their political identification over time more often than partisans (ranging from 36% to 64% for
voters  and  59%  to  83%  for  non-voters)  with  the  exception  of  non-voting  Republicans  on  the
Initial Party ID Scale. These findings suggest that independent voters and non-voters who iden-
tify in one political classification in one election are less likely to identify themselves in the same
manner in the next election. Their identification may depend on specific candidates or issues on
the ballot (Reilly et al., 2023) or may derive from short-term interest rather than a long-standing
loyalty (Miller, 1991). This finding supports Fiorina's (2016) assertion about independent voting
behavior: “whatever they are, they are an important component of the electoral instability that
characterizes the contemporary era. Their critical contribution to contemporary elections lies in
their volatility” (p. 10).

We next investigated how frequently respondents to the ANES scales vote “straight tickets,”
always  choosing  candidates  from  the  same  party  or  “mixed  tickets”  where  some  Republican
and some Democrat candidates are chosen. Our expectation that those identifying as Demo-
crat or Republican would mostly choose candidates from their own party was confirmed, while
independents demonstrated more variety in their choices. For the Initial Party ID scale, over
70%  of   partisans  voted  straight  ticket,  while  65%  of  independent  respondents  did,  with  the
independent  straight-ticket  voters  divided;  35%  voting  for  Democrats  and  30%  for  Republi-
cans.  While  the  Seven-Point  Scale  showed  similar  results,  Independent-Democrats  showed  a

REILLY anD HUnTInG

19

higher percentage of straight-ticket support of Democratic candidates than the Weak Demo-
crats, while Independent Republicans were slightly less likely to vote a straight ticket than Weak
Republicans.

Finally, we explored the degree to which individuals change their voting choices over time
(two-wave  voters).  Partisan  voters  exhibited  the  same  voting  behavior,  either  straight-ticket
or  mixed  voting  in  each  wave,  in  70%  of  the  cases.  For  independents,  this  percentage  drops
to  57%.  The  Initial  Party  ID  analysis  found  that  approximately  30%  of  partisans  changed
between  waves  compared  to  43%  of  independents;  while  the  Seven-Point  Scale  showed
Independent-Independents were the most likely to change their voting profile at 48%, followed
by Independent-Republicans at 45%. Weak Democrats voted for a straight Democratic ticket at
a rate of 45%, while Independent-Democrats did so at a rate of 47%. Independents were more
likely to go from mixed to straight voting and much more likely to vote for a mix of the two
parties in both waves. For the Summary 3-Category, approximately a third of partisans changed
their voting patterns over the two waves compared to 48% of independents.

Our research on independent voting behavior included analyses of voting patterns over time.
The findings confirm that independents do indeed move in and out of independent status when
tracked over multiple elections. Further, this study lends support to the notion that there is a
good deal more fluidity in voting patterns of independents. This was the case when analyzing
data  across  all  three  ANES  scales.  Our  analysis  of  the  ANES  Seven-Point  Scale  showed  that
Independent-Republicans  and  Weak-Republicans  resembled  each  other's  voting  patterns  on
straight/split ticket and voting change over time analyses (Mayer, 2008; Petrocik, 2009; Smith
et al., 1995); however, this was not the case with Independent-Democrats and Weak-Democrats.
Independent-Democrats, who did not affiliate with any party in the Initial Party Response ID
question, were more likely to vote only for Democrats than Weak-Democrats that specifically
identified  as  Democrat.  Similarly,  Independent-Republicans  were  as  likely  to  vote  a  straight
Republican ticket in both waves than Weak Republicans. Respondents who were coded as either
5 (Independent-Republican) or 6 (Weak Republican) showed essentially the same voting behav-
ior,  while  those  coded  as  3  (Independent-Democrat)  were  more  Democratic  in  their  behavior
than the supposedly more liberal Weak Democrats at scale point 2. This indicates that caution
needs to be exercised when treating the Seven-Point Scale (VCF0301) as a continuous variable.

It may be that traditional ways of measuring voter identification do not capture the independ-
ent voter due to voter composition of the electorate and our hyperpolarized political environ-
ment. Due to the research of Keith and others (1992), it became most common for researchers
with  the  ANES  to  utilize  a  three-point  or  five-point  scale  that  classified  independent-leaning
Democrats,  or  independent-leaning  Republicans,  as  partisans  (VCF0302,  Initial  Party  ID
response).  This  resulted  in  a  significant  reduction  in  the  number  of  self-described  independ-
ents. Reilly and Hedberg (2022) have argued that in light of the more recent work of Klar and
Krupnikov (2016) and Zschirnt (2011), which showed the importance of the independent iden-
tity, classifying self-identified independents as partisans seems counterproductive in examining
their influence on partisans, especially when respondents elected to self-identify as leaners. The
authors  collapsed  three  groups—respondents  who  selected  option  3,  4,  or  5—as  independ-
ent,  thus  treating  leaners  as  Independents  (VCF0303,  Summary  3-category  scale).  Similarly,
Fiorina (2016) has long been an opponent of classifying leaning independents as partisans and
leaving pure independents in the middle ID category arguing that “We can think of no other case
in political science where analysts change a respondent's explicit response to a survey item on
the basis of information from other items—especially one generally used as the dependent varia-
ble” (Abrams & Fiorina, 2011, p. 5). Perhaps it is time to develop new explanatory constructs to
capture independent voter classification.

20

CONCLUSION

THE FLUID VOTER

Our study contributes to previous literature on the independent voter by showing their voting
patterns  are  volatile,  unpredictable,  and  distinct  from  partisans.  Additionally,  when  analyz-
ing  voting  behavior  over  time,  our  research  confirmed  that  a  sizeable  number  of   independ-
ents move in and out of  independent status from one election to another. This volatility was
observed in all three measures of  party affiliation used by the ANES survey data. When inde-
pendents are followed over multiple elections, they have been found to have no firm partisan
loyalties.

Despite our contributions, our study has several limitations. First, as with any survey, the
voting classification and behavior details are all based on self-reports, which are suspectable to
response bias. Second, although ANES is a rich dataset with a long history to draw from, it does
have limitations for this sort of analysis. There is limited information about voter choices in the
data. The survey asks for party choices on just four races: president, Congress, Senate, and gover-
nor. With the survey waves spaced two years apart (except for the 2016–2020 waves), respondents
will not be able to provide answers to presidential, senatorial, and most governor's races in both
waves, which limits the data available for analysis. These four races, especially at the presidential
level,  are  susceptible  to  a  “celebrity  effect”  where  a  high-profile  candidate's  perceived  charm
(or repulsiveness) may overwhelm a voter's policy-based preferences when selecting a candidate.
Data that included more frequent and down-ballot races would provide a better picture of the
relationship between the stated political identification of voters and their voting choices (Bitzer
et al., 2021). Finally, it is also difficult to draw solid conclusions about the behavior of non-voters
from these data. These non-voters may be latent voters who generally lie dormant but turn out
at the polls when there is an issue or candidate that particularly motivates them. Without a very
long time-series survey, it is difficult to say with what frequency these latent voters are activated
or what motivates changes in their political identification.

There  is  a  lot  that  still  needs  to  be  learned  about  this  emerging  group  of  voters.  Future
research should explore the fluidity of Black and Latino voters as well as the increasing genera-
tional divide. Most importantly, there is a need to continue to track independent voting behavior
over time and more analysis on the voting patterns of independents is needed down ballot at the
state and local level.

ORCID
Thom Reilly

  https://orcid.org/0000-0001-8614-0482

REFERENCES
Abrams, Samuel J., and Morris P. Fiorina. 2011. “Are Leaning Independents Deluded or Dishonest Weak Partisans?”

https://cise.luiss.it/cise/wp-content/uploads/2011/10/Are-Leaners-Partisans.pdf.

ANES Data Center. 2015. “Party Identification 3-Point Scale (Revised in 2008) 1952-2012.” American National Election

Studies. https://electionstudies.org/resources/anes-guide/top-tables/?id=22.

ANES Data Center. 2021. “American National Election Studies.” https://electionstudies.org/data-center/.
ANES Data Center. n.d. “Time Series Cumulative Data File.” American National Election Studies. https://electionstud-

ies.org/data-center/anes-time-series-cumulative-data-file/.

Bitzer, J. Michael, Christopher A. Cooper, Whitney Ross Manzo, and Susan Roberts. 2021, November 4–5. “The Rise
of the Unaffiliated Voter in North Carolina.” Prepared for Presentation at the State if the Parties 2020 and Beyond
Virtual Conference. Ray C. Bliss Institute of Applied Politics, University of Akron.

Bitzer,  J.  Michael,  Christopher  A.  Cooper,  Whitney  Ross  Manzo,  and  Susan  Roberts.  2022.  “Growing  and  Distinct:
The Unaffiliated Voter as Unmoored Voter.” Social Science Quarterly 103(7): 1587–601. https://doi.org/10.1111/
ssqu.13225

Brody, Richard A. 1978. “Change and Stability in the Components of Partisan Identification” Paper Prepared for the

NES Conference on Party Identification.

REILLY anD HUnTInG

21

Brody, Richard A. 1991. “Stability and Change in Party Identification: Presidential Off-Years.” In Reasoning and Choice,
edited by Paul A. Sniderman, Philip E. Tetlock, and Richard A. Brody, 179–205. Cambridge: Cambridge Univer-
sity Press.

Brody, Richard A., and Lawrence S. Rothenberg. 1988. “The Instability of Partisanship: An Analysis of the 1980 Pres-

idential Election.” British Journal of Political Science 18(4): 445–65. https://doi.org/10.1017/S0007123400005214.

Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American Voter. Hoboken,

NJ: Wiley.

DeBell, Matthew. 2010. How to Analyze ANES Survey Data. ANES Technical Report Series no.nes012492. Palo Alto,

CA, and Ann Arbor, MI: Stanford University and the University of Michigan.

Fiorina, Morris P. 1977. “An Outline for a Model of Party Choice.” American Journal of Political Science 21(3): 601–25.

https://doi.org/10.2307/2110583.

Fiorina, Morris P. 2016. “Independents: The Marginal Members of an Electoral Coalition.” Washington, DC: Hoover

Institution. https://www.hoover.org/research/independents-marginal-members-electoral-coalition.

Fiorina,  Morris  P.  2017.  Unstable  Majorities:  Polarization,  Party  Sorting,  and  Political  Stalemate.  Washington,  DC:

Hoover Institution Press.

Gallup. 2022. “Party Affiliation.” Gallup Polls. https://news.gallup.com/poll/15370/party-affiliation.aspx.
Gruber, Jeremy, and John Opdycke. 2020. “The Next Great Migration: The Rise of Independent Voters.” Open Prima-

ries. https://openprimarieseducationfund.org/wp-content/uploads/2020/11/ROI_Report_R1.pdf.

Ingraham, Christopher. 2021. “How to Fix Democracy: More Beyond the Two-Party System, Experts Say.” The Wash-

ington Post. https://www.washingtonpost.com/business/2021/03/01/break-up-two-party-system/.

Jackson,  John  E.  1975.  “Issues,  Party  Choices,  and  Presidential  Votes.”  American  Journal  of  Political  Science  19(2):

161–85. https://doi.org/10.2307/2110431.

Keith, Bruce E., David B. Magleby, Candice J. Nelson, Elizabeth Orr, Mark C. Westlye, and Raymond E. Wolfinger.

1992. The Myth of the Independent Voter. Oakland, CA: University of California Press.

Klar, Samara. 2014. “Partisanship in a Social Setting.” American Journal of Political Science 58(3): 687–704. https://doi.

org/10.1111/ajps.12087.

Klar, Samara, and Yanna Krupnikov. 2016. Independent Politics: How American Disdain for Parties Leads to Political

Inaction. New York: Cambridge University Press.

Magleby, David, Candice Nelson, and Mark Westlye. 2011. “The Myth of the Indepedent Voter Revisited.” In Facing the
Challenge of Democracy: Explorations in the Analysis of Public Opinion and Political Particiaption, edited by Paul
M. Sniderman and Benjamin Highton, pp. 238–63. Princeton, NJ: Princeton University Press.

Mayer, William G. 2008. The Swing Voter in American Politics. Washington, DC: Brookings Institution Press.
Miller, Warren E. 1991. “Party Identification, Realignment and Party Voting: Back to Basics.” American Political Science

Review 85(2): 557–68. https://doi.org/10.2307/1963175.

Page, Benjamin I., and Calvin C. Jones. 1979. “Reciprocal Effects of Policy Preferences, Party Loyalties, and the Vote.”

The American Political Science Review 73(4): 1071–89. https://doi.org/10.2307/1953990.

Petrocik, John R. 2009. “Measuring Party Support: Leaners Are Not Independents.” Electoral Studies 28(4): 562–72.

https://doi.org/10.1016/j.electstud.2009.05.022.

PEW Research Center. 2019. Political Independents: Who They Are, What They Think. U.S. Politics & Policy. https://

www.pewresearch.org/politics/2019/03/14/political-independents-who-they-are-what-they-think/.

PEW Research Center. 2022. Americans’ Views of Government: Decades of Distrust, Enduring Support for Its Role U.S.
Politics & Policy. https://www.pewresearch.org/politics/2022/06/06/americans-views-of-government-decades-of-dist
rust-enduring-support-for-its-role/.

Reilly, Thom, and E. C. Hedberg. 2022. “Social Networks of Independents and Partisans: Are Independents a Moderat-

ing Force?” Politics & Policy 50(2): 225–43. https://doi.org/10.1111/polp.12460.

Reilly, Thom, Jacqueline S. Salit, and Omar H. Ali. 2023. The Independent Voter. London: Routledge.
Sides, John. 2013. “Three Myths about Political Independents.” The Monkey Cage. https://themonkeycage.org/2009/12/

three_myths_about_political_in/.

Smith, Andrew E., Alfred J. Tuchfarber, Eric W. Rademacher, and Stephen E. Bennett. 1995. “Partisan Leaners Are NOT

Independents.” The Public Perspective. https://ropercenter.cornell.edu/sites/default/files/2018-07/66009.pdf.

Teixiera, Ruy. 2012. “The Great Illusion.” The New Republic. https://newrepublic.com/article/100799/swing-vote-untapp

ed-power-independents-linda-killian.

Zschirnt, Simon. 2011. “The Origins & Meaning of Liberal/Conservative Self-Identifications Revisited.” Political Behav-

ior 33(4): 685–701. https://doi.org/10.1007/s11109-010-9145-6.

22

THE FLUID VOTER

AUT HOR BI OG RAP HIES

Thom Reilly is a Professor and Co-Director for the Center for an Independent and Sustaina-
ble Democracy in the School of Public Affairs at Arizona State University. He is the former
Chancellor  of  the  Nevada  System  of  Higher  Education  and  County  Manager  for  Clark
County,  Nevada.  Reilly's  research  focuses  on  public  pay  and  benefit  schemes,  nonpartisan
governance,  the  independent  voter,  and  child  welfare.  He  is  the  author  of  several  books
including The Independent Voter (Routledge Press, 2023) with co-authors Jacqueline Salit and
Omar Ali, The Failure of Governance in Bell, California (Lexington Press, 2016), and Rethink-
ing Public Sector Compensation (M.E. Sharpe, 2012).

Dan Hunting is Senior Researcher at Arizona State University's Lodestar Center for Philan-
thropy  &  Nonprofit  Innovation.  His  research  interests  include  economic  impacts  of  the
nonprofit  sector,  voter  dynamics,  workforce  development,  education  funding,  and  urban
growth. Hunting has authored several foundational works that have informed Arizona policy
discussions including Finding & Keeping: Educators for Arizona's Classrooms an analysis
of the state's teacher shortage, and Sun Corridor: A Competitive Mindset (co-authored with
Grady Gammage, Jr.) which described the complex connections between the economies of
Phoenix and Tucson.

How to cite this article: Reilly, Thom, and Dan Hunting. 2023. “The fluid voter:
Exploring independent voting patterns over time.” Politics & Policy 00: 1–22. https://doi.
org/10.1111/polp.12517.