3595626_A New Measure of Party Strength

8/3/2019 3595626_A New Measure of Party Strength

http://slidepdf.com/reader/full/3595626a-new-measure-of-party-strength 1/13

University of Utah

A New Measure of Party StrengthAuthor(s): James W. Ceaser and Robert P. SaldinSource: Political Research Quarterly, Vol. 58, No. 2 (Jun., 2005), pp. 245-256Published by: Sage Publications, Inc. on behalf of the University of UtahStable URL: http://www.jstor.org/stable/3595626 .

Accessed: 13/01/2011 22:08

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=sage. .

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 service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Sage Publications, Inc. and University of Utah are collaborating with JSTOR to digitize, preserve and extend

access to Political Research Quarterly.

http://www.jstor.org/action/showPublisher?publisherCode=sage

http://www.jstor.org/action/showPublisher?publisherCode=utah

http://www.jstor.org/stable/3595626?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp



http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/stable/3595626?origin=JSTOR-pdf

http://www.jstor.org/action/showPublisher?publisherCode=utah




A New Measure of Party Strength

JAMESW CEASER,ROBERTP.SALDIN, UNIVERSITYFVIRGINIA

Inorder o measurehestrength f thepartiesn eachstate,theMajorParty ndex(MPI)was builtby averag-ing theresultsof the six major lections hat takeplace n thefiftystates.This ndex allowsus to describe he

absoluteandcomparativeartisaneaningof each state n each electionandidentify rendsof party trengthover time within ndividual tates,amongregions,and within the nationasa whole.TheMPI hedsconsider-ablelighton threegeneraldevelopments:1) a nationalchange romDemocratic ominancen the 1980s toa Republicandge by 2002, (2) significant egionalrealignmentsn the South and New England,and (3) a

strong rend owardgreater onsistencybetweenpartisan otingat the federalandstate evels.

Party strength n the UnitedStates s generallymeasured

by public opinion polls in which respondentsreporttheirown partisandentification.Thereare well-known

advantageso thisapproach. t permitsanalyses hatrelatean

individual's arty preference o attitudinaland demographic

factors,therebyallowingfor the generationand testing ofhypothesesaboutthecausesof partyattachment. elf-identifi-

cationcanalso be correlatedwithreported otes,which allows

forinquiries nto the relative mportanceof party preferencefor electoralbehavior.For these reasons,poll surveyshave

become the preferredtandardn academicresearch.But this

method has admittedweaknesses. Polls are expensiveand

theyoftenhave a significantmarginof error,whichin the case

of generalnationalpolls increasesgreatlyas one begins to

investigate articular ub-groupsorgeographicalegions.Par-

tisan self-identification y itself, moreover,does not tell ushow people actuallyvote. For example, many white South-ernersafter1950 identified hemselvesas Democrats r inde-

pendentseven as theyhadbecome reliableRepublican otersat the presidential evel. Finally,partisanself-identificationitselfsays nothingabout the habitualbehavior-if it exists-of independents.Manyanalysts odaysuspect hatmostvoters

who call themselvesndependentsn facthave a fairlydistinct

partisan eaningwhen it comes to casting heirballots.'Anotherapproachto assessing partystrengthis to look

at election resultsand measurepartystanding by the voteseach party'scandidates received. This method also has its

strongpoints. It is inexpensive(at least forscholars),as the

governmentpicks up all of the costs of the researchprojectby holding elections. The marginof error s fairly ow,withthe inaccuracies

beinglimited to fraud and the now well-

documented difficulties encountered in counting -and

recording votes (hanging chads, undervotes, votingmachine failure,and the like).2 Election results also have

Bruce E. Keith,et al., TheMythof theIndependent oter Berkeley:Uni-

versityof CaliforniaPress,1992).2 For a discussion of some of these problems,see Brady,HenryE.,Justin

Buchler,MattJarvis,andJohn McNulty.2001. "CountingAll the Votes:The Performanceof Voting Technology n the United States."Availableat:http://ucdata.berkeleyedu/.

PoliticalResearchQuarterly, ol.58, No. 2 (June2005): pp. 245-256

the advantageof directly measuring the phenomenon: if

one is interestedin party strength,it makes greatsense to

look at how citizens actuallyvote. Still, there are weak-

nesses in using electoral results. The statistics are tied to

aggregates (collective units), not to individuals, which

makes itimpossible

to connect the vote to attitudinalvari-

ables and difficultto relateit to many demographicfactors

(other,of course, than geography tself).3Anotherproblemis that a vote for a particularcandidate s by no means the

same thing as an expressionof support for a party; t may

only reflecta preference ora particularndividual,as when

millions of Democrats pulled the lever for the war hero

Dwight Eisenhower in 1952 and 1956. Election results

from any single election may thereforereveal little about

"real" r "normal" artystrength.All approachesare bound to have strengthsand weak-

nesses,and thequestionof a measurement's orthshouldbe

judged on the practical grounds of its helpfulness as an

investigativeool. Despitethe drawbacksust noted, the ideaof measuringparty strengthon the basis of election results

seems attractiveenough to hold considerableinterest for

politicalscientists-not as a substitutefor,but as a supple-ment to, polling methods. Yet this approach is virtuallyabsent from contemporarypoliticalscience. Frequently,of

course, one comes across rule-of-thumbdescriptionsthat

refer to election outcomes. Pick up any book on elections

and you will likely readstatementssuch as "Massachusetts

voted Democraticn the last fourpresidential lections,"hus

qualifying t as a solidlyDemocratic tate.Generalhistorical

studies also frequentlyrely on such characterizations s in

Blackand Black's1992) characterizationf the South:"For

morethansix decades,from1880-1944, the elevenstatesofthe old Confederacy . . regularlyvoted as a solid block in

favorof the presidential andidateof the DemocraticParty."4Such references,however,arecompletelyad hoc; they have

no common standard and offer no basis for systematic

3 For a recent method that can account for ways of using aggregatedata

for some of these purposes,see GaryKing. 1997. A Solution o theEco-

logical nference roblem: econstructingndividual ehaviorfrom ggregateData.Princeton,NJ:PrincetonUniversityPress.

4 Earl Black and MerleBlack,The VitalSouth(Cambridge:HarvardUni-

versityPress, 1992) 4.

245



246POLITICAL RESEARCH UARTERLY

communicationamongthose studying partystrength.What

is needed, clearly, s a general ndex to which all can refer.The idea behind an index-think for exampleof the Dow

Jones index-is to capturein one number a measurethat

suppliesa broadpictureand that can providea baselinefor

charting rendsand movements n partystrength.A few politicalscientists n the periodbetween 1960 and

1989 sought to devise such an index. These efforts did not

gainwide currencyat the time, and none of them has been

updated or remainsin use today.Partof the reason is no

doubt the aforementioned appeal of survey data. Butanotherreason,we would argue, s to be found in the flawsin these indexes. Withimprovements,we believeit is possi-ble to constructa measurethat overcomessome of the pastdifficultiesand that can serve today as a helpful tool for

political analysis. The objective is to construct a measurethat capturesthe underlyingstrengthof the partieswhile

managingto absorband smooth out distortions of particu-lar elections.

Anymeasurementusing electoralresultsmust, as noted,proceedon the basis of geographicalunits. This unit couldbe as small asthe precinct.But forthe purposesof a generalindex for the whole of Americanpolitics, the state is the

properstartingpoint. States serve as the unit for selectingpresidentialelectors,senators,and governors.Additionally,state majoritiesgenerallydetermine the majority party in

state legislatures. In order to measure the underlyingstrengthof the parties n each state,we have built an indexthat averagesthe results of the elections for the six majoroffices that take place in each of the fiftystates:president,U.S. senator,U.S. representative,governor, state senator,and staterepresentative.

OurMajorPartyIndex (MPI)allowsone to describe theabsolute and comparativepartisan eaning of each state ineach election year and to identify trends of party strengthover time within individualstates,among regions,and (byweightingpopulations)within the nation as a whole. It isalso possible, by breakingthe index into subindexes, tomeasure the consistency of partisanvoting between stateand federaloffices and to observe whether partisanvotingpatterns at these two levels are moving furtherapart orcloser together.The MPIcan be used, as will be demon-

strated, to shed considerablelight on some of the majordevelopments n modern electoralpolitics.

PREVIOUSMEASURESOF INTERPARTY OMPETITION

Pastindexes of party strength employed differentmeth-ods to determine which elections to observe and how to

weight the differentcomponentsin relationto one another.The diversity in approaches resulted in part from theresearchers'different interests and purposes, but it also

highlightssome of the difficulties n using electoral resultsto come to grips with measuringthe abstractionof partystrength.Previous ndexes were generallybased on averag-ing aggregatepartyvote totals for particularofficesover a

totals for federalelectionsonly.5Bycontrast,Ranney 1965)and King (1989) measured the level of partycompetition

only in state-levelcontests.6A few efforts combinedresults

from both levels, as we shall do here (Hofferbert1964;David 1972; Bullock 1988).7

Anotherpoint of differencerevolvedaround the type of

data employed and the frequency of observation. Some

indexes employedthe raw vote totals or percentagesof thevote each party won (Hofferbert 1964; Ranney 1965;

David1972;King1989). Others reliedsolelyon the number

of seats each partyheld (Bullock 1988). The frequencyof

observationalso varied. Some indexes were based on long

periods of aggregateanalysis (Hofferbert,Ranney,King),while others took new measurements every two years(David, Bullock). The Cox Index used a combination of

theseapproaches.Theseindexes were alsoused fordifferent

purposes.Hofferbert ndKingsoughtto determine he levelof competitionin each state and ascertainwhich partywas

in control,while others were more concernedwith observ-

ing partisanshipover time and in analyzingthe reasonsforchange(Cox, Ranney,David,Bullock).

These previous attempts to examine party competitiondemonstratethe diversityof opinion over what should be

measuredand how such measurementsshould be carried

out. Theyalso reveal,however,some methodologicaldraw-

backs,even when takingaccount of the purposesforwhich

they were intended. Threeof these indexes deserve further

attention because their successes and failures nformed the

developmentof our new measure.

The Cox Indexwas an earlyeffort o measurepartycom-

petition in and between the states. Relying solely on vote

totals for federaloffices, Cox labeled each state as either

Democraticor Republicanand classifiedit as either "safe"for a party,"generally"or a party,or "marginal" hen nei-

ther partywas dominant.8 As a result of only looking at

national elections, Cox concluded that most states were

"marginal,"r not alignedwith a party.While this conclu-

sion may have had validityat the federal evel in the post-World WarII years, it certainlydid not describethe situa-

tion within the states. Someof the states,particularlyn the

South, were competitivein presidentialelections,but were

strongly supportiveof one partyat the statelevel.

Cox's index was also methodologicallyflawed in that

each biennial observation ncluded only the elections that

5 EdwardF Cox, "TheMeasurementof Party Strength,"WesternPolitical

Quarterly 3 (1960).6 AustinRanney,"Parties n StatePolitics,"n Herbert acoband Kenneth

N. Vines,Politics n theAmerican tates:A Comparativenalysis FirstEdi-

tion) (Boston:Little, Brown and Company, 1965). James D. King in

"InterpartyCompetition in the American States:An Examination of

Index Components,"Western oliticalQuarterly 2:1 (1989).7 Richard . Hofferbert,"Classification f AmericanStatePartySystems,"

Journalof Politics 26:3 (Aug. 1964), 550-567. Paul T. David, Party

Strengthn the UnitedStates:1872-1970(Charlottesville: niversityofVir-

giniaPress,1972). CharlesS. BullockIII,"RegionalRealignmentroman

OfficeholdingPerspective,"ournalofPolitics50:3 (Aug. 1988) 553-574.8 CoX.articular period of time. Cox's (1960) index used vote

246 POLITICALESEARCH UARTERLY



A NEW MEASUREOF PARTY TRENGTH 247

occurredin that particularyear.Thus, congressionalraces

were included in every measure,while presidentialvotingwas only partof the measure n everyotherbiennialobser-vation. As a result, the measurementvaried enormouslybased on whether a presidentialelectionwas held, makingit difficult to draw conclusions or demonstrate rends overtime. The index failed to supply a comparablemeasure for

each observation.

Ranney'sindex is perhaps the most well known. In

contrast to Cox, Ranney only measured intra-state con-

tests and excluded federal offices. His index also differedfrom Cox's in that it considered the longevity of partycontrol of the governorshipsand the legislaturesin addi-tion to partisanvote totals for governor,state senate, andstate house. Ranneyused his measure to assign each stateto one of five categories:"OneParty Republican,""Modi-fied One-Party Republican," "One Party Democratic,""ModifiedOne-PartyDemocratic,"and "Two-Party."9he

RanneyIndex has a number of limitations. One problem

is that it measures only state-level elections and gives aninsufficient measure of party competition writ large.10Another drawback is that the Ranney Index cannot beused to examine individual years because of its consider-ation of a party's ongevity of control.

Finally, Paul T. David's (1972, updated 1974, 1976,1978) index averagedthe vote for governor,U.S. Senate,and U.S. House elections and classified states as "pre-dominantlyDemocratic,""predominantlyRepublican,"or

"competitive."1'His index is more comprehensive thanCox's or Ranney'sbecause it considered both state andfederal elections. In addition, David's index took thesethree factorsinto consideration at each biennial observa-

tion, thus avoiding the inconsistency of Cox's measure-ment. (The results of the last governor's race wereincluded in the index, even if a gubernatorialrace did nottake place in that year.) Nonetheless, David's index alsohas problematic features.Although it includes both fed-eral and state electoral competition, it omits the most

important of all elections in America: the contest for the

presidency.In addition, given the three races David uses,his index assigns two-thirds of the weight to national

voting and one-third to state voting. No compellingreason is given for this weighting. David'sindex, finally,seems skewed because, within the intra-statecomponent,he includes elections for the governorshipbut omits those

for the state legislatures, which seem just as important.For example, in 1972 Southern Republicans intermit-

tently controlled governorships, but won only a handfulof seats for state legislaturesand did not control a singlestate legislativebody.At the time these states were clearlymore Democratic than states that had intermittentRepub-lican governors with occasional or frequent Republican

9Ranney.

10King,84.

1 David.

state legislatures.The David Index missed this importantelement in state partisanship.l2

THEMAJOR ARTYNDEX

The MajorPartyIndex (MPI) s intendedto measurethelevel and extent of interpartycompetitionin and between

the states. It builds on previous findings by providing ameans to identifystatesdominatedby one party(asCoxand

RanneyIndexes did) as well as to observe importantchar-

acteristicsof the national electorate in specific years andover time. It also improveson most past indexesby includ-

ing both federaland stateelections. Eachcomponentwithin

these two arenas is weighed evenly-indeed, there is no

compelling reason to assign them different values. Of

course, the constructionof any index has an arbitrary le-ment to it but the MPIhas the most logicalmakeup possi-ble. Althoughthe index cannot be said to offera perfectrep-resentation of Americanparty strength-no measurement

can-the hope is that it will offer a more comprehensiveand more reliablepicturethanany index thus farproposed.

Design

TheMajorParty ndexis comprisedof six weightedcom-

ponents calculated on even numberedyearsfor each state,from1990 to 2002. Threeof thesecomponentsarebased on

elections for national officesand three for state-leveloffices.Withinthe federalportion,half of the weight is assignedto

presidentialelections and half to Congressionalelections.

Similarly,he statecomponentis evenlydivided betweenthe

governorshipand the state legislature.This is an appropri-

ate weightingscheme because it attachesequal importanceto the two arenasof government n our federalsystemand,within each arena,to the executive branch and the legisla-tive branch. The MPItakes all of the majorelective officesinto account, therebyfulfillingthe aim of being a compre-hensive measure.

The index is based on a two-partyevaluation. Third-

party candidates have been eliminated in all six compo-nents. A value for all componentsis entered for each bien-nial observation,which is accomplishedby using the most

recentpresidentialandgubernatorial lectionresults f none

was held in a given year.Thecomponentsand theirweightsare listed in Table1.

The firstcomponent of the MPIand the firstnationallybased measure is the two-partyvote for presidentin eachstate'smost recent presidentialelection. Thus in 2000, the

12 In addition, calculatingthe index for three parties rather than two

lessens comparativeopportunities.Byincludingthirdparties n his for-

mulation, David decreases the marginal difference between the two

major parties,making them more difficult to compare. By using two-

partyvote totals the majorpartiesbecome mirror magesof each other,

increasingand decreasing n proportion o one another.The changes n

partisansupportare thus amplifiedand a more realisticand meaning-ful analysisof the major parties s possible.

A NEWMEASUREFPARTYTRENGTH 247



248 POLITICAL ESEARCH UARTERLY

= TABLE1

COMPONENTS F THEMAJORPARTYNDEX MPI)

President ........ 25%

Congress ........ 25% (Senate: 12.5%; House: 12.5%)

Governor ........ 25%

State Legislature . . . 25% (Senate: 12.5%; House: 12.5%)

value is based on the election result of thatyear.For 2002the same value is entered becausetherewas no presidentialelection in 2002. This factor accounts for 25 percentof thetotal index value foreach state.

Thesecond MPIcomponent,also at the national evel, isthe two-partyvote in each state'stwo most recent UnitedStatesSenate electionsand accounts for 12.5 percentof thetotal index value.'3Totake an example,Idaho'sU.S. Senatevalue is calculatedby averaging he results from the 1998and 2002 Senate elections. By taking both of the Senate

elections into account, despite the time lag on the former,the MPIattemptsto reflectpartisanvotingfromyearto yearwithout overemphasizing current partisan swings. Byincluding a Senate result every year, it evens out results

amongdifferent tates.The thirdcomponent,and the finalnationalmeasure, s

the total two-partyaverageof all votes in each state'sbien-nial U.S. House elections. Virginia's alue, for example, isobtainedby adding Republicanand Democraticvotes in all

congressionaldistricts and calculatingeach party'spercent-age of this total. This method provides a more accuratemeasure of the overallstatepartisanchoice than would beobtained by averagingthe two-partypercentagesof each

districtbecause districtsdiffer n populationandturnout.Inaddition, it reduces the impact of uncontested seats. This

component of the MPIaccounts for 12.5 percent of eachstate's otal score.14

The fourth component of the MPI,and the first state-level measure, is the two-party vote percentagein eachstate'smost recentgubernatorial lection. This componentaccounts for 25 percentof each state's otal.l5

13 Values for Louisianaare not always present because of the state'sunusualvoting laws. For much of the 1990s, the state held a "blanket

primary"he summerbeforethe usual November elections. If a candi-date received a

majorityof the

vote,he or she

was elected and the seatwas not on the Novemberballot. In these instances,Louisiana's enatevaluewas omitted and the 12.5 percentwas dividedequallyamongtheother components. Jim Jeffords, Vermont'sindependent senator, iscounted as a Republicanbecause he was elected under that party'sbanner. Should he seek reelection as an independent, he will becounted as a Democratbecausehe caucuses with thatparty.

14 Vermont'sndependentrepresentative,BernieSanders, s counted as aDemocratwhile Virginia'sVirgilGoodeis counted as a Republican venin thoseyearsbeforehe officially oined thepartyThisallows fora moreaccuratevote countbecausetheparties hatbenefitfromthesemembersnevermounted a legitimatechallenge.

15 Thereis one exceptionto using the most recentelection. Twostates-Vermontand New Hampshire-elect governorsevery two yearsratherthan every four years. For consistency and comparability, four-year

The fifthmeasure,also at the statelevel, is the two-party

percentageof all seats in eachstate'sSenate.TheMajorPartyIndex value is determined in this case by dividing thenumberof Republican eatsby the sum of Democraticand

Republicanseats. Therewere two reasons in this case for

using seats(as opposed to votes)as the basis of calculation:vote totals for statelegislativeelections aredifficult to find,and manyof these racesareuncontested(a largenumberofuncontested seats skews the two-partyvote totals).l6This

componentis weightedas 12.5 percentof the total index.The finalcomponentof the MPI,and the thirdstate-level

measure, s the two-partypercentageof seats in statehouses.This scoreis calculated n the sameway as that for the statesenateand is worth 12.5 percentof the total indexvalue.17

Thesum of these components s then calculated o deter-mine the biennialMajorParty ndex value in each state.TheMPIis based on Republicanscores in a two-partyevalua-tion. The Democraticvalue is the inverseof the Republicanscore. Thus, as the numerical value increases, a state

becomes more Republicanand, conversely,as the valuedecreases,a state becomes more Democratic.The formulafor each state in each even numberedyear,then, is:

MAJOR ARTYNDEX= ((Mostrecent2-PartyRepublicanPresidentialVote)*0.25)+ ((Average f the TwoMostRecentRepublican2-PartyVotes for the U.S.

Senate)*0.125)+ ((Republican2-PartyPercentof allU.S. House Votes)*0.125)+ ((MostRecent2-PartyRepublicanVote for Governor)*0.25)+ ((2-Party

RepublicanPercentageof Seats in the State

Senate)*0.125)+ ((2-PartyRepublicanPercentageofSeatsin the StateHouse)*0.125).

Table2 lists the four most Republicanand Democraticstates over the pastthreebiennialobservationsnote thatthevalues istedin the tableare the difference f the MPI coreof

eachstate from50, expressedas positivenumbersforeaseof

comparison).On the Republicanside, Idaho has had the

highest MPIratingsince 1998, with a 68.5 score in 2002.

Indeed,Idaho is a haven for the GrandOld Party, nd if all

value is obtainedby averaging he two-partyvote total from the two

most recent gubernatorialelections. Vermont's2002 gubernatorialvalue, for instance,is obtainedby averaging he 2002 and 2000 two-

partyvote totals.16 The data of popularvotes forstatelegislativeelections is unavailable n

any one central ocation for electionsbefore 1996. It would have been

possible, with some difficulty, o have collected this data,but we were

mindful also of the possibility that at some point certain researchers

maywish to extend the index backin time before 1990. Requiring he

popularvote totalswould have made such an extension difficult f not

impossible.One of the advantagesof this index, as we see it, is its rela-tive ease in preparation.The asymmetriesntroducedby using different

bases(popularvote sharesand seatshares) s not, in ourview,veryseri-ous given the broad descriptive purposes for which the index is

intended.17 In the case of Nebraska,which has a unicameral,nonpartisan egisla-

ture, the state House and Senatevalues are omitted and the othervari-ables areproportionally e-weighted.





TABLE

MOSTREPUBLICANNDDEMOCRATICTATES S MEASURED YMPI (1998-2002)*

1998 2000 2002

MostRepublican Idaho,69.1 (19.1) Idaho,69.4 (19.4) Idaho,68.5 (18.5)

Utah, 68.7 (18.7) Kansas,68.2 (18.2) Alaska,67.7 (17.7)

Kansas,66.3 (16.3) Wyoming,67.2 (17.2) Utah 67.6 (17.6)

Montana,61.8 (11.8) Utah, 64.9 (14.9) Wyoming,64.8 (14.8)

MostDemocratic Hawaii,31.9 (18.1) Mass,31.4 (18.6) Mass,31.5 (18.5)RhodeIs., 32.8 (17.2) Hawaii,34.9 (15.1) RhodeIsland,34.8 (15.2)WestVirg,33.2 (16.8) RhodeIsland,34.9 (15.1) Hawaii,36.2 (13.8)Mass.,34.0 (16.0) Maryland,35.4 (14.6) Maryland,39.3 (10.7)

*Thevaluesin parentheses ndicatethe differenceof each state'sMPIscore from50, expressedas apositivenumbers for ease of comparison.Thus, in 1998Idahohad an MPIof 69.1 (19.1 points above50) and Hawaiihad an MPIof 31.9 (18.1 points below 50). This means that Idaho was slightly more Repub-lican thanHawaii was Democratic.

the nation were like Idaho,GeorgeW Bushwould have noneed for KarlRove. No Democraticpresidential ominee hascarried he statesinceLyndon ohnsonedgedBarryGoldwa-terby 5,000 votes in 1964. Similarly, o Democrathas beenelected to the U.S. Senate since FrankChurch n 1974. On

averageRepublican residential ndsenatecandidatesgamerover 60 percentof the vote. Occasionally dahoDemocratshavebeen electedto the U.S.House,thoughnone have in thelast twelveyears.The current tatelegislaturehas a Republi-canmajority f 28 seats to 7 in the Senateand54 to 16 in theHouse.Moststatewideofficeshavealso been heldby Repub-licans, although the Democrats controlled the governor'smansion from1970 to 1994-due in largepartto the popu-

larityof CecilAndrus,a conservativeDemocrat,who servedas the chief executive for sixteen of those years. In the 42statewideelections since 1984, Democrats have only wonthirteen-and six of thesewere for the minoroffice of Con-troller. nonlyfourof theseelectionswas thevotemargin essthan fivepercentbetween the two majorpartycandidates.18In short,no one canbe surprised hatIdaho has become the

postercase forRepublican trength.Massachusetts as had the lowestMPIratingsince 2000,

making t themost Democratic tate,forthe two most recentbiennial observations.Massachusetts,as noted, has goneDemocratic n all presidentialraces since 1988. It gave thewidest marginof any state to Bill Clinton in 1996 and the

second largest to Al Gore in 2000. Its two Senators areDemocrats TedKennedyandJohnKerry), nd all ten Housemembers are Democrats,with none receivingless than 60

percent of the vote in the 2002 election (six were unop-posed). On BeaconHill, in the statelegislature,Republicanscompriseless than 15 percent of the membership n bothhouses. Notwithstandinghis fact,Massachusetts ow has a

Republicangovernor asdo a few otherstronglyDemocratic

1 Todd Lochnerand GaryMoncrief,"Idaho:ElectoralReformat the Mar-

gins,"Election Reform:Politicsand PolicyConference,29 May2003.

states);and, in a greatanomaly,t has had a Republicangov-ernor since 1990, giving it one of the nation's ongest con-secutiveperiodsunderRepublicangovernors.

THE MAJORPARTY NDEXAPPLIED

The MajorPartyIndexnot only allows us to describethe

partisan eaningof eachstateeverybiennialyear,but also to

identify trends in partisan strengthover time within indi-vidual states, among regions, and within the nation as awhole. The MPI heds lighton threedevelopments n recentAmericanhistory: (a) a national change from Democraticdominancein the 1980s to a Republicanedge by 2002, (b)

significant regional realignments n the South (in favorofthe Republicans)and in New England (in favor of the

Democrats), and (c) a trend toward greater consistencybetweenpartisanvotes at federaland state levels.

The MPIdisplaysthe rise of the RepublicanPartyat the

national evel.Appendix1 givesthe MPIvalue foreach stateat each biennial observation rom 1990 to 2002. Appendix2 displaysthe figuresfor2002 brokendown by each of thesix components.19Republicanstates,defined as those witha final valueabove50, areshaded.A decadeago, Democrats

were clearlythe dominant partywith an advantage n 33statescompared o only 17 for the Republicans.Butoverthelast ten years Republicans gained states in every election

except 1998. The 2002 MPIindicates that the Republicanresurgencehas not yet ended. When the 2002 MPI s con-trastedwith the results fromthe early1990s, it is clearthata majorpartisanreversalhas takenplace.Now 34 statesfallon the Republican ide of the ledger,while only 16 remainDemocratic see Table3).

The most significant change in the state breakdownbypartisan eaningoccurred n 1994. Thatyear'selections, of

19A data file for all the componentsfor all statesin each yearis availableon request.





TABLE

STATEADVANTAGEYPARTY1990-2002)

1990 1992 1994 1996 1998 2000 2002

RepublicanStates 20 17 27 28 27 32 34DemocraticStates 30 33 23 22 23 18 16

course,were dominatedby the Republicans,who won con-trol of both chambersof Congressfor the first time since1954. Republicansalsomade notablegainsat the statelevel,both in gubernatorialand state legislativeelections. The

change in the partisanbreakdown,with Republicansgain-ing the MPIadvantage n ten states, reinforces our confi-dence in the measure. While the index smoothes out thevalue by spreadingit over a number of election contests

(includingsome in previousyears), t is sufficiently ensitiveto quicklyregister argeand broadchanges.

Computingpartisanadvantageby the number of states,

however,tells only part of the story.Any kind of nationalmeasure must also take into account the states'respectivepopulations. While the Republicans currently enjoy an

advantage n 34 states, this does not necessarilyindicatethat they arenow the dominantpartynationally.Severalofthe largeststatesremain n the Democrats'column, includ-

ing three of the five most populous states: California 1),New York(3), and Illinois (5). On the other side of the

ledger, seven of the Republicanstates are among the tensmallest states in the country, including Wyoming (50),Alaska(48), North Dakota(47), South Dakota(46), Mon-tana(44), and New Hampshire 41).

Weighting he data to takeelectoral trength nto account

allows for a proportional xamination.Goingbackto 2000,the yearof the last presidentialelection, the dataindicatedalmostperfectelectoralparitybetweenthe parties.Republi-can statesaccountedfor277 of the 535 stateelectoralvotes

(51.8 percent),while the Democratic states had a total of258 (48.2 percent).(TheDistrictof Columbia,with its threeelectoral votes for the presidency, is not included; theDemocrats have a complete lock on the District'svotes,

TABLE

DEGREEFPARTYTRENGTHYSTATE002

Even(50) AL

Democrat Republican

0-2 DE, KY,ME,MI, GA,IA, LA, MN, MS,OR PA,NC, TN, WI

2-4 NY IN, NE, MO

4-6 WA,IL,VT,CT, FL,OK,SC

NJ,AR

6-10 CA,NM,WV SD, TX, NV, OH, AZ,VA,ND, CO MT

+10 MD, HI, RI,MA D, AK, UT, WY,KS,NH

having won them in every election since they have been

counted.)Furtherevidenceof the paritybetweenthe partieswas of course evidentin GeorgeW Bush'snarrowvictory, nwhich he won 271 electoralvotes to Al Gore's266 and in

which the candidateswere separatedby only 540,000 pop-ular votes in Gore'sfavor (the MPI is not yet sufficientlyrefined to takeSupremeCourt votes into account).

But using the 2002 Major Party Index and the new

(2004) electoralcollegefigures,the shift toward the Repub-licans now becomes more evident. Republicanshave an

advantage n states accounting for 307 electoralvotes, or

57.4 percent,while Democrats control states that cumula-tively carry228 of the 535 electoralvotes or 42.6 percent.20

This analysis,however,is based on dichotomizingstates

into the simplecategoriesof Republicanand Democratic.Itdoes not take into account the closeness of the divisionbetween the parties. Further inspection indicates that

althoughRepublicansare dominantin a majorityof states,their hold on these statesis not necessarilystable. Table 4

displays each state's2002 MPI score from highest (most

Republican) o lowest (most Democratic)and groupsstates

togetherbased on their distance from equal party control

(50). Nine Republicanstates have a score of less than 52

(two points or less from even partycontrol),while the same

is true for only five Democraticstates. Within a four-pointrangearetwelveRepublicans tates but only six Democraticstates.Thus,despitethe 34 to 16 stateadvantage he Repub-licans currentlyenjoy, their control of many states is farmore tenuous thanis the case with the Democrats.

There is a more comprehensiveway to indicatenational

party strengthby one simplemeasure.It is to take the MPIof each state, weight it accordingto its electoralstrength,and then calculate a nationalaverage.This measureis theDowJonesof politicalfortunes,representingn our view the

single best generalmeasure of party strength.Even if the

meaningof theabsolute evelsmaybe questioned,the meas-urements of relativechange are instructive. The index is

shown in Chart1.21

20 Thiscomparisons still slightlybiasedtoward he Republicans.Electoralvotes are determinedby the number of Congressional istricts n a stateand the number of United States Senators.Congressionaldistricts are

apportionedbased on populationbuteverystatehastwo Senators.Thus,smaller stateswill have more electoralvotes under this system than if

apportionmentwas determinedstrictly by population.Weightingthe

populationbased on electoralvotes, though,is arguablymoreinsightfulbecausethe presidency s the only officevotedupon by the entirecoun-

try,and the winner is determinedbasedon the electoralcollege.21 Despite showing two lines, there is really only one number to this

index. The Democratic coreis calculatedas 100 minus the Republican




A NEWMEASUREOF PARTYTRENGTH 251

CHART

NATIONALWEIGHTEDAVERAGE

60.00

58.00

56.00

54.00

52.00

o

c-(:_

50.00

48.00

'-- -' - 4 - Republican:

I -*""Democratic:? . ? ? ? a . a *

/ ,-

qlk/

1990 1992 1994 1996 1998 2000 2002

Years

The Major PartyIndex also highlightsone of the most

striking developmentsof Americanpolitics in recent times:

strong (and partly offsetting) regional realignments.TheSouth's realignment is the greatest.22Table 5 displaysRepublicangains in the region. In 1990, the firstyearcov-eredby the MPI,everySouthernstateexceptSouth Carolina

was in the Democraticcolumn. Since then there has been ahuge shift toward the RepublicanParty.Every Southernstate increased ts MPIvalue, with the result that all of thestates with the exceptionof Arkansasnow show a Republi-can advantage Alabamas almost dead even). A few states

experiencedtrulydramatic hiftsduringthese twelveyears.For instance,Georgiaand Texas ncreased theirMPIvalues

by 12.2 and 12.0, respectivelyAccounts of Southernpoliticshave tracedthe outlines of

this transformation.The Southern realignment towards

Republicansbeganin the 1950s, at the presidentialevel, inthe two Eisenhowerelections. The strengthof the partyin

presidentialelectionsbeganto spreadthereafter, nd by the

time of the three "Reaganelections" (1980, 1984, and1988), Republicanswon the electoralvote of every South-ern state except Georgiain 1980 (Jimmy Carter'shome

state).No non-SouthernDemocrathas captureda Southern

score. Datapoints are as follows: 1990: 46.55R 53.45D; 1992 45.33R,54.67D; 1994 49.50R, 50.50D; 1996 49.58R, 50.42D; 1998 49.42R,

50.58D; 2000 50.41R, 49.59D; 2002 52.32R, 47.68 D.22 For the purposesof this paper,the South is definedin accordancewith

the longstandingpolitical science consensus grounded in VO. Key's(1949) work:Alabama,Arkansas,Florida,Georgia,Louisiana,Missis-

sippi, NorthCarolina,SouthCarolina,Tennessee,Texas,andVirginia.

Statesince 1968, andAl Gore,a Southerncandidate, ost allof the Southern states in 2000. The realigning processworked its way graduallyto representationn the Senate,the House and the governors'mansions. A thresholdwascrossed in the 1994 election. In 1992 a majorityof both thesenators and House members from the South were Demo-

crats,ashad been the case for more than a century;but afterthe 1994 elections, the majority n both cases were Repub-licans. A similar, f slower, trend has followed at the statelevel. In 1990 the Republicansdid not control a single leg-islatureof a single statein the South,whereasby 2002 theyhad a majorityn 10 of the region's22 legislativebodies.The

good news for Republicans n the South is that they have

gained tremendouslyover the past half-centuryand eventhe past decade. The bad news may be that, for national-level offices,they arenearingtheirpeak (thoughsubstantial

gains remain in the offing for some of the state legisla-tures).23Theonly statein the South to avoida full tip to theGOP has been Arkansas,althoughit too has been moving

slowly in that direction. Part but only part-of its resist-ancewas due to the fact that it was the home stateto Presi-dent Clintonin 1992 and 1996.

New England also experienced a realignmentbut-incontrast to the South-in favor of the Democrats. Table 6

displaysthe New Englanddata.In 1990 two of the six New

England States were Republican:Connecticut and New

Hampshire.By 2002, only New Hampshire-the Arkansas

23Republicans till have gains to be realizedin a few states, and a pro-posed redistrictingplan in Texas could net a substantialnumber ofadditionalHouse seats.

46.00

44.00

42.00

40.00

.$

IA

,,~ - AL - -...o. 1~ ~ ~ ~





= TABLE

CHANGEOF PARTY TRENGTHN THESOUTH1990-2002

South 199(

Alabama 42.7Arkansas 35._

Florida 48.L

Georgia 39.Louisiana 29.

Mississippi 42.

North Carolina 48.

South Carolina 52.-

Tennessee 44.]

Texas 47.(

Virginia 47.2

*MoreRepublicann 2002 than 1990.

7

4

37

7

L

L

2

1996

47.039.6

51.4

48.145.5

50.8

48.7

53.8

50.2

53.252.5

2002

50.0*44.1*

55.5*

51.5*51.8*51.7*

51.1*

56.0*

50.7*

59.6*58.2*

TABLE

CHANGEOF PARTY TRENGTHN NEW ENGLAND1990-2002

New England 1990 1996 2002

Connecticut 51.8 44.9 44.4*Maine 46.6 42.2 48.5

Massachusetts 40.9 40.1 31.5*New Hampshire 61.2 53.1 61.1*Rhode Island 32.0 33.1 34.8Vermont 49.1 39.5 44.1*

*MoreDemocratic n 2002 than 1990.

of New England-remained in the Republican column.Moretelling,however, s the degreeto which the Democrats

solidified their hold on the region. Four of the six statesregistered smore Democraticn 2002 than n 1990, with an

averagechangeof 2.7 points. Someof these shiftswere sig-nificant.Massachusetts,orinstance,decreased ts MPIscore

by 9.4 points while Connecticut became more Democratic

by 7.4 points. This realignmentmatchesthe ideologicaltiltto the left in New England.ManyprominentRepublicansnthe region-such as SenatorsLincolnChafee,SusanCollins,and OlympiaSnowe-are much more moderatethan their

partisancolleaguesfrom otherpartsof the country.

Given the national trend towards the Republicans,Democraticgains in New Englandare impressive-even if

they aremore modest than Republicangains in the South.

Americanpolitics scholars for years have considered the

South to be a "special ase,"but it maynow be time to con-

sider New England n thatlight. In theirvoting for Democ-

rats,some of the New Englandstatesstronglyresemblethe

states of the South during the heyday of one-partydomi-nance. In fact,New Englandqualified n 2000 as the regionthat is furthest romthe nationalaverageor midpoint.

A third and final trend that can be studied by our data

relates to the complex relationshipbetween national and

state-levelvoting. This analysisproceeds by breakingthe

MPIdown into two sub-indexes,one forfederalofficesand

one for state offices. In compiling these sub-indexes the

executive and legislative branches are weighted equally.These two indexes replicate, n effect,the intentions of the

Cox Index (federal)and the RanneyIndex(state).Our con-

cern here was to study the states havinga "schizophrenic"

profile,which we defined as having:(a) an MPIthatwas ofa differentvalencefor eachsubindex, i.e., Republicanat one

level and Democratic at another, and (b) a difference

betweenthesescores of morethan six points (the latterpro-vision is introducedto eliminatestatesthat werein a condi-

tion of essential competitivenessat both levels). A schizo-

phrenic profile indicates what amounts to a different

partisanpatternon the two levels of government.In 1990,17 states that fell into this category,whereasby 2002 the

numberhad decreasedto only ten.24This suggestsgrowing

congruence in the partisan voting patternsbetween state

and federalvoting, and it would be interesting o see if the

trend noticed here is an important movement that took

placechieflywithin the lastdecade.Itremains,of course,tobe seen, if a less schizophrenicelectorate s also a happierand healthierone.

24 The seventeen states that fell in this category n 1990 were Alabama,

Alaska,California,DelawareFlorida, Idaho, Indiana,Kentucky,Missis-

sippi, Nevada,New Jersey,New Mexico,North Carolina,Pennsylvania,SouthDakota, Texas,andVirginia.The ten states that met this standard

in 2002 were Alabama, Florida, Indiana,Kentucky,Michigan,Missis-

sippi, New York,North Carolina,Oklahoma,and Tennessee.




APPENDIX

MAJORARTYNDEX YRANKORDER

"ORDER REPUBLICANTATESSHADED), EMOCRATICTATESUNSHADED)"

1990

N. HampshireUtah

Arizona

WyomingKansas

Indiana

Delaware

SouthDakota

Alaska

SouthCarolina

Nebraska

Connecticut

Missouri

Ohio

Colorado

Idaho

IowaWisconsin

Vermont

MichiganFlorida

NorthCarolina

Montana

PennsylvaniaTexas

New Mexico

VirginiaNevada

NewJerseyCalifornia

Maine

Illinois

OregonMinnesota

WashingtonNorthDakota

1992

49.8

49.1

48.9

48.4

48.1

47.9

47.8

47.6

47.2

47.2

47.1

47

46.9

46.6

46.3

45.7

45.5

45

44.4

Utah

Kansas

N. Hampshire

ArizonaSouth Carolina

Idaho

Iowa

WyomingNebraska

Connecticut

Alaska

Ohio

Wisconsin

VirginiaNew JerseySouth Dakota

TexasIndiana

MichiganMontana

Colorado

Florida

OregonIllinois

California

Mississippi

PennsylvaniaMaine

Minnesota

Missouri

North Carolina

Delaware

New Mexico

North Dakota

Vermont

Alabama

1994

49.8

49.6

4949

48.6

47.7

47.2

46.9

46.7

45.8

45.7

45.5

45.1

44.6

44.6

44.6

43.6

43.4

43.4

42.1

41.9

41.9

Utah

Idaho

Kansas

N. HampshireWyomingOhio

Arizona

Alaska

Wisconsin

VirginiaIowa

New Jersey

MichiganSouth Dakota

Montana

South Carolina

IndianaIllinois

Florida

Colorado

Connecticut

PennsylvaniaNorth Carolina

Nevada

Minnesota

Oklahoma

Oregon

Mississippi

WashingtonCalifornia

Tennessee

New Mexico

Delaware

Maine

Nebraska

Texas

1996

49.6

49.5

49.3

49.2

49.2

48.8

48.8

48.6

48.1

47.6

46.5

46.4

Utah

Idaho

Montana

KansasWyomingAlaska

Ohio

NorthDakota

South Dakota

Arizona

Indiana

South Carolina

Iowa

Texas

N. HampshireWisconsin

Oklahoma

VirginiaColorado

Florida

Michigan

Pennsylvania

MississippiTennessee

Illinois

New Jersey

OregonNew Mexico

North Carolina

Nevada

Minnesota

Kentucky

GeorgiaCalifornia

WashingtonAlabama

1998

"Yt.O

49.5

48.7

48.7

48.5

48.4

48.2

48.1

47.9

47.4

47

Idaho

Utah

Kansas

MontanaSouth Dakota

WyomingArizona

Nebraska

Alaska

Texas

North Dakota

PennsylvaniaColorado

Tennessee

Ohio

Florida

NevadaOklahoma

MichiganIowa

Indiana

VirginiaWisconsin

South Carolina

NewJersey

Kentucky

MississippiNew Mexico

GeorgiaN. HampshireIllinois

Missouri

Connecticut

Minnesota

Maine

Alabama

2000

69 Idaho

58 7 Kansas

6.3 Wyoming

5 8 Utah5. 2 South Dakota

5:Z Arizona

758 Alaska

57.9 Texas

7 Nebraska

56.8 Tennessee

56 Oklahoma;5 9 Montana

5,f6 North Dakota

55. Ohio

54 Pennsylvania3 7 Colorado

:531 Florida529 Virginia52 5 New Mexico

l52 : Iowa

:523 Michigan5 South Carolina

:5.5 Nevada

.4 Wisconsin

49.6 Indiana

49.5 N. Hampshire49.5 Missouri

49.2 Louisiana

48.9 Maine

48.5 Georgia47.9 Mississippi47.1 New Jersey

47 North Carolina

46.9 Minnesota

46.8 Alabama

46.7 Illinois

A

-94 Uta

:82 Ida7 2 Ka

549 Wy53.8 Ari

5 2 Ala

51 2 Sou

588 N.

588 Oh

i5: Mo

569 Iow

575 Sou

564. Ne

55.; Wi

556 Ind

:;552 Tex

5.5 Vir4 - o

7; Mi532 Flo

53e Pen2 No

Ne

.i: Ten

522 Ne

Co

i51.0Ok

Ne

i Illi

49.6 Mi49.4 No

49.3 Mi

49.0 Mi

48.8 Del

48.4 Or48.0 Ma



APPENDIX(continued)

1990 1992 1994 1996 1998 2000 A

Tennessee 44.1 Nevada 41.9 New York 46 Nebraska 46.5 North Carolina 46.5 Arkansas 47.6 Ca

Alabama 42.7 Tennessee 40.8 Missouri 45.8 Delaware 46.2 New York 46.2 New York 47.5 Ala

Mississippi 42.7 Kentucky 40.5 North Dakota 45.4 Missouri 46 Delaware 45.1 Delaware 47.3 WaNew York 42.4 Georgia 40.1 Alabama 45 Louisiana 45.5 Washington 44.8 Kentucky 46.2 Ke

Kentucky 41.8 Oklahoma 39.7 Kentucky 43.9 Connecticut 44.9 Louisiana 44.7 Connecticut 45.7 GeMassachusetts 40.9 Washington 39.1 Georgia 43.9 New York 43.2 Arkansas 44.6 Washington 45.5 Ne

Oklahoma 39.9 New York 38.5 Massachusetts 42.3 Maine 42.2 Oregon 44 Oregon 44.6 Ver

Georgia 39.3 Massachusetts 37.3 Maryland 41.2 Maryland 40.6 California 42.1 Vermont 43.4 Lo

Maryland 38.4 Maryland 35.1 Vermont 40.3 Massachusetts 40.1 Vermont 39.5 California 40.8 Ar

Arkansas 35.3 Louisiana 33.5 RhodeIsland 36.6 Arkansas 36.9 Maryland 36.1 WestVirginia 36.0 MaWestVirginia 34.1 RhodeIsland 32.6 WestVirginia 34.8 Vermont 35.4 Massachusetts 34 Maryland 35.4 Ma

RhodeIsland 32 Arkansas 31.3 Louisiana 34.1 WestVirginia 35.1 WestVirginia 33.2 Rhode Island 34.9 We

Hawaii 31.5 WestVirginia 30.7 Arkansas 33.1 Rhode Island 33.1 Rhode Island 32.8 Hawaii 34.9 Rh

Louisiana 29.7 Hawaii 29.4 Hawaii 32.1 Hawaii 32.7 Hawaii 31.9 Massachusetts 314 Ha



A NEW MEASURE CE PARTYSTRENGTH 255

APPENDIX 2

2002 MAJOR PARTYINDEX (PERCENT REPUBLICAN)

President Governor U.S. House U.S. Sen State Sen StateHouse Score

Idaho 70.9% 57.4% 65.1% 68.9% 80.0% 77.1%Alaska 67.9% 57.8% 81.3% 83.8% 57.8% 67.5% 67Utah 71.7% 59.6% 60.5% 66.8% 75.9% 74.6%

Wyoming 71.0% 49.0% 62.1% 74.9% 66.7% 75.0% 548Kansas 61.1% 46.0% 68.1% 83.7% 75.0% 64.0% 631New Hampshire 50.7% 60.6% 59.2% 61.4% 75.0% 70.3%South Dakota 61.6% 57.5% 54.0% 43.4% 73.5% 70.0%Texas 61.0% 59.1% 50.2% 61.5% 66.3% 58.6% 59.6Nevada 51.9% 75.6% 60.1% 54.0% 61.9% 45.2% 59.5Ohio 52.0% 60.2% 57.6% 59.6% 66.7% 62.6% 58 9Arizona 52.9% 49.0% 55.2% 85.8% 56.7% 65.0% 583

Virginia 54.0% 47.4% 63.2% 76.2% 57.5% 66.3% 58 2NorthDakota 64.6% 55.0% 47.6% 37.1% 70.0% 70.2% 58.Colorado 54.5% 65.2% 55.3% 58.5% 51.4% 56.9%Montana 63.6% 52.0% 66.0% 42.7% 58.0% 53.0% 56.4South Carolina 58.2% 53.0% 61.8% 50.9% 54.3% 58.9% 56w0Florida 50.0% 56.6% 56.0% 42.5% 65.0% 67.5% 555Oklahoma 61.1% 49.6% 59.6% 64.5% 41.7% 47.5%Nebraska 65.5% 71.5% 88.6% 67.0% 53 7Indiana 58.0% 42.4% 56.3% 51.4% 64.0% 49.0% 52 /Missouri 51.7% 49.5% 53.8% 49.8% 58.8% 55.2% 525Iowa 49.8% 45.8% 55.0% 57.0% 58.0/ 54.0% 5. 9

Louisiana 53.9% 67.8% 64.7% 40.8% 33.3% 32.4% 518

Mississippi 57.8% 49.4% 51.6% 83.8% 34.6% 29.4% 51 7Minnesota 48.7% 55.0% 47.8% 49.2% 46.9% 61.2%

Pennsylvania 47.9% 45.5% 55.1% 58.7% 58.0% 53.7% 51 5Georgia 56.0% 52.70 53.8/ 46.50 53.6/ 40.8/ 5 5

North Carolina 56.5% 47.2% 55.9% 51.2% 44.0% 50.0% 5 I

Wisconsin 49.9% 47.8% 56.0% 43.2% 54.5% 58.6% 5 .Tennessee 51.9% 48.4% 53.1% 61.1% 45.5% 45.5% 7Alabama 57.5% 50.1% 54.6% 61.4% 28.6% 40.0% SCMichigan 47.6% 48.0% 46.6% 43.9% 57.9% 57.3% 49.6

Oregon 49.8% 48.6% 43.9% 47.5% 50.0% 58.3% 49.6Delaware 43.1% 40.1% 73.0% 42.6% 38.1% 70.7% 48.9Maine 47.4% 46.8% 41.7% 63.7% 48.6% 45.6% 48.5

Kentucky 57.7% 26.8% 68.9% 57.4% 55.3% 35.0% 48.2New York 37.2% 59.7% 41.3% 42.6% 59.7% 31.3% 46.1

Washington 47.0% 40.5% 46.1% 45.8% 51.0% 46.9% 45.6Illinois 43.8% 46.6% 46.6% 45.2% 45.8% 44.1% 45.3Vermont* 44.6% 51.2% 33.2% 47.9% 36.7% 51.7% 45.1Connecticut 40.7% 56.1% 47.8% 34.2% 41.7% 37.7% 44.4

NewJersey 41.8% 42.5% 42.1% 46.7% 50.0% 45.6% 44.1Arkansas 52.9% 53.3% 42.8% 44.8% 22.9% 30.0% 44.1New Mexico 50.0% 41.0% 37.7% 51.7% 41.5% 38.6% 43.9California 43.7% 46.8% 41.7% 42.1% 35.0% 40.0% 42.5WestVirginia 53.1% 48.5% 29.8% 28.7% 29.4% 32.0% 40.4

Maryland 41.4% 51.7% 35.1% 33.1% 29.8% 30.5% 39.3Hawaii 40.2% 52.3% 33.6% 21.8% 20.0% 29.4% 36.2RhodeIsland 34.2% 54.7% 30.4% 39.8% 15.8% 14.9% 34.8Massachusetts 35.3% 52.6% 13.1% 33.7% 15.0% 14.5% 31.5

A NEW MEASURE OF PARTYSTRENGTH 255



POLITICALESEARCHUARTERLY

REFERENCES

Black,Earl,and Merle Black. 1992. TheVital South.Cambridge,MA:HarvardUniversityPress,

Bullock, Charles S. III. 1988. "RegionalRealignmentfrom an

OfficeholdingPerspective."ournalofPolitics50 (August):3.

Cox, Edward E 1960. "Measurement f PartyStrength."Western

PoliticalQuarterly 3 (4) 10: 22-42.David,Paul T. 1972. PartyStrengthn the UnitedStates:1872-1970.

Charlottesville:Universityof VirginiaPress.

Hofferbert,Richard . 1964. "Classificationf AmericanStateParty

Systems." ournalofPolitics26 (August)3: 550-67.

Keith,BruceE.,DavidB.Magleby,Candice .Nelson,ElizabethOrr,MarkC. Westlye,andRaymondE.Wolfinger.1992. TheMythoftheIndependentoter.Berkeley:Universityof CaliforniaPress.

Key,V O. SouthernPoliticsn Stateand Nation.New York:Knopf,

VintageBooks.

King,James D. 1989. "InterpartyCompetitionin the American

States:An Examinationof Index Components."WesternPoliti-

cal Quarterly 2: 1 .

Lochner,Toddand GaryMoncrief.2003 "Idaho:ElectoralReform

at the Margins."ElectionReform:Politicsand Policy Confer-

ence. May29.

Ranney,Austin. "Partiesn State Politics."In HerbertJacob and

KennethN.

Vines, eds.,Politicsn the American tates:A Com-

parativeAnalysis,1st ed. Boston,MA:Little,Brown.

Received: January 28, 2004

Accepted for Publication: April 22, 2004

[email protected]

256

3595626_A New Measure of Party Strength

Documents

Transcript of 3595626_A New Measure of Party Strength