Resurgent Mass Partisanship: The Role of Elite Polarization
Author(s): Marc J. Hetherington
Source: The American Political Science Review, Vol. 95, No. 3, (Sep., 2001), pp. 619-631
Published by: American Political Science Association
Stable URL: http://www.jstor.org/stable/3118237
Accessed: 21/05/2008 11:29
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=apsa.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].
http://www.jstor.org
Vol. 95, No. 3
American Political Science Review
,
Role
of Elite
The
Mass
Partisanship:
MARC J. HETHERINGTON Bowdoin College
Resurgent
F
September 2001
Polarization
orthe most part, scholarswho studyAmericanpoliticalparties in the electoratecontinueto
characterizethem as weak and in decline.Partieson the elite level,however,have experienceda
resurgenceoverthe last two decades.Such a divergencebetweenelitebehaviorand mass opinionis
curious,giventhatmostmodelsofpublicopinionplace thebehaviorof elitesat theircore.Infact, Ifind that
overthelasttwodecades.Greaterpartisan
a noteworthy
resurgence
partiesin theelectoratehaveexperienced
the
has
in
Americans,whichin turn
for ordinary
polarization Congress clarified parties'ideologicalpositions
has increasedpartyimportanceand salienceon the mass level.Althoughpartiesin the 1990s are not as
centralto Americansas theywerein the 1950s,theyarefar moreimportanttodaythan in the 1970sand
1980s. Thepartydeclinethesisis in need of revision.
W
ith fewexceptions(see Keithet al. 1992),the
scholarlyconsensuson contemporaryAmerican politicalparties in the electorate centers
on party decline. There is disagreement about its
sources-whether people are more negative (Nie,
Verba, and Petrocik 1979) or neutral (Wattenberg
1984)-and aboutits abruptness-whetherprecipitous
(Wattenberg1984) or less steep but still meaningful
(Konda and Sigelman 1987)-but the conventional
wisdom is that parties have long been irrelevantto
many. Bartels (2000) cites a litany of scholarlywork
that suggestspartydecline in the electoratewill persist
into the new century.1Accordingto some, that trendis
potentiallyirreversiblebecause of the antipartyelectoral changes implemented in the 1960s and 1970s
(e.g., Aldrich 1995, 245-53; Beck 1997, 385).
The centralityof party decline in the thinking of
public opinion scholarsis curious, especiallybecause
Congressscholarsdiscoveredyearsago that partiesare
resurgenton the elite level (e.g., Poole and Rosenthal
1997; Rohde 1991). Since most theories of public
opinion changefocus on the behaviorof elites (Brody
1991; Carminesand Stimson 1989;Page and Shapiro
1992; Zaller 1992), party resurgence in Congress
should be consequentialin understandingmass attitudes towardparties.Mass behaviorshould reflect, at
least to some degree, elite behavior.Therefore,mass
party strength should have increased as a result of
greaterpartisanshipat the elite level.
I will demonstratethat the measuresscholarshave
used as evidence of mass party decline now point to
party resurgence.In most cases the movement has
been extraordinary,especially in view of the glacial
pace characteristicof most public opinion change.
Moreover,it can best be explainedby the increasein
ideological polarization along congressional party
lines. I will show that elite polarizationhas clarified
public perceptions of the parties' ideological differences, which has led to a resurgenceof partiesin the
electorate.
INDIVIDUAL-LEVEL
EVIDENCEFOR PARTY
DECLINE
Scholarshave detailedpartydecline usingdata at both
the aggregateand individuallevel, but I confine my
analysis to the latter, using data collected by the
National Election Study (NES). To public opinion
scholars,the most familiarevidenceof partydecline is
the rapid increase in political independenceand the
accompanyingdecreasein strongpartisanshipafterthe
1950s. The percentageof independentleaners nearly
doubledbetween1960and 1980,and the percentageof
strong partisans dipped by more than one-third
(Wattenberg1984).2One prominentexplanationfor
party decline is that, in a candidate-centeredera,
parties have become irrelevantto many people. As
evidence,Wattenberg(1984) cites a rapid increasein
the percentageof Americanswho are neutraltoward
both parties,as tappedby likes/dislikesquestionsin the
NES survey.
AlthoughKonda and Sigelman(1987) expressconcerns about Wattenberg'smeasures,they find further
support,albeit muted, for the neutralitythesis. They
measured party engagement as the total number of
party likes and dislikes providedby respondentsand
MarcJ. Hetheringtonis AssistantProfessorof Government,Bowdiscoveredthat engagementdeclined substantiallybedoin College,Brunswick,ME 04011 ([email protected]).
tween 1952and 1984.In laterwork,Wattenberg(1994,
I thankSuzanneGlobetti,in particular,for her helpfulcomments
andPaulFreedman,BruceLarson,RobertLuskin,JayMason,Keith
1996, 1998) focuses on the Perot candidacies and
Poole, Daron Shaw,and JonathanWeilerfor their sound advice.I
arguesthat the partiesare still in decline.An apparent
also wishto recognizethe top-notchresearchassistanceprovidedby
of major party failure is the fact that Ross
indicator
Rob FordandJoshHelfatas well as the computingexpertiseof Eric
Perot receivedmore votes in 1992than any thirdparty
ConsorChown.The data were obtainedfrom the Inter-university
tiumfor Politicaland SocialResearch,whichbearsno responsibility candidatesince Theodore Roosevelt eightyyears earfor their use.
1 Bartels (2000) demonstratesthat party identificationhas an increasinglylargeeffecton presidentialand congressionalvote choice
and is largelyalone in challengingthe partydeclinethesis,although
other publicopinionscholarsnote the resurgenceon the elite level
(e.g., Aldrich1995;Beck 1997).
2
Bartels (2000) finds that these trends have reversed recently,
especially among the politicallyactive. The proportionof party
identifiersamongvoterswashigherin 1996thanin anyelectionsince
1964.
619
Elite PolarizationSeptember
Mass Partisanship:
Partisanship:The
The Role
ResurgentMass
Role of Elite
Polarization
Resurgent
2001
September 2001
FIGURE1. Feelings about the Parties, Likes-Dislikes, and Feeling Thermometers, 1952-96
60
50
40
-
..C*,~\
e
y?
-3--30
---
Pos-Neg (Likes-Dislikes)
- * --Pos-Neg (Feeling Thermometers)
- Z - -Neu-Neu (Feeling Thermometers)
/
.~-
/yn~~
Neu-Neu (Likes-Dislikes)
20
x .-x
xo.
-X~
*x.X..
-x.
10
0
!,
,
,
i
i
i
i
,
,
'Ix
,
1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996
Year
lier. In addition,split-ticketvotingreachedan apexfor
the NES survey era. Of course, Perot's historically
strong showing absent congressional Reform Party
candidatespotentiallyexplains the increase in ticket
splitting,although this phenomenon is often considered a symptom of party decline (e.g., Beck 1997;
Keefe 1998).
INDIVIDUAL-LEVEL
EVIDENCEOF PARTY
RESURGENCE
Because the conventionalwisdomhas a stronghold, I
must demonstratethat a resurgenceof party at the
mass level has occurred.I do so by relyingon manyof
the measures developed in the 1980s to show party
decline. Key to Wattenberg's(1984) argumentis an
increase in the percentage of people with neutral
feelings towardboth parties, as measuredby the net
numberof likes and dislikes offered by NES respondents. If people give more reasons for liking a party
than dislikingit, they are consideredpositive toward
thatparty.If theyprovidemore dislikesthanlikes,they
are considerednegative.An equal numberof likes and
dislikesor no responsesat all indicateneutrality.
The solid lines in Figure1 trackchangesin the most
partisanand most neutralcategories.The percentage
of those neutral toward both parties declined by 6
points between 1980 and 1996, and the percentageof
those positive towardone party and negative toward
the other increased by the same amount. Positivenegative replaced neutral-neutralas the modal category in 1988 and continuedas the mode throughthe
rest of the time series. Although the proportionof
positive-negativesin 1996 does not approachthat of
1952, a movementtowardgreaterpartisanshipis still
evident.
620
This use of the likes-dislikesmeasure has several
problems.In additionto obscuringdifferencesbetween
categoriesand overstatingneutrality(see DeSart 1995;
Kondaand Sigelman1987;Stangaand Sheffield1987),
the measure lacks a stated neutral point. People are
classified as neutral if they unwittinglybalance the
numberof likes and dislikes or, perhapsmore problematically,provide no answers at all. Feeling thermometers are more attractivebecause they have an
explicit neutral point, 50 degrees, and almost all respondentsprovidevalid answers(Craig1985).3
I can use the NES partythermometersto constructa
measure of affect similar to Wattenberg's.I classify
those who answer50 degreesto both thermometersas
neutral-neutral,those who answerabove50 degreesto
one party and below 50 degrees to the other as
positive-negative,and so forth. The results appearas
the broken lines in Figure 1. The most noteworthy
Only
findingis the recentupsurgein positive-negatives.
about35%fell into this most partisancategoryin 1980,
comparedto nearlyhalf in 1996,which representsan
increaseof 40%.
Konda and Sigelman(1987) measurepartyengagement as the total number of likes and dislikes that
respondentsprovideaboutthe parties.Figure2 reveals
thatpartiesare far moresalientin the 1990sthanin the
1970s and 1980s; the mean number of responses in
1996 was higher than in any year except 1952 and 1968.
By this measure,the salienceof partyhas increasedby
45% since 1980.
The NES changed the phrasing of the party thermometer questions
in 1978. Previously, respondents were asked how they felt about
"Republicans" and "Democrats." Subsequently, they have been
asked about the "Republican Party" and the "Democratic Party." In
1980, the NES asked both versions, and the difference in means was
quite large. Therefore, I cannot extend the analysis back any farther.
3
American Political Science Review
Poiia
eiwVl
Vol. 95, No. 3
cec
5o
AmericanI~~~
I
FIGURE2. Mean Total Number of Likes and Dislikes about the Parties, 1952-96
5
4.5
4
4
c
n
4U
>.92
3.5
3.5
3.4
3.32
\
3
3
.2
2.99
2.5
.
l
l
l.
1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996
Year
Another indicator of party-centric attitudes is
straight ticket voting. Wattenberg(1994) notes that
votingfor a presidentialcandidateand House member
of differentpartiesreacheda high for the NES era in
1992, but Figure 3 demonstratesthat 1992 was an
anomaly. Even if third party voters are included, a
higherpercentageof Americansreportedvoting for a
presidentialand House candidateof the same partyin
1996thanin anyyearsince 1964.The tendencyis much
cleareramongmajorpartypresidentialvoters.Straight
ticketvotingfor presidentand House in this grouphas
increasedprogressivelysince 1980.Even in 1992,when
partisanshipwas supposedlyat a low point, Bush and
Clintonvoterswere more inclinedto vote for the same
party in their House election than were Bush and
Dukakisor Reagan and Mondalevoters.4
Although these data suggest a dramaticresurgence
in party, Perot's historicallylarge vote share in 1992
appearsto suggestthe opposite. Scholarshave shown,
however, that such factors as the third party candidate's personalcharacteristics(Rosenstone,Behr, and
Lazarus 1996) and respondents'trust in government
(Hetherington1999) better explainthird partyvoting
than does strengthof partisanship.Moreover,Perot's
4
2.71
It can be arguedthat thirdpartypresidentialvoters are the least
partisanandhence leastlikelyto vote a straightticket.Therefore,by
focusingonlyon majorpartypresidentialvoters,I mayoverstatethe
increasein partyloyalty.If thiswere true,however,a higherlevel of
straightticketvoting amongmajorpartypresidentialvoters should
haveoccurredin 1992,whenPerotreceived19%of the vote, thanin
1996,whenhe receivedjust 9%.Instead,of majorpartypresidential
votersin the twoyears,78%votedfor a Housecandidateof the same
partyin 1992,whereas82%didso in 1996.If the focuson majorparty
voters merelyeliminatedthose who mighthave been least likelyto
vote a straightticket had Perot not run, I would have found more
straightticketvotersin 1992than 1996becausetwiceas manyof the
supposedlyleast partisanthirdpartyvoterswere eliminated.
personalfortuneallowedhim to overcomemanyof the
handicaps-such as ballot access laws, small advertising budgets,and dismissivenews reporting-faced by
most third party candidacies(Rosenstone, Behr, and
Lazarus1996). In short, Perot's showingresultedprimarilyfrom factorsother than partydecline.
WHATCAUSES MASS OPINIONTO
CHANGE?
Mass opinion in the aggregatetends to move glacially
if at all (Page and Shapiro1992). When it does move,
it usually responds to changes in the information
environmentprovidedby elites. Althoughthe authors
of TheAmericanVoterpartiallyblame cognitivelimitations for Americans'lack of ideologicalsophistication,
they also recognizedthe importanceof elite-levelcues:
"Thereare periodsin whichthe heat of partisandebate
slackens and becomes almost perfunctory,and the
positionsof the partiesbecome relativelyindistincton
basic issues. In times such as these, even the person
sensitiveto a range of political philosophiesmay not
feel this knowledgeto be helpful in an evaluationof
current politics" (Campbell et al. 1960, 256). V.O.
Key's (1966) echo chamberanalogy further suggests
that elite behavior will set the terms by which the
masses think about politics (see also Nie, Verba, and
Petrocik1979;Page 1978).If politiciansprovidepartyoriented or issue-orientedcues, then the public will
respond in a party-centricor issue-centricmanner.
They are unlikelyto do so withoutsuch cues.
Indeed, the most sophisticatedrecent theories of
public opinion place elite behavior at the center of
individualopinion change (Brody1991;Carminesand
Stimson 1989; Zaller 1992). For example, Carmines
621
Resurgent Mass Partisanship: The Role of Elite Polarization
i
FIGURE 3.
Percentage
September 2001
of Voters Casting a Straight Ticket for President and House, 1952-96
90
85
80
\
='^
o 75
o)
--
\
.
.'
f
\
\
/
'\.
^
>-
,'
-U-
.
,,,,
.' \
/
.'
--All
Presidential
Voters
MajorParty
Presidential
Voters
70
70
'"
?
''
'
65
60
l
,
, . .,
1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996
Year
and Stimson (1989) identify changes in the behavior of
Republican and Democratic elites as the engine for an
issue evolution on race in the 1960s. Similarly, Brody
(1991) argues that we can best understand presidential
approval by observing elite behavior. He maintains that
elite consensus generally predicts higher approval ratings, and elite division usually means lower approval
(see also Mermin 1999).
Zaller (1992, 311), who develops the connection
between elite behavior and mass opinion most completely, concludes that even those most attentive to
politics "respond to new issues mainly on the basis of
the partisanship and ideology of the elite sources in the
messages." If people are exposed to a heavily partisan
stream of information, which will be more likely if
elites are behaving in a partisan manner, then it follows
that respondents will express opinions that reflect the
heavily partisan stream. Because greater ideological
differences between the parties on the elite level should
produce a more partisan information stream, elite
polarization should produce a more partisan mass
response.
PARTYRESURGENCEON THE ELITE
LEVEL
One measure that taps changing elite behavior is Poole
and Rosenthal's DW-NOMINATE scores for members of Congress.5 These scores allow for both between-member and between-year comparisons. An increasing ideological distance between Democratic and
Republican elites should produce a more partisan
5 DW-NOMINATE scores are the most commonly used estimate of
the ideological position of members of Congress. Members' ideal
points are derived using a dynamic, weighted, nominal three-step
estimation procedure based on all nonunanimous roll call votes
taken in each Congress (see Poole and Rosenthal 1997 for details).
622
information environment for ordinary Americans, especially in view of the media's well-known bias toward
framing politics in terms of conflict (Graber 1997, chap
4).
To measure party polarization in the House, I divided members by party, calculated the mean DWNOMINATE score on each dimension for each caucus, and calculated the weighted Euclidean distance
between them.6 Figure 4 tracks the distance between
the House Republican and Democratic caucuses from
the 81st Congress, which began a few days into 1949, to
the 104th Congress, which ended a few days into 1997.
Polarization declined steadily from the late 1940s into
the late 1960s and remained relatively constant until
the late 1970s, a trough that coincides with the decline
of party in the electorate. Congressional behavior then
changed in the late 1970s. With the 95th Congress,
ideological polarization between the parties began a
steady rise.7
These changes in congressional behavior correspond
closely with, but slightly precede, the increases in
measures of mass partisanship described above. For
example, half the growth in elite polarization occurred
6 Because Poole and Rosenthal
compute the coordinates with a
weighted utility model, any use of the DW-NOMINATE scores to
calculate a distance requires that the second dimension be weighted
by .3 (see Poole and Rosenthal 1997). Scores for the Senate,
although less polarized, follow much the same pattern as those for
the House (Poole 1998).
7 Scholars
suggest several alternatives to the measure of polarization
used here. For example, Aldrich, Berger, and Rohde (1999) employ
a number of measures of both polarization and homogeneity,
including the median distance between the parties, the intraparty
homogeneity along the NOMINATE score's first dimension, and the
proportion of members of one party who ideologically overlap the
other party on the first dimension. It is worth noting that the mean
Euclidean distance measure that I employ here is correlated with
these three measures at .99, .99, and .97, respectively, when I use data
from the 85th to the 103d Congress.
Vol. 95, No. 3
I
P
Review
American
Politicall Science
m
I
I
4. Mean EuclideanDistance between Republicanand DemocraticPartyDW-NOMINATE
FIGURE
Scores, House of Representatives,1949-97
,
0.3
,
Lc
c 0c,oRc ,o
ao
,
,-c o--~oc~ oo
~,-~
,
c
sb6e 0Cs~~~b
8 Bp?~~~~~`
8' 90
--c
00
cao, 0
,i>, c toA'm
o,
0co
, to-0
, ,
,-
0c
0J
c0 o~,-<,i>-~
,~ ooOK
o,o/
l9b 9 sb~~~~~~~~~~'~~
ce* of
q
8
9i'l 9o9f ,o
betweenthe last Congressin the Carteradministration
and the first Congressin the second Reagan administration. The start of the substantialincreases in the
total numberof likes and dislikesand the trendtoward
straightticket voting took hold in the election cycles
that followed.On the heels of the second spike in elite
polarization,which occurredduringthe second Congress of the Bush administration(the 102d), all measuresof partystrengthrespondedin kind,with the rise
in the percentageof positive-negativesthe most dramatic. Elite polarization,therefore, appears to be a
potentialengine for change at the mass level.
INCREASEDCLARITYOF PARTYIMAGES
AS AN INTERMEDIATE
STEP
I place elite polarizationat the heartof the explanation
for party resurgenceand hypothesizea set of causal
dynamicsbetween elites and ordinaryAmericanssimilar to those posited and demonstratedby Carmines
and Stimson (1989, 160) regardingracial issue positions. More partisanelite behaviorcausedby polarization shouldclarifypartypositionsfor the public,which
in turnshouldinfluencethe importanceand salienceof
parties.8
One way to test whether clarity has increased is
simply to ask people whether they see important
differencesbetween what the parties represent.The
NES does so, and the solid line in Figure5 tracksthis
trend. From 1960 to 1976, the percentagewho per-
8 In
exploringthe influenceof party activistson party ideologies,
Aldrich (1995, chap. 6) employedvariationson several of these
measuresof clarityandreachedresultsconsistentwithmine.He does
not suggest,however,that greaterclarityreinvigoratedpartisanship
at the masslevel.
,
0<,-*
0
f
6O
ceived importantdifferencesrangedfrom the high 40s
to the middle50s.A markedupturnbeganin 1980,and
58% or more have seen importantdifferencesevery
year since. The percentagereached63% in 1996, the
highestlevel in the series.9
The "important differences" responses suggest
greater clarity has occurred but not why. Since the
polarizationin Congresshas been ideological,ideological differencesare a likelyreason.One measureof this
is whetherthe publiccan arraythe partiescorrectlyon
a liberal-conservativescale. If ideological clarity has
increased,then people should be both better able to
place the Democratsto the left of the Republicansand
more likelyto perceivea largerdistancebetweenthem.
The broken line in Figure 5 demonstrates that
people in the 1990sare better able to arraythe parties
ideologically.From 1984until 1990,only about50%of
the publicdid so correctly,but this figurereached63%
in 1996.10In additionto arrayingthe partiescorrectly,
respondentsperceive a widening ideological gulf between them. Accordingto data from the NES Cumu9 Wattenberg(1990) identifiesbut dismissesthis trend,noting that
peoplearenot also moreinclinedto thinkone of the partiesis better
able to solve their importantproblems.A potentialexplanationfor
this contradictionis that people perceivegreaterpartypolarization
but are not necessarilyenthusiasticaboutit (Dionne 1991;Hibbing
andTheiss-Morse1995).In that sense, people maythinkthe parties
will do a differentjob, not an idealjob.
10I can make safe comparisonsonly
startingin 1984.Before then,
respondentswho refused to place themselves on an ideological
continuumor said they did not knowwere not asked to place the
parties.Beginningin 1984,a follow-upquestionaskedsuch respondents "if they had to choose,"what theywouldconsiderthemselves
to be. Onlythose who refusedthe followup as well were not asked
to place the parties,whichreducedmissingdata by more than half
between1980and 1984.
623
Mas Partisanship:
Patsnhp
Th Role
Roeoof Elite
lt Polarization
oaiainSpebr20
Reugn Mass
The
Resurgent
September 2001
FIGURE 5. Percentage of Respondents Who Perceive Important Differences between the Parties
and Correctly Place the Parties Ideologically, 1960-96
65
60
~~~~~-
?*g=~~~~~~~~/.;
2 55
0Q)
/
*See
.
*
/
---'
ImportantDifferences
CorrectlyPlaces Democrats
to Leftof Republicans
50
,
45
1960
1964
1968
1972
1976
,
1980
1984
,
1988
1992
1996
Year
lative File (Sapiroet al. 1997), the mean signed ideological distance between the parties rose from 1.52
points in 1984 to 1.94 points in 1992 and 1996, an
increaseof 28%.11I use the signedratherthanabsolute
ideologicaldistancebecause elite polarizationshould
also help people arraythe partiescorrectly.The use of
absolute distance would make equivalentthe placement of Democratsone unit to the left or to the right
of Republicans,which would obscure the increasing
proportionof correctplacements.
Similarto the patternrevealedby the indicatorsof
partyresurgence,the increasesin partyclarityoccurred
soon after increasesin elite polarization.Apparently,
as partyelites beganto clarifyideologicalcues, citizens
became less inclinedto see the partiesas Tweedledee
andTweedledum.Whenpeople perceivethatwho wins
and loses will lead to distinct futures, they should
develop more partisan feelings and become more
inclinedto organizepoliticsin partisanterms.
WHYPARTYCLARITYHAS INCREASED
As parties in Congresshave become more polarized
alongpartylines,people havebecomemore inclinedto
1 To
compare data from 1996 with other years in the NES Cumulative File, the NES provides a weight to correct for a too highly
educated 1996 sample. I do this for all other descriptive analyses, but
I use the unweighted data for perceived ideological distance because
the weighting overcorrects due to a rapid decline in missing data for
this item over time. Specifically, only 9.5% of cases are missing in
1996, compared with 20% or more in preceding years. If I employ the
weight, it reduces the mean for education (2.48) among those with
valid responses in 1996 below the 1984 mean (2.52). Among all
weighted respondents, however, the mean for education rose a
statistically significant .09 points between 1984 and 1996. Even
without weighting the data for perceived ideological distance, the
mean for education in 1984 (2.52) is still too high relative to that of
1996 (2.54), which likely accounts for the leveling off of perceived
ideological distance between 1992 and 1996.
624
see importantdifferencesbetween the parties, place
them correctlyin an ideologicalspace, and perceivea
wider ideological distance between them. I will test
whether there is a causal connection between elite
polarizationand these mass responses.
I employa pooled cross-sectionaldesign,using data
gatheredby the NES in both presidentialand off-year
elections between 1960 and 1996. This design allows
me to mergecontextualinformationovertime,namely,
the aggregatemeasuresof ideologicalpolarizationin
the House, with the surveydata. Due to data limitations, I am often confinedto the seven surveystaken
between1984and 1996.The resultsare consistent,and
often stronger, when I include appropriatedummy
variablesto confrontthese data limitationsand extend
the analysisback farther.
The firstdependentvariableis whethera respondent
Sees ImportantDifferencesbetween what the parties
represent.It is coded 1 if the respondentclaimsto see
importantdifferences,0 otherwise.The second dependent variableis CorrectIdeologicalViewof the Parties,
which is coded 1 if the respondentplaces the Democratic Party to the left of the Republican Party, 0
otherwise.12The third dependentvariableis Perceived
IdeologicalDistancebetweenthe Parties.It is measured
as the signed differencebetween where respondents
place the Republicanand Democraticparties on the
scale.
NES's seven-pointliberal-conservative
These three measures should be a function of a
number of different attitudinaland contextualvariables. Most importantfor my purposesis Elite Polarization.I tap this as the mean Euclideandistancein the
DW-NOMINATEscoresbetweenthe Democraticand
12 To conserve
cases, I include in the analysis both those who placed
themselves ideologically and those who did not. This means that all
those who failed to place themselves and thus were not asked to
place the parties are coded 0.
American
PoliticalScience
Science Review
Review
American Political
RepublicanHouse caucuses,which I lag by one Congress for two reasons.First, a Congressofficiallyends
after most postelectionsurveyshave been completed,
so using a contemporaneousterm would suggestthat,
for example, congressionalbehavior in 1993 affects
1992 attitudes,which makes no temporalsense. Second, time elapses before the publicperceiveschanges
in elite behavior.Publicopinion on race, for instance,
did not react immediatelyto the parties' change in
position (Carminesand Stimson1989).In mergingthe
contextualwith the individual-leveldata, I give each
1996respondentthe meanEuclideandistancefromthe
1993-95sessionof Congress,each 1994respondentthe
mean Euclidean distance from the 1991-93 session,
and so forth.13
A numberof attitudinalmeasuresalso mayaffectthe
dependent variables,so they are added as controls.
Strengthof Ideologyand Strengthof Partisanshipshould
playimportantroles.Those who place themselvesnear
the poles of the seven-point scales demonstratean
understandingof ideologyand partisanshipand hence
should be more inclinedto see differencesthan those
who place themselvesat mid-scale.In addition,several
social characteristicsare relevant. Those with more
Educationwill be less inclined to provide mid-scale
responses than those with less (Delli Carpini and
Keeter 1996),whichincreasesboth the probabilitythat
they will see differencesbetween the parties and the
extent of distancethey see. Age should have a similar
effect; older respondents, who have more political
experiencethanyoungerones, shouldbe more inclined
to see importantdifferences,arraythe partiescorrectly,
and see a widergulf betweenthem.In contrast,Women
and AfricanAmericansexhibit less political expertise
than males and whites, respectively(see e.g., Mondak
1999), so they should be less likely to see party
differences.14
I also must account for contextualfactors. People
Elections,so
pay less attentionandvote less in Off-Year
they should tend to see the parties as less distinctin
nonpresidentialyears. Years characterizedby Divided
Governmentmay make a difference.On the one hand,
both parties have a prominentvoice in government,
which may increase people's ability to identify the
parties ideologically.On the other hand, people may
have a hardertime decidingwhethera presidentof one
partyor a Congressof the other is drivingthe ideolog13
I specifya model in whichcausationrunsfrom the elite to mass
level,but Rohde(1991)suggeststhe reverse.Ourgoalsdiffer.Rohde
exploreschangesin the directionof white southerners'partyidentification,whereasmy concernis the strengthof partisanattitudes.It
is more likely that mass-level strength intensified over time in
responseto a more partisanelite environmentthan that a sudden,
unexplainedinfluxof strongerpartisansin the electoratepavedthe
wayfor the likes of Newt Gingrichand Jim Wright.Moreover,even
the directionalchangesin southernpartisanshipdescribedby Rohde
wererootedin elite behaviorchangeson civilrightsissues(Carmines
and Stimson1989).
14 I would have liked to include an objectivemeasureof political
knowledge,but the NES did not provide a consistentbatteryof
knowledgeitems until 1988. Scholarsoften use formal education,
whichI includein the model,as a proxyfor politicalknowledge(e.g.,
Sniderman,Brody,and Tetlock 1991).
No. 3
Vol.
Vol. 95, No.
ical direction of the country.Indeed, in 1990, fewer
thanhalf of Americanscould even identifywhichparty
controlledthe House, despite the fact that the Democratshad done so for nearly40 years,so twovoices may
only serve to confuse citizens.
In sum, I estimate the followingmodels to analyze
the influenceof elite polarizationon three measuresof
mass-levelclarityabout the parties.
Pr(sees importantdifferences)
= f(elite polarization,strengthof partisanship,
education,age, black,female, off-year
election, dividedgovernment).15
(1)
Pr(correctlyplaces the partiesideologically)
= f(elite polarization,strengthof ideology,
strengthof partisanship,education,age,
black,female, off-yearelection,
dividedgovernment).
(2)
PerceivedIdeologicalDistance
= f(elite polarization,strengthof ideology,
strengthof partisanship,education,age,
black,female, off-yearelection, divided
(3)
government).
The first two dependentvariablesare binary,so ordinary least-squares (OLS) estimates will be biased.
Hence, I use logistic regression to estimate these
models and use OLS to estimatethe third.
The resultsin the firstcolumnof Table 1 suggestthat
elite polarizationhas a significanteffect on whether
people see importantdifferencesbetweenthe parties.16
In fact, all variablesperformas expected, except for
age, which is insignificant,and race, which is positive.
That the Democratshave been much friendlierto the
interests of African Americansappears to matter to
this groupin identifyingimportantdifferences.
Achieving statistical significance in a sample of
19,000 is no great feat. More important,the effect of
elite polarizationis substantivelyimportantas well. If I
accountfor the 1960contextof dividedgovernmentin
a presidentialyear and set the other variablesto their
1960meanvalues,the predictedprobabilityof seeing a
differencebetween the partiesis .512, which is almost
identical to the 50.9% of respondentswho reported
seeing a difference in 1960. If I hold all variables
constantat their 1960means,againaccountfor divided
governmentand election context, but increase elite
polarizationto its 1996level, the predictedprobability
of seeing importantdifferencesrises to .611, an in15 Becausethe NES
only startedto ask people to place themselves
ideologicallyin 1972,I dropstrengthof ideologyfromthe important
differencesequation,so I canincludedatafrom1960,1964,and 1968.
Droppingstrengthof ideologyshouldnot affectthe resultsunduly,
giventhat the partialcorrelationbetweenelite polarizationand the
percentagewho see importantdifferencesbetweenthe partiesis an
extremelyrobust.87, controllingfor off-yearelectionyears.
16 Becausemy measureof elite
polarizationis not independentfrom
year to year, there may be some concern about autocorrelation.
Regressiondiagnosticsrevealedno suchproblems.For instance,the
Durbin-Watsonstatisticsfor the OLS models presentedbelow are
1.99 and 1.97, respectively,which indicates not even a hint of
autocorrelation.
625
September
2001
September 2001
Elite Polarization
Polarization
The Role
Role of Elite
Mass Partisanship:
Partisanship:The
ResurgentMass
Resurgent
TABLE1. Perceptionsof IdeologicalClarityand Measures of PartyStrengthas a Functionof
ElitePolarization,PoliticalAttitudes,Social Characteristics,and ContextualFactors
Variable
Elitepolarization
S(I)
Sees
Important
Differences
1960-96
Param.Est.
(Std. Err.)
1.707***
(0.189)
(VII)
(V)(VI)
/n/.)(IV)
TotalPartyLikes
Respondent Is
(11s)
Dems.
Places
(III)
and Dislikes
a Positive-Negative
Perceived
to Left
1984-96
1984-96
of Reps.
IdeologicalDist.
1984-96
FullModel Reduced Form FullModel Reduced Form
1984-96
Param.Est.
Param.Est. Param.Est. Param.Est. Param.Est. Param.Est.
(Std. Err.) (Std. Err.) (Std. Err.) (Std. Err.)
(Std. Err.)
(Std. Err.)
2.144**
2.918***
2.017***
3.103***
3.087***
2.065***
(0.722)
(0.815)
(0.491)
(0.468)
(0.434)
(0.444)
Perceived ideological
0.280***
--0.136***
distance
(0.009)
(0.015)
Education
0.383***
(0.016)
0.803***
(0.022)
0.626***
(0.023)
0.206***
(0.023)
0.266***
(0.021)
0.928***
(0.037)
1.101***
(0.031)
Strengthof partisanship
0.469***
(0.016)
0.268***
(0.020)
0.135***
(0.024)
0.606***
(0.024)
0.648***
(0.022)
0.466***
(0.037)
0.601***
(0.031)
0.610***
(0.023)
0.461***
(0.025)
0.336***
(0.024)
0.374***
(0.022)
0.475***
(0.039)
0.683***
(0.035)
Strengthof ideology
Race (AfricanAmerican)
0.186***
(0.049)
-0.458***
(0.058)
0.052***
(0.011)
-0.569***
(0.069)
-0.125
(0.067)
-0.112*
(0.059)
-0.035**
(0.013)
-0.016
(0.011)
-0.225*
(0.110)
0.166***
(0.021)
-0.309***
(0.092)
0.174***
(0.017)
Age
-0.006
(0.009)
Sex (female)
-0.305***
(0.031)
-0.310***
(0.039)
-0.074
(0.044)
-0.036
(0.042)
-0.067
(0.039)
-0.803***
(0.069)
-0.862***
(0.060)
Dividedgovernment
-0.138***
(0.035)
-0.551 ***
(0.040)
-0.069
(0.071)
-0.200*
(0.078)
-0.194**
(0.075)
-0.203**
(0.069)
-0.334**
(0.126)
-0.351**
(0.111)
-0.176***
(0.045)
-0.290***
(0.051)
-0.232***
(0.049)
-0.272***
(0.045)
-0.224*
(0.092)
-0.240**
(0.079)
-2.531***
(0.107)
1956.75***
-4.518***
(0.313)
3555.78***
-2.332***
(0.347)
-5.138***
-2.105***
-5.001***
(0.565)
(0.310)
(0.337)
1863.49*** 2012.54***
Off-yearelection
Intercept
X2
Adjusted R2
-.13
SEE
-2.321-
0.086"**
(0.013)
14,109
Numberof cases
11,394
19,206
Election
National
Source:American
File,1948-96.
Studies,Cumulative
11,255
13,405
-3.509***
(0.496)
-
.22
.23
3.239
3.139
8,967
11,191
*p < .05, **p< .01, ***p< .001; one-tailedtests.
crease of .099, and is not much different from the
63.0%who reportedseeing a differencein 1996.
Among the other independentvariables,between
1960 and 1996 only education changed such that it
would increase the probabilityof seeing important
differencesbetweenthe parties.In a simulationsimilar
to the one above, holding elite polarizationand all
other variablesat their 1960 means, and accounting
only for the increasein education,the predictedprobability of seeing importantdifferencesrises by .058.
Although both factors are important,increasedelite
polarizationis better than 50% more importantthan
increasededucationin explainingchange over time.
The resultsin the second columnof Table 1 suggest
that elite polarizationhas clarifiedmassperceptionsof
Betweenthe 97th
the parties'ideologicaldifferences.17
17 The same
pattern of results were obtained when I reestimated the
models by successively dropping each year from the analysis, which
suggests that the results are not a function of a single observation. In
626
and 103d Congress,the mean Euclideandistancebetween partymembers'DW-NOMINATEscoresin the
House grew from .530 to .690. When I increasedelite
polarizationby this amount,set dividedgovernmentto
one, set off-yearelection to zero, and held all other
variablesconstantat their 1984 means, the predicted
probabilitythat a respondentwouldcorrectlyplace the
DemocraticPartyto the left of the RepublicanParty
rose from .540 to .620. The increase of 8 percentage
points caused by elite polarizationbetween 1984 and
1996 likely accounts for a large portion of the 9.6
percentagepoint changethat actuallyoccurred.
addition, the results are not time bound. When I estimated a model
using data back to 1972, the first year the ideology questions were
asked, and included a dummy variable for pre-1984 cases to account
for the differing response rates to the ideology questions, the effect of
elite polarization remained statistically significant (13= 1.329, p <
.001) and was substantively even larger, taking into account elite
polarization's greater range over the longer period.
American Political Science Review
Across their ranges, variables such as education,
strengthof ideology, and strengthof partisanshipall
have larger effects than elite polarization,but none
increasedby as much as 5% between 1984 and 1996.
Indeed,only the increasesin educationand strengthof
ideology were statisticallysignificant.When I performed parallelsimulationsfor these two variables,I
found that the predicted probabilityof arrayingthe
partiescorrectlyrose by a paltry.018 and .011 points,
respectively.
The same pattern of results emerges in explaining
perceivedideologicaldistancebetweenthe parties,and
the effect of elite polarizationis again substantively
These resultsappearin the thirdcolumn
important.18
of Table 1. Multiplyingthe parameterestimateby the
.16-pointincreasebetween 1984and 1996producesan
increaseof .323 in perceivedideologicaldistance.The
dependentvariableincreasedby .4 between 1984 and
1996,and greaterelite polarizationaccountsfor about
80% of the change,other thingsbeing equal.
For the thirdequation,the attitudinalvariablesand
social characteristicsperformedas expected. It is important to note, however, that only education and
strength of ideology rose significantlybetween 1984
and 1996. Multiplyingtheir respectiveparameterestimates by their differences in means provides their
contributionto the increase of .4. These calculations
yield increases of .094 and .059, respectively.Both
effectspale in comparisonto that of elite polarization.
EXPLAININGPARTYRESURGENCE
As Carminesand Stimson (1989) would predict, the
results thus far suggest that elite polarization has
clarifiedpublic perceptionsof the parties'ideological
positions.What differencedoes greaterclaritymake?
Some suggestthat perceptionsof polarizedpartiesmay
cause dissatisfaction(e.g., Dionne 1991;Fiorina1996;
Hibbing and Theiss-Morse 1995; King 1997), but I
contend that greaterideologicalclarityshould invigorate partisanattitudes.When people believe that parties providechoices not echoes (Key 1966;Nie, Verba,
and Petrocik1979;Page 1978), they ought to be more
concernedaboutwho dictatespublicpolicy.As people
come to realize that Democratsand Republicanswill
pursue substantiallydifferentcourses, attachmentto
one side or the otherbecomesmoreconsequential,and
partyimage becomes more salient (see also Carmines
and Stimson 1989).
To test the influenceof elite polarizationon party
affect and salience, I estimated models for two measures of partyresurgence:whethersomeone is a Positive-Negative
using the partyfeeling thermometersand
the TotalNumberof PartyLikes and Dislikesprovided
by a respondent. I used the same right-handside
variablesas above for the same reasons,but I added
18 WhenI estimateda model
usingdatabackto 1972and includeda
dummyvariablefor pre-1984cases, the effect of elite polarization
remainedstatisticallysignificant(3 = 2.035,p < .001) and was
substantivelymuch largerthan that estimatedfrom the 1984 and
latercases.
Vol. 95, No. 3
the thirdmeasureof partyclarity,PerceivedIdeological
Distance,to the right-handside as well. If people see
sharperdistinctionsbetweenthe parties,partiesshould
be more importantand salientto them. This specification allowselite polarizationto haveboth a directeffect
on partyaffectand an indirecteffectthroughperceived
ideologicaldistance.
In functionalform, the models are as follows.
Pr(Respondentis a Positive-Negative)
= f(elite polarization,strengthof ideology,
strengthof partisanship,education,age,
black,female, off-yearelection, divided
government,perceivedideologicaldistance). (4)
Total Numberof PartyLikes and Dislikes
= f(elite polarization,strengthof ideology,
strengthof partisanship,education,age,
black,female, off-yearelection, divided
government,perceivedideologicaldistance). (5)
Again, the first dependent variable is binary,which
prompts the use of logistic regression. The second
dependentvariableis an intervalscale, which allows
the use of OLS.
The resultsappearin the fourthand sixthcolumnsof
Table 1. Elite polarizationis again positivelysigned
and statisticallysignificantfor both dependent variables. Its effect, moreover, is substantial. When I
increasedelite polarizationfrom its 1984 to its 1996
level, accountedfor dividedgovernmentand a presidential election year, and held all other variables
constant at their 1984 mean values, the predicted
probabilitythata respondentwillbe a positive-negative
rose from .325 to .442. In addition,elite polarization
has an indirect effect through perceived ideological
distance. Recall that elite polarizationexpands the
distancebetweenthe partiesby .323.When I increased
perceivedideologicaldistanceby this amountaboveits
1984mean, the probabilityof givingresponsescategorizedas positive-negativerose by another.010,bringing
the total effect of elite polarizationto .127, ceteris
paribus. Again, the effect of more education and
stronger ideology did not approachthat of growing
elite polarization.
The same pattern of results emerges for the total
number of likes and dislikes mentioned about the
parties,as shown in the sixth column of Table 1. The
parameterestimateof 2.144 for elite polarizationsuggests that its increase of .16 between 1984 and 1996
caused an estimated increase of .343 points in the
number of likes and dislikes mentioned.Accounting
for the rise of the .323 points in perceivedideological
distancecausedby elite polarizationadds another.090
points to the dependentvariable.Thus, the total effect
of elite polarizationis .433, or more than 60% of the
increaseof .72 in the dependentvariable.In comparison, the total effectscausedby a rise in educationand
strengthof ideology between 1984 and 1996 are each
less than one-quarterof that of elite polarization.
A largepercentageof respondentsdo not answerthe
ideologicalself-placementquestionand,as a result,are
not asked to place the parties,so includingperceived
627
Mass Partisanship: The Role of Elite Polarization
ResurgentReugnI
It
IoSetm
Ip IThIoeo
IM
ideological distance on the right-handside creates a
good deal of missingdata. This may undermineconfidence in the results,especiallysince those missingtend
to be less sophisticatedthan those who provide responses.I thereforeestimatedreducedformsfor equations 4 and 5, droppingperceivedideologicaldistance
from the models.Its effect shouldbe capturedby elite
polarization,which allows the recapture of several
thousandlost cases (Markus1988). The resultsin the
fifth and seventhcolumnsof Table 1 demonstratethat
elite polarizationremainssignificantwith or without
the missingdata.19Taken together, these results suggest that increasinglystrong partisanorientationson
the mass level are a function of growingideological
polarizationon the elite level.
RECEPTIONOF ELITEPOLARIZATION
Althoughthe resultsthus far providestrongevidence
that elite ideologicalpolarizationhas produceda more
partisanelectorate, an even more rigoroustest is to
account for people's differingability to absorb this
information.A more ideologically polarized House
should produce a more ideologicallypolarized issue
environment,but those with more political expertise
should reflectit better than those with less.
Many suggest that a measure of objectivepolitical
knowledgeis the best indicatorof political expertise
(e.g., Delli Carpiniand Keeter 1996;Price and Zaller
1993;Zaller 1992).Unfortunately,the NES onlybegan
askinga detailedbatteryof factualquestionsin 1988.
The use of educationas a proxyfor knowledgeis not
ideal (see Luskin 1987), but many have done so (see
e.g., Popkin 1994; Sniderman,Brody, and Tetlock
1991).Indeed,people with more educationshould,on
average,have better developed cognitivetools, which
should allow them to absorbmore political information. If a knowledge battery were available over a
sufficientlylong period,I wouldexpectthe resultsto be
even strongerthan those presentedbelow.20
I replicatedeach of the full modelsfromTable 1 and
introducedan interactionbetween elite polarization
and education. Education is coded 1 for those who
completedgrade school, 2 for those who attendedor
graduatedfrom high school, 3 for those who attended
college, and 4 for college graduatesand those with
graduatedegrees.The interactionshould carrya positive sign, which indicatesthat those who can best use
the informationgeneratedby a more polarizedenvi-
Septemberr 2001
I
ronment are the most inclined to see differencesbetween the partiesand providemore partisanopinions.
Since the effectsof each variablenot includedin the
interactionsare almostidenticalto those presentedin
Table 1, I includeonly the estimatesfor the interaction
andits componentpartsin Table2. In all fivecases,the
interactionis properlysigned and statisticallysignificant.To interpretthe interaction,I calculatedthe total
effectof elite polarizationfor the perceivedideological
distance and the total number of likes and dislikes.
Although the pattern is the same for the three logit
models, the interpretationof the nonadditive OLS
The total effectof elite
modelsis more straightforward.
polarizationis derivedas follows:
EElite Polarization= P1 +
33(Educationi),
(6)
where EElite Polarization is the total effect of elite polarization,and Educationiis the ith respondent'slevel of
education.In the perceivedideologicaldistanceequation, the estimatefor 1, is -1.209 and for f3 is 1.271,
and in the likes/dislikesequation,P1 is -4.209 and 33
is 2.504.
Given these estimatedeffectsfor perceivedideological distance,elite polarizationhas no effect(E = .062)
on those who completedgradeschool (education= 1),
and its effect for those with a high school education
(education = 2) is a relativelysmall 1.333. For those
with at least some college (education= 3) and college
graduates (education = 4), however, the effect is
substantial:2.604 and 3.875, respectively.The results
for the total numberof likes and dislikesare similar.
That is, elite polarizationhas no effect on people with
less cognitive training,but among college attendees
and graduates,the effect is a hefty 3.303 and 5.807,
respectively.
In sum,these resultsshouldincreaseconfidencethat
elite polarizationis drivingthe impressiveincreasein
party-centricthinkingon the mass level. People with
the greatestabilityto assimilatenew information,those
with more formaleducation,are most affectedby elite
polarization.These findingsare all the moreimpressive
in view of the fact that education is not an optimal
proxyfor politicalknowledge.21
CONCLUSION
The resultspresentedhere suggestthat partiesin the
electoratehave reboundedsignificantlysince 1980and
that the partydecline thesis is in need of revision(see
also Bartels2000).Althoughthe environmentdoes not
mirror
that of the 1950s,Americansin the 1990s are
19There are roughly2,200 more cases for the positive-negative
more
to thinkaboutone partypositivelyand one
likely
equationthanfor the likes/dislikesequation.This is largelybecause
negatively, less likely to feel neutral toward either
only half sampleswere asked the likes/dislikesquestionsin 1986,
1990,and 1996.In addition,I estimatedan equationusingdataback
party,and better able to list why they like and dislike
to 1952, droppingboth ideological polarizationand strength of
the parties than they were ten to thirtyyears ago.22
= 2.180, <
remained
Elite
p
significant(P
polarization
ideology.
.001), whichsuggeststhat the resultsare not time bound.
20 The NES has askedone factualitem in almosteverysurveysince
1960:whichpartycontrolledthe House of Representativespriorto
the election.When I specifiedan interactionbetweenwhetherthe
respondentansweredthis questioncorrectlyand elite polarization,
the same basic patternof resultsemerged.Becausethis is a single
item and one that manyare likelyto get rightby guessing(Luskin
n.d.), I opted to use educationinstead.
628
results for the interactiveterms were replicatedwhen I
includeda dummyvariablefor pre-1984cases and extended the
analysesfor partyplacementandperceivedideologicaldistanceback
to 1972.The resultsfor the total numberof likes and dislikesalso
were replicatedwhen I includeddata backto either 1972or 1952.
22
Althoughthe resultsare not presentedhere due to spaceconsiderations, I found that elite polarizationalso has increased the
21 The
Vol. 95, No. 3
American Political Science Review
TABLE2. Replication of Models in Table 1, Adding an Interaction between Elite Polarization and
Education
Variable
Elitepolarization
Education
Elitepolarization
x Education
Sees
Important
Differences,
Parameter
Estimate
(StandardError)
0.068
(0.408)
0.061
(0.086)
0.620***
(0.164)
Places
Democratsto
Leftof
Republicans,
Parameter
Estimate
(StandardError)
-0.261
(1.304)
0.273
(0.311)
0.886*
(0.520)
Perceived
Ideological
Distance,
Parameter
Estimate
(StandardError)
-1.209
(1.485)
0.135
(0.331)
1.271*
(0.552)
Respondent
Is a PositiveNegative,
Parameter
Estimate
(StandardError)
-0.094
(1.448)
-0.537*
(0.320)
1.239**
(0.532)
Total Party
Likes and
Dislikes,
Parameter
Estimate
(StandardError)
-4.209*
(2.280)
-0.570
(0.504)
2.504***
(0.839)
Source: AmericanNationalElectionStudies, CumulativeFile, 1948-96.
*p < .05, **p< .01; ***p< .001; one-tailedtests.
Consistentwith most theories of publicopinion,these
mass-levelchangeshave resultedfrom changesin elite
behavior.Greaterideologicalpolarizationin Congress
has clarified public perceptions of party ideology,
which has produceda more partisanelectorate.
AlthoughI have focusedon the strengthas opposed
to direction of partisanship,the results of this study
mayhave implicationsfor the latteras well. In discusssome suggest that such shorting macropartisanship,
terminfluencesas changesin economicconditionsand
presidentialapprovalratingshave profoundeffectson
the distributionof Republicansand Democrats (e.g.,
Erikson,MacKuen,and Stimson1998;MacKuen,Erikson, and Stimson 1989), whereas others argue that
their effects are minimal(e.g., Green, Palmquist,and
Shickler1998).Because strengthof partisanshipin the
aggregate has fluctuated markedlyover the last 50
years, heterogeneityin the time series is likely. When
people hold their partisan ties more intensely, the
probabilityof party identificationchange is reduced.
Hence, the effect of short-termforces on macropartisanship should be smallerwhen strengthof partisanship is relativelyhigh and largerwhen it is relatively
when people perceive that the parties representdistinct ideologies, which allowsvoters to make rational
calculations about alternative futures (e.g., Downs
1957;HinichandMunger1994).Mystudysuggeststhat
voters are much better able to make such ideological
distinctionsthan in the past. On average,partisanship
allows less sophisticatedAmericansto connect their
valuesandinterestswithvote choice (Delli Carpiniand
Keeter 1996), so voters should be able to participate
more effectivelyas a result.
The election of 2000 provides further empirical
evidence of mass partyresurgence.The proportionof
Republican and Democratic voters is nearly equal
today (Miller 1998), so a national election strongly
influencedby partyshouldbe veryclose, andAmerican
electionsdo not get muchcloserthan the 2000 contest.
Preliminarydatafromthe 2000 NES suggestthat more
than 90% of both Democraticand Republicanidentifiers voted for their party's presidential candidate
(Burnset al. 2001), and partisansof everystripewere
significantlymore loyalto theirparty'sstandard-bearer
in 2000thanin either 1992or 1996(Pomper2001, 138),
which also were two highlypartisanelections (Bartels
low.23
2000). In addition,straightticket voting for president
The resurgenceof partyis, of course,good news for
and House remainedabove 80% among majorparty
those who trumpetthe unique role that partiestradi- presidentialvoters (Burnset al. 2001),whichproduced
tionally have played in organizing political conflict a razor-thinRepublicanmajorityin the House that
(e.g., Shattschneider1975). Voting theorieswork best
reflects the closeness of the presidentialrace. Public
reactionto the election was also stronglypartisan.In
tendencyof respondentsto vote for presidentialand House candidates of the same party, controllingfor strengthof partisanship, earlyDecember,more than 90% of Republicanscriticized Al Gore's legal efforts in Florida (Saad 2000),
strengthof ideology,race, age, sex, whethera House seat is open,
whethera Houseseat is contested,whethera thirdpartypresidential despite clear indicationsthat a pluralityof Floridians
candidateis running,and whetherthe House incumbentis of the
intendedto vote for him (Bradyet al. 2001), and more
partyoppositethe respondent'spartyidentification.
than 80% of Democrats approved of Gore's legal
23 Green,
Palmquist,andShickler(2000)note thatthe proportionof
challenge(Saad 2000).
Democraticidentifiersdid not rise substantiallyin the 1990s,despite
It is easy to overlook a party resurgence when
increasesin consumerconfidenceandconsistentlyhigh
extraordinary
presidentialapprovalratings,whichcastsfurtherdoubton the import
symptomsoften associatedwith weak parties,such as
of short-termforces. They argue that this may be the result of
third party candidaciesand divided government,are
questionableestimationdecisionsmade by MacKuen,Erikson,and
features of the political environment.Such
regular
Stimson (1989), but it also may have become harder to move
in the 1990sbecausestrengthof partisanshiphas
phenomena, however, result from other factors in
macropartisanship
addition to weak partisanship.Ross Perot certainly
grown.
629
oeo of Elite
lt Polarization
asPriasi:
h Role
The
Resurgent MassReugn
Partisanship:
oaiaionSptmer20
benefited from party independence, but his success was
mostly a function of his personal style and fortune
(Rosenstone, Behr, and Lazarus 1996). To the extent
that weak parties do advantage third party efforts,
moreover, it is noteworthy that third parties have
drawn progressively fewer votes over the last three
presidential elections, a period when my results suggest
that partisanship in the electorate has strengthened.
In addition, strong parties do not automatically
produce unified government (Fiorina 1992). Although
the late nineteenth century was America's most partisan era, divided governments were the norm. Today,
candidate quality and fundraising play a dominant role
in understanding which voters split their tickets (Burden and Kimball 1999). If either party gains a significant identification advantage among regular voters,
unified government will almost certainly result. Until
then, we are likely to see close presidential elections
and small majorities for one or the other party in both
houses of Congress.
APPENDIX A. QUESTION WORDING
Partisanship
Generallyspeaking,do you usually think of yourself as a
Republican, a Democrat, an Independent, or what? (If
Republicanor Democrat.)Wouldyou call yourselfa strong
(Republican/Democrat)or a not very strong (Republican/
Democrat)?(If independent,other, or no preference:)Do
you thinkof yourselfas closer to the Republicanor DemocraticParty?
Ideology Questions
We hear a lot of talk these days about liberalsand conservatives.I'm going to showyou (1996:Here is) a seven-point
scale on whichthe politicalviewsthat people mighthold are
arrangedfrom extremelyliberal to extremelyconservative.
Wherewouldyou placeyourselfon this scale, or haven'tyou
thoughtmuch about this?
Wherewouldyou place the DemocraticParty?
Wherewouldyou place the RepublicanParty?
Sees Important Differences
September 2001
Likes/Dislikes
Is there anything in particularthat you like about the
Democratic Party? What is that?
Anything else [you like about the Democratic Party]? Up
to five mentions.
Is there anything in particular that you dislike about the
Democratic Party? What is that?
Anythingelse [you dislike about the DemocraticParty]?
Up to five mentions.
Is there anything in particular that you like about the
Republican Party? What is that?
Anything else [you like about the Republican Party]? Up
to five mentions.
Is there anythingin particularthat you dislike about the
Republican Party? What is that?
Anythingelse [you dislike about the RepublicanParty]?
Up to five mentions.
REFERENCES
Aldrich, John H. 1995. WhyParties? The Originand Transformationof
Party Politics in America. Chicago: University of Chicago Press.
Aldrich,JohnH., MarkM. Berger,andDavidW. Rohde.1999."The
HistoricalVariabilityin ConditionalParty Government,18771986."Paperpresentedat Conferenceon the Historyof Congress,
StanfordUniversity,Palo Alto, CA, January14-15.
and Voting Behavior,1952Bartels,LarryM. 2000. "Partisanship
1996."American Journal of Political Science 44 (January): 35-50.
Beck,PaulAllen. 1997.PartyPoliticsin America,8th ed. New York:
Longman.
Brady, Henry E., Michael Herron, Walter Mebane, Jr., Jasjeet
Skehon,KennethW. Shotts,andJonathanWand.2001."Lawand
Data: The ButterflyBallot Episode."PS: Political Science &
Politics 34 (March): 59-70.
Brody, Richard A. 1991. Assessing the President: The Media, Elite
Opinion, and Public Support. Stanford, CA: Stanford University
Press.
Burden,BarryC., andDavidC. Kimball.1999."A New Approachto
the Study of Ticket Splitting."American Political Science Review 92
(September):533-44.
Burns,Nancy, Donald R. Kinder,Steven J. Rosenstone,Virginia
Sapiro,and the NationalElectionStudies.2001.NationalElection
Studies, 2000: Pre-/Post-Election Study [data set]. Ann Arbor:
Universityof Michigan,Centerfor PoliticalStudies[producerand
distributor].
Campbell,Angus,PhilipE. Converse,WarrenE. Miller,andDonald
E. Stokes.1960.TheAmericanVoter.New York:Wiley.
Carmines,EdwardG., andJamesA. Stimson.1989.IssueEvolution:
Race and the Transformationof American Politics. Princeton, NJ:
Do you thinkthere are anyimportantdifferencesin whatthe
Republicansand Democratsstandfor?
PrincetonUniversityPress.
Craig,StephenC. 1985."TheDecline of Partisanshipin the United
States:A Reexaminationof the NeutralityHypothesis."Political
Party Feeling Thermometers
Delli Carpini,MichaelX., and Scott Keeter.1996.WhatAmericans
I'd like to get your feelings toward some of our political
leaders and other people who are in the news these days
(1990:who have been in the news). I'll read the name of a
personand I'd like you to rate that personusingthe feeling
thermometer.Ratingsbetween50 and 100degreesmeanthat
you feel favorablyand warm toward the person; ratings
between0 and 50 degreesmean thatyou don'tfeel favorably
towardthe personand thatyou don'tcare too muchfor that
person.You would rate the personat the 50 degree markif
warmor cold towardthe person.If
you don'tfeel particularly
we come to a personwhose name you don't recognize,you
don'tneed to rate thatperson.Justtell me andwe'llmoveon
to the next one.
The Democratic Party.
The RepublicanParty.
630
Behavior 7 (March): 57-78.
Know about Politics and Why It Matters. New Haven, CT: Yale
UniversityPress.
DeSart,Jay A. "InformationProcessingand PartisanNeutrality:A
Reexaminationof the PartyDeclineThesis."Journalof Politics57
(August):776-95.
Dionne, E. J., Jr. 1991. Why Americans Hate Politics. New York:
Touchstone.
Downs, Anthony. 1957. An Economic Theory of Democracy. New
York:Harper.
Erikson,Robert S., MichaelB. MacKuen,and JamesA. Stimson.
A Response to Green,
1998. "WhatMoves Macropartisanship:
Palmquist, and Schickler." American Political Science Review 92
(December):901-12.
Political
Fiorina,MorrisP. 1992."AnEra of DividedGovernment."
Science Quarterly107 (Autumn): 387-410.
Fiorina, Morris P. 1996. Divided Government, 2d ed. Needham
Heights,MA: Allyn & Bacon.
ceceRve
Review
Political Science
AmericanAmrcnPliia
Graber,Doris A. MassMediaandAmericanPolitics,5th ed. Washington,DC: CQ Press.
Green, Donald P., Bradley Palmquist,and Eric Schickler.1998.
AmericanPolitA ReplicationandCritique."
"Macropartisanship:
ical ScienceReview92 (December):883-900.
Green,DonaldP., BradleyPalmquist,andEricSchickler.2000."The
PS:PoliticalScience& Politics
ComingDemocraticRealignment."
33 (June):199-200.
Hetherington,MarcJ. 1999. "The Effect of PoliticalTrust on the
PresidentialVote: 1968-96."AmericanPoliticalScienceReview93
(June):311-26.
Hibbing,John R., and ElizabethTheiss-Morse.1995. Congressas
PublicEnemy:PublicAttitudesTowardAmericanPoliticalInstitutions.New York:CambridgeUniversityPress.
Hinich,MelvinJ., and MichaelC. Munger.1994.Ideologyand the
Theoryof PoliticalChoice. Ann Arbor: Universityof Michigan
Press.
Keefe,WilliamJ. 1998.Parties,Politics,andPublicPolicyinAmerica,
8th ed. Washington,DC: CQ Press.
Keith, Bruce, E., David B. Magleby,CandiceJ. Nelson, Elizabeth
Orr,MarkC. Westlye,andRaymondE. Wolfinger.1992.TheMyth
of theIndependentVoter.Berkeley:Universityof CaliforniaPress.
Key, V. 0. 1966. The ResponsibleElectorate.Cambridge:Harvard
UniversityPress.
King, David C. 1997. "The Polarizationof AmericanParties and
Mistrustof Government."In WhyPeopleDon't TrustGovernment,
ed. Joseph S. Nye, Jr., Philip D. Zelikow, and David C. King.
Cambridge,MA: HarvardUniversityPress.Pp. 155-78.
Konda,ThomasM., and Lee Sigelman.1987."PublicEvaluationsof
the AmericanParties,1952-1984."Journalof Politics49 (August):
814-29.
AmerLuskin,RobertC. 1987."MeasuringPoliticalSophistication."
icanJournalof PoliticalScience31 (November):856-99.
Luskin,Robert C. N.d. "FromDenial to Extenuation(and Finally
Beyond): Political Sophisticationand Citizen Performance."In
ThinkingaboutPoliticalPsychology,ed. JamesH. Kuklinski.New
York:CambridgeUniversityPress.Forthcoming.
MacKuen,MichaelB., Robert S. Erikson,and JamesA. Stimson.
1989. "Macropartisanship."
AmericanPoliticalScienceReview83
(December):1125-42.
Markus,GregoryB. 1988. "The Impactof Personaland National
EconomicConditionson the PresidentialVote: A Pooled CrossSectional Analysis."AmericanJournal of Political Science 32
(January):137-54.
Mermin,Jonathan.1999.DebatingWarandPeace:MediaCoverage
of
in thePost-Vietnam
Era.Princeton,NJ:Princeton
U.S.Intervention
UniversityPress.
Miller,WarrenE. 1998."PartyIdentificationand the Electorateof
the 1990s."In ThePartiesRespond:Changesin AmericanParties
and Campaigns,3d ed., ed. L. Sandy Maisel. Boulder, CO:
Westview.Pp. 109-27.
Mondak,JeffreyJ. 1999."Reconsideringthe Measurementof Political Knowledge."PoliticalAnalysis8 (1): 57-82.
Nie, NormanH., Sidney Verba, and John R. Petrocik.1979. The
Vol.
ol 95,
5 No.
o 3
ChangingAmericanVoter.Cambridge,MA: HarvardUniversity
Press.
Elections:
Page,BenjaminI. 1978.ChoicesandEchoesin Presidential
RationalMan and ElectoralDemocracy.Chicago:Universityof
ChicagoPress.
Page,BenjaminI., andRobertY. Shapiro.1992.TheRationalPublic:
Fifty Yearsof Trendsin Americans'PolicyPreferences.Chicago:
Universityof ChicagoPress.
Pomper, Gerald M. 2001. "The PresidentialElection." In The
Electionof 2000, ed. Gerald M. Pomper.New York: Chatham
House. Pp. 125-54.
Poole, KeithT. 1998."Recoveringa BasicSpacefroma Set of Issue
Scales."AmericanJournalof PoliticalScience42 (July):954-93.
A PoliticalPoole, KeithT., and HowardRosenthal.1997.Congress:
EconomicHistoryof RollCallVoting.NewYork:OxfordUniversity
Press.
and
Popkin,SamuelL. 1994.TheReasoningVoter:Communication
Persuasionin Presidential
2d ed. Chicago:Universityof
Campaigns,
ChicagoPress.
Price, Vincent, and John Zaller. 1993. "Who Gets the News?
AlternativeMeasuresof News Receptionand TheirImplications
for Research."PublicOpinionQuarterly
57 (Summer):133-64.
House.
Rhode,DavidW. 1991.PartiesandLeadersin thePostreform
Chicago:Universityof ChicagoPress.
Rosenstone,StevenJ., Roy L. Behr, and EdwardH. Lazarus.1996.
ThirdPartiesinAmerica:CitizenResponseto MajorPartyFailure,2d
ed. Princeton,NJ: PrincetonUniversityPress.
Saad,Lydia.2000."Gore'sImageSlippingas ElectionContestDrags
On."GallupPoll NewsRelease.December6.
Sapiro,Virginia,StevenJ. Rosenstone,WarrenE. Miller,and the
NationalElectionStudies.1997.1948-1996 Cumulative
Data File
[dataset].Ann Arbor:Universityof Michigan,Centerfor Political
Studies[producerand distributor].
Schattschneider,E. E. 1975. The Semisovereign
People:A Realist's
Viewof Democracyin America,2d ed. New York:Holt, Rinehart,
and Winston.
Sniderman,PaulM., RichardA. Brody,and PhilipE. Tetlock.1991.
Reasoningand Choice:Explorationsin PoliticalPsychology.New
York:CambridgeUniversityPress.
Stanga,John E., and JamesF. Sheffield.1987."The Mythof Zero
Partisanship:AttitudesTowardthe AmericanPoliticalParties."
AmericanJournalof PoliticalScience31 (November):821-55.
Wattenberg,Martin P. 1984. The Decline of AmericanPolitical
Parties,1952-1980.Cambridge,MA: HarvardUniversityPress.
Wattenberg,Martin P. 1990. The Decline of AmericanPolitical
Parties,1952-1988.Cambridge,MA: HarvardUniversityPress.
Wattenberg,Martin P. 1994. The Decline of AmericanPolitical
Parties,1952-1992.Cambridge,MA: HarvardUniversityPress.
Wattenberg,Martin P. 1996. The Decline of AmericanPolitical
Parties,1952-1994.Cambridge,MA: HarvardUniversityPress.
Wattenberg,Martin P. 1998. The Decline of AmericanPolitical
Parties,1952-1996.Cambridge,MA: HarvardUniversityPress.
Zaller, John. 1992. TheNatureand Originsof Mass Opinion.New
York:CambridgeUniversityPress.
631