Southern Political Science Association

Transcription

1 Suthern Plitical Science Assciatin A New Partisan Vter Authr(s): Jseph Bafumi and Rbert Y. Shapir Surce: The Jurnal f Plitics, Vl. 71, N. 1 (Jan., 29), pp Published by: Cambridge University Press n behalf f the Suthern Plitical Science Assciatin Stable UR: Accessed: 25/3/213 19:1 Yur use f the JSTOR archive indicates yur acceptance f the Terms & Cnditins f Use, available at. JSTOR is a nt-fr-prfit service that helps schlars, researchers, and students discver, use, and build upn a wide range f cntent in a trusted digital archive. We use infrmatin technlgy and tls t increase prductivity and facilitate new frms f schlarship. Fr mre infrmatin abut JSTOR, please cntact supprt@jstr.rg.. Cambridge University Press and Suthern Plitical Science Assciatin are cllabrating with JSTOR t digitize, preserve and extend access t The Jurnal f Plitics. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

2 A New Partisan Vter Jseph Bafumi Rbert Y. Shapir artmuth cllege Clumbia University The American electrate tday is different frm that described in The American Vter. Bth the 195s era f idelgically inncent party vting and the subsequent perid f partisan dealignment are ver. Sme plitical scientists began t describe the New American Vter as a new partisan evlutin ccurred. What has nt been fully appreciated in the twentieth/twenty-first century histry f vting studies is hw partisanship returned in a frm mre idelgical and mre issue based alng liberal-cnservative lines than it has been in mre than 3 years. This is visible in the strength f partisan vting, in the relatinship between partisanship and idelgy, and in the strength f the relatinship f partisanship and self-reprted liberal-cnservative idelgy t the public's ecnmic, scial, racial, and religius attitudes and pinins. Nt nly has the public respnded in a striking way t changes in plitics and its cntext, but the current transfrmatin has als appeared t be strikingly enduring and difficult t shake, based n survey evidence fr this new partisan vter. he nging debates abut "plitical plarizatin" r "partisan plarizatin" and the natin's Republican "red states" and emcratic "blue states" have pinted t an American plitics and electrate very different frm thse described in The American Vter (Campbell et al. 196). An increasing number f schlars have tracked the changes that have ccurred and have begun t assess their causes and implicatins fr American plitics and plicymaking (Firina and Abrams 28; Nivla and Brady 26, 27). t is nw time t take stck f the lng-term changes that have ccurred in the American electrate. We first review questins asked and answered in the research that began with The American Vter's analysis f pstwar plitical behavir. Fast frwarding, we shw that while 195s-style partisan vting lks alive and well, the 195s' idelgically inncent party vting is ver. Whatever partisan dealignment stensibly ccurred after the 195s als ceased, althugh independent vters have remaine decisive in determining electin utcmes. Sme plitical scientists began t describe a New American Vter (e.g., Miller and Shanks 1996) as a new partisan evlutin ccurred. What has nt been fully appreciated in the twentieth/twenty-first century histry f vting studies is hw partisanship has returned in a frm that is bth mre idelgical and mre issue based alng liberal-cnservative lines than The Jurnal f Plitics, Vl. 71, N. 1, January 29, Pp. 29 Suthern Plitical Science Assciatin it has been in mre than 3 years. This is visible in the strength f partisan vting, in the relatinship between partisanship and idelgy, and in the strength f the relatinship f partisanship and self-reprted liberal-cnservative idelgy t the public's ecnmic, scial, racial, and religius attitudes and pinins. Nt nly has the public respnded in a striking way t changes in plitics and its cntext, but the current transfrmatin has als appeared t be enduring and difficult t shake, based n survey evidence fr this new partisan vter. The Changing, Unchanging, and New American Vter While it is impssible t summarize mre than 6 years f survey research n the American vter, tw f the mst imprtant debates in this researc have cncerned the "demcratic cmpetence" f the American public and hw vters are influenced by lnger-term partisan factrs and shrter-term electin-specific factrs. We are cncerned with the secnd debate, althugh it has an imprtant bearing n the first. While usually cntrasted with each ther, the presidential vting studies led by Paul azarsfeld and his clleagues at Clumbia University's Bureau f Applied Scial Research (azarsfeld, Berelsn and di: 1.117/S SSN This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

3 2 JOSEPH BAFUM AN ROBERT Y. SHAPRO Gaudet 1944; Berelsn, azarsfeld and McPhee 1954) entry f a new generatin f vters. The relatinship and Angus Campbell and his clleagues at the emerg- between partisanship and vte chice thus appeared ing Survey Research Center at the University f t weaken. The authrs als gt bgged dwn in a Michigan (Campbell et al. 196) shared an interest messy debate abut whether the plitics f the 196s in and emphasis n the partisan-emcratic versus made vters mre idelgically attuned and cnsis- Republican bases f vting. tent r "cnstrained" the Cnverse (1964) sense. When azarsfeld's grup did nt find the shrt- One change frm the 195s that Pmper emphasized term campaign and cmmunicatin effects they had was that the grunding f partisanship in pinins expected, they fcused n the mre prevalent sci- n plicy issues became mre apparent, s there was ecnmic bases f partisanship and the imprtance a clear substantive plicy dimensin t self-identified f interpersnal cmmunicatin that reinfrced the partisanship. These findings reflected better n the scilgical influences n vting r prvided new electrate's cmpetence, and the argumenthat infrmatin n the current campaign. n cntrast, plitical cntext mattered was persuasive, especially Campbell et al. (196) emphasized the deeper psych- as data frm 1972 t 1976 shwed that by 1976, the lgical aspects and influences f partisanship invlv- effects f plitical cnflict-ver civil rights, the ing the enduring effect f the New eal realignment Vietnam War, law and rder, and ther issues-wre (and its sciecnmic bases), affect, generatinal ff smewhat (Nie, Verband Petrcik 1979; Pmper transmissin, and hw, in effect, psychlgical bal- 1975). ancing and aviding dissnance led vters t adhere S had the American vter changed in any t partisan predispsitins at electin time. Electin- fundamental way after all? One immediate respnse specific variables, such as candidate characteristics t Nie, Verba and Petrcik (1979), Pmper (1972, and specific majr issues (e.g., war, the ecnmy), 1975) and thers was that methdlgical and data had smaller effects, althugh they culd be decisive in limitatins raised questins abut whether any sigprducing deviatins frm the partisan balance in the nificant change ccurred in the 196s cncerning the electrate. public's idelgical thinking and the issue cntent What was striking abut the findings f these f partisanship. Subsequent t The American Vter studies was nt the centrality f partisanship but the (196), the Natinal Electin Study (NES) changed limited effect f plicy issues n vting. This, alng its questin frmat, s "changes" in the 196s and with vters' limited factual knwledge, helped t 197s culd have been artifacts f the differences in spark the debate regarding the public's plitical measurement. Smith's (1989) The Unchanging Amercmpetence, which the next wave f vting studies ican Vter cntributed significantly t this debate and examined further. Sme f these studies argued that critique. Further, Pmper's findings f the increasing the findings f the 194s and 195s may have been relatinship between partisanship and plicy pin- "time-bund." Pst-New eal electins t the 196s ins were based n nly six plicy questins. S what were, relatively speaking, nt idelgically tumultu- culd we cnfidently say abut the changing Amerus, in cntrast the 193s r earlier perids (fr ican vter by the mid-197s? Prbably that partisanwhich adequate natinal survey data were nt avail- ship was smewhat less imprtant than in the 195s, able). The 196s perid saw expanded plitical cn- thathere was sme evidence fr a weakening f party flict precipitated by the civil rights mvement, the ties-a dealigning in the electrate-and that greater Vietnam war prtests, and ther emerging left-right changes were pssible if the plitical cntext changed idelgical cnflicts. further, especially in a systematic and sustained way. Tw imprtant and widely debated plitical sci- When Miller and Shanks (1996) revisited The ence wrks that examined the effects f the new American Vter in The New American Vter, they plitical cntext were The Changing American Vter emphasized the cntinued and increasingly impr- (Nie, Verba and Petrcik 1979) and the "ssue Vt- tant rle f partisanship (see als Hetheringtn ing" sympsium in the 1972 American Plitical Science (21), Green, Palmquist and Schickler (22) and Review, led by Pmper's (1972) "Frm Cnfusin t Bartels (2)) alng with electin-specific cncerns Clarity" (see als Pmper (1975)). Nie, Verba and such as plicy preferences, candidate evaluatins, Petrcik (1979) shwed a decline in the number f perceptins f current cnditins, and retrspective party identifiers as the number f self-identified evaluatins, all f which had been studied extensively independent vters increased. This culd be related since the 195s. One further cnsideratin that Miller t disenchantment with the tw majr parties, the and Shanks placed n the same stage in their causal lapse f time since the New eal realignment, and the sequencing f variables were "plicy predispsitins" This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

4 A NEW PARTSAN VOTER 3 and-whether separate r part f these predispsi- The New eal divisins were transfrmed as new tins-liberal-cnservative idelgy, as measured by issues came t the fre in American self-placement alng a plitics and seven-pint scale. The NES public discurse. As nted abve, during the 196s started t measure this in the 197s (related t the and 197s, members f the electrate became less empirical study f spatial theries f vting). T the bund by past partisan lyalties (and thse f their extenthat Miller and Shanks (1996) saw idelgy as parents) as the effect f the 193s realignment faded imprtant, it had t d with a general verarching and new issues, cnflicts, and resulting cleavages liberal/cnservative perceptin triggered by nging emerged (Nie, Verba and Petrcik 1979). These plitics and nt the pressures tward idelgical changes have given way t an electrate that is mre cnstrainthat Nie, Verba and Petrcik (1979) and strngly driven by liberal/cnservative idelgical Smith (1989) had tracked with n clear cnclusin. cncerns (Abramwitz and Saunders 1998). This Perceived liberalism-cnservatism f this srt culd idelgical psitining has been driven by a set f be analyzed in all subsequent NES and ther surveys, new issues (racial, scial, religius) and by leadership as researchers acknwledged that a large segment f that has prduced visible partisan divisins (Carmines the public was able t understand and respnd t and Stimsn 1989; Niemi and Jennings 1991; Adams these labels (e.g., Knight and Eriksn (1997)). 1997; Wlbrecht 2; ayman 21). These divisins have been mre pervasive and enduring than any changes that have ccurred since the 195s. n the rest f this paper, we systematically A New Partisan Vter examine what has ccurred. While sme f this verlaps with the imprtant research and writing f T what extent is the American vter in the early 21st thers, we have framed ur analysis in the lnger century different frm the American vter f past histry f the study f the American vter, and we decades? While plitical scientists in the 196s were examine bth partisan and related influences n limited in the extent which they culd reliably track vting. We present sme f the latest available data changes in idelgically based partisanship and vt- situated within ther recent research. After presenting, we are better able nw t examine what has ing evidence fr a level f partisan vting that is happened in the last 3 years. As we bserved at the unparalleled since the 195s, we examine t what utset, current debates regarding "partisan plar- exten this is a new srt f partisanship-ne that is izatin" and red state/blue state plitics describes substantively different frm partisanship f the past. an American plitics and electrate that are different We find that this partisanship has vters mre frm thse described in The American Vter. There is strngly anchred than ever befre by left/right ne similarity, hwever: the imprtance f partisan- idelgical thinking. This idelgy is still steeped in ship. But the plitical cntexts are very different. The ecnmic issues, but it has becme increasingly 195s was a perid in which there was a dmestic rted in scial issues and religius values. t als, cnsensus n an enlarged American welfare state even mre s than earlier, has an imprtant undercmpared t the pre-new eal era and a Cld War pinning in racial issues.' cnsensus in freign plicy. American plitics is currently situated at a transfrmatin that has brad implicatins fr American plitics. The natin's plitical parties, at the elite and activist level, have Resurgent Partisanship becme mre idelgically cherenthan they were in the mid-197s. The partisan plarizatin that has The evidence that partisan and idelgical plarccurred at the elite level (Rhde 1991; Aldrich 1996; izatin has increased in the United States since the McCarty, Ple and Rsenthal 26) has becme 197s can be fund in measures f interparty diverincreasingly evident in the mass electrate (Bartels gence and intraparty cnvergence in legislative be- 2). The strength f party identificatin in predict- havir, which have reached levels unseen in 6 years ing the vte has grwn cmparable t, if it has nt (Rhde 1991; Aldrich 1996; McCarty, Ple and exceeded, what it was in the era f party vting, the Rsenthal 26). The relatinship between elites 195s (Campbell et al. 196). Althugh its predictive and mass public pinin is a dynamic ne in which strength is reminiscent f anther day, plitical,t may extend t what used t be thught f as nn-partisan partisanship tday is f a different srt (Hetheringtn freign plicy, but this is beynd the scpe f this paper (Blch- 21). Elkn and Shapir 25; Shapir and Blch-Elkn (26, 27). This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

5 4 JOSEPH BAFUM AN ROBERT Y. SHAPRO we wuld cnjecture that elite level plarizatin FGURE 1 Standardeviatin f the seven-pint might either lead t, r result frm, changes amng partisan identificatin self-placement the mass public.2 Either way, we wuld expect t see item frm 1952 t 22. The evidence f public pinin plarizing alng partisan variability in partisanship begins very and idelgical lines. Where elite level plarizatin high but takes a dwnward turn leads, we wuld expect mre clearly defined platbeginning in the mid-196s t the 197s. Plarizatin then frms and diverging issue stances ver time between reemerges the emcratic and beginning in the 198s. Republican parties-and especially their leaders in gvernment-t cntribute t O cuj plarizatin amng partisans in the public at large. T what extent, then, has the American electrate -u plarized alng party lines and in idelgically a definable ways? The evidence that this has ccurred -l O CU is striking, beginning with what is suggested by trends in partisanship and vting in presidential electins. First, a simple graph f the standar deviatin in seven-pint partisan identificatin taken frm the Natinal Electin Studies (NES) data is telling. Figure plts the standard deviatin, as a prxy fr plarizatin f party identificatin in the mass public, ver all years in which NES asked the questin frm 1952 share f the electrate tended t be t 24, the last available survey (there was n 26 nticeably higher (abut 15% in 1976).4 NES cngressinal electin study).3 As the figure shws, partisan plarizatin f this srt has made a Figure 2 ffers further evidence fr this, shwing nticeable cmeback in recent years. n the the trend fr strng, weak, and independent partisans beginning as well as fr f the series, plarizatin is quite high, and then, in pure independents. Beginning with pure the mid-196s, it begins t drp ff independents, we see that their ranks were lw in the substantially. 195s, increased substantially in a perid f partisan By the 198s the trend reverses and the standardeviatin increases. As f 24, the level f dealignment, and shrank back dwn again in the late partisan 197s. f vters are plarizatin has nt reached the heights f the becming mre partisan, we wuld 195s, but it has risen t well beynd what ne expect declines in pure independents t result in increases in wuld expect if partisan dealignment had endured. t independent partisans. This is evident fr bth emcrats and may be that this plarizatin is driven by a small Republicans. Since the 197s, segment f the public, while the rest remain mre independent partisans have grwn substanneutral independents. Hwever, the evidence tially. Meanwhile, at the extremes f the scale, strng suggests that this is nt the case. Fr example, accrding t the Republicans have grwn substantially and strng emcrats have trended NES, the number f pure independents amng vters slightly upward since the 197s. This is in the 24 presidential electin was clse t twice as particularly revealing in light f the fact that the number f emcrats relative t Republicans many (abut 1%) as in the 1952 electin (arund in the electrate has declined ver this time, 5%), whereas during the intervening perid, this leading t clsely balanced prprtins f partisans. ast, weak Republicans have remained fairly stable ver 2This interesting and enduring questin in plitical science asks: time, while weak emcrats have declined, suggest- elites change first and then the general public fllwsuit, r ing that this is where emcrats have lst supprt. d elites realign themselves fr electral purpses in respnse t Generally, the number f partisans has grwn, while newly emerging r widening cleavages amng the American fewer Americans place themselves in the middle f public? argely, it will depend n the issue. Fr example, psitins n racial issues may be mre tp-dwn, while mve- the scale. We see, then, that grwing plarizatin is ment by the religius right may be characterized as bttm-up. evident, but t what extent has this increase in See, fr example, Sundquist (1983); Carmines and Stimsn partisanship influenced hw peple vte? (1989); Jacbs and Shapir (2) fr a discussin f elite/public interactins. The authrs f The American Vter first em- 3The data are frm the American Natinal Electins Studies phasized in sweeping terms the imprtance f (NES) cumulative file. Fr descriptive statistics n party identificatin and all ther individualevel variables used thrughut 4Pure independents are thse respndents wh placed themselves the study, see appendix A. in the middle f the seven-pint partisanship scale. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

6 A NEW PARTSAN VOTER 5 FGURE 2 Plt f respnses t the seven-pint NES partisan self-placement item. emcrat Pure ndependent CO) C) O CY) d 6 d, Weak C', % %C\C Republican i Strn ''.a O. 6.. ndependent C 6." ndependent partisanship in explaining and predicting the vte was strng in the early pst-wrld War perid, but and affecting hw peple perceive and react t it declined smewhat as a dealigning perid ccurred plitics (Campbell et al. 196). This ptent influence until abut the end f the 197s. Beginning the f partisanship seemed less relevant as evidence f 198s, the effect f partisanship began t grw dealignment was fund in the 196s and 197s substantially as a predictr f the vte. By the 24 (Wattenberg 1994). This raised questins fr plitical presidential electin its effect was n a par with r scientists wh thught party allegiances served many exceeded its impact in the 195s. At the mean f the imprtant galvanizing and mbilizing functins and prbability curve, a ne unit change in partisanship helped t fster prcesses f plitical representatin resulted in abut a 3 percentage-pint shift in the in a republican demcracy (Aldrich 1996). Multi- vte frm a emcratic t a Republican candidate in variatevidence indicates that partisanship, cntrl- the firstw (1952, 1956) and last tw (2, 24) ling fr demgraphic variables, has grwn substantially electinstudied, hlding ther variables cnstant. as a predictr f the vte since the dealignment perid This cntrasts with an analgushift f abut 2 f the late 196s and 197s (Miller and Shanks 1996). percentage pints in Partisan vting has grwn Figure 3 presents a series f lgistic regressin significantly since the perid that was thught t be cefficients (and their standard errrs) fr predic- part f a ptentially lnger-term dealignment. tins f the vte fr the Republican presidential But hw imprtant, verall, is the rle f particandidate in each presidential electin frm 1952 t sanship? Hw much difference des partisanship 24.s The cntrl variables in this multivariate make cmpared t ther predictrs f vte chice? analysis include sex, age, educatin, religin, incme, f partisanship matters a great deal, what is the regin (suth), and a statistical interactin term t prcess explaining these changes ver time? Few allw fr a differing effect f partisanship fr white wuld disagree with the imprtance f partisanship sutherners, wh have undergne a majr shift in as a predictr f the vte.8 Even casual bservers f partisan allegiance frm the emcratic t the Republican party.6 The effects f mst predictrs in this multivariate analysis are dampened by the inclusin 7This is ften called the marginal effect and is equal t the slpe f the f party identificatin. The effect f partisanship itself prbability curve at its mean. Other predictrs are als held t their mean. 5Each year represents a separate regressin equatin. 6The data are frm the American Natinal Electins Studies (NES) cumulative file. Republican vters are cded 1 while emcratic vters are cded in the utcme variable. Partisanship is measured n a seven-pint scale. Age is divided by 1 s that age squared has a reasnable range. 8Much researchas fcused n the stability f partisanship as a series. Fr example, researchers ask whether it can be cnsidered an exgenus plitical measure r nt. While individuals' partisanship based n panel data studies has been shwn t be ne f the mst stable plitical rientatins attitudes (Cnverse and Markus 1979; Green, Palmquist and Schickler 22), there is sme evidence f shrt-term fluctuatins (Firina 1981; Franklin and Jacksn 1983; MacKuen, Eriksn and Stimsn 1989). This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

7 6 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 3 gistic regressin predicting presidential vte chice frm 1952 t 24. Each year represents a separate regressin equatin. The missing parameter estimate fr whites, sutherners and their interactin in 1964 can be explained by a cllinearity prblem stemming frm all blacks in the NES sample vting emcratic in that year. After a lull, the effect f partisanship has grwn t r exceeded 195s levels. Female Grades 9-12 Sme Cllege Cllege r Mre Cathlic -. m C.a., 6 6-i a ) )-'1 1' \T O T Jewish Rel-Other Age Age Squared ncme 6Oi- C:) n d 6 6O c c;6 6 SO 1.1 UO a cr 6 T T-- T-- T-, White Suth White*Suth Party,,h cucu c'j electral plitics readily nte the nearly unanimus supprt that strng partisans give t their party's presidential candidate. The extent which partisanship matters may nnetheless be surprising when cmpared t ther characteristics f vters. Figure 4 shws the explanatry pwer f a multivariate versus a bivariate vte chice equatin predicting the vte.9 The full vte chice equatin includes all the predictrs listed abve. The bivariatequatin includes nly party identificatin. What we see is that the equatin that includes all demgraphic predictrs rarely has a much better fit than the equatin with party identificatin alne. Even in the stensibly weak days f party vting (1972), the full vte chice equatin explains nly abut 13 percent mre f the variability in the vte than party identificatin alne. 9The explanatry pwer is defined as 1-(deviance/null deviance) and is labeled "Pseud R Squared." The deviance is equal t -2 times the lg likelihd. Clearly, party identificatin is the wrkhrse in the series f regressins viewed here. nterestingly, terms f explanatry pwer, party identificatin reaches its highest level in 1996 and 24, nt in the early perids f the series. A New Partisanship Hw d we explain apparent fluctuatin in the pwer f party identificatin t predict the vte? Again, the plitical histrical cntext prvides mst f the answer. n the mid-2th century, the cuntry had just survived years f severe ecnmic depressin fllwed by a wrld war. The depressin era spurred a majr realignment in the grup bases f party supprt (invlving immigrants, urban residents, black Americans, sutherners, blue-cllar wrkers, and thers) that weighed heavily in favr f the This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

8 A NEW PARTSAN VOTER 7 FGURE 4 Variance explained in presidential identificatin and the cntrls listed earlier.12 Parenvte chice equatins frm 1952 t tal party was asked with the same questin wrding 24. The pints labeled "FU" in NES during fur years frm the 195s t the shw the variance explained after dealignment perid. Althugh this des nt cnstiaccunting fr race, gender, tute a lng series, the effects f the demgraphics educatin, age, incme, party trend as we identificatin and regin while the might have expected. Fr example, females became mre pints labeled "P." shw the likely t identify with the variance explained with emcratic nly party party ver time, whereas white suthidentificatin. t is clear that erners became much mre party likely t self-identify as identificatin is the wrkhrse in the Republicans. Thus, even with a shrt series, shifts in first equatin but is weakest during the pwer f parental party t predict partisanship the perid f partisan dealignment. can be infrmative. n 1958, during the perid in which the impact f scializatin n partisanship expected t be clearest, the party f the respndent's parents is a strnger predictr f the vte cmpared t a decade r mre later. A difference f ne cz ) FU FU categry in parent's party affiliatin (n a five pint C) 5 FU FU FU FU FU P FUP. P P FU F FU scale) in 1958 is assciated with a change f greater t S P. FU FU P P. RP. P. P F Fu p. -C FU than.7 n the seven-pint partisan self-placement R p. P. P.' scale. This effect declines t abut.5 in 197. C)P. Parental scializatin had its greatest impact n party identificatin early in the series, absent new realign ing issues. The plitical calm als resulted in the imprtance f party identificatin in predicting the vte, as we saw in Figure 3. Partisanship acquired frm parents helped indirectly t anchr vte chice emcrats (Key 1955).1 After the depressin and decisins during this first perid fr which we have Wrld War, the 195s were a perid f relative NES data. calm in which the New eal calitin essentially held Whatever equilibrium there was in partisanship tgether, althugh the Republican party had re- and vting did nt last. As discussed earlier, what bunded and was a cmpetitive frce in presidential fllwed was a perid f greater plitical turbulence vting." Cnverse (1964) regarded this as an era f thrugh the 196s and int the 197s. The civil rights idelgical inncence. t was the scial psychlgical mvement, the Vietnam War, scial unrest, plitical aspects f partisanship that anchred the electrate assassinatin and mre led t increased cnflict and (Campbell et al. 196). Absent new issues t shake up plitical antagnism.13 delgy began t take n the party system nce mre, partisanship remained new meanings this perid (Nie, Verband Petrcik stable and cntinued t strngly predict the vte. The 1979), and whether the ld idelgical inncence demgraphic grup-based plitics f the time, as well persisted became an pen questin. The lack f fully as the relative plitical calm, was expected t scialize cmparable data makes it difficult t cmpare the new entrants int the American plitical system int 195s with later perids in this regard, althugh the existing partisan divisins with partisan lyalty remain- aggregate pinin changes that ccurred are well ing high. Vters wuld tend t inherit r therwise knwn (Page and Shapir 1992; Mayer 1992; Stimsn take up the party attachment f their parents. 1991). New issues and the persistent salience f racial Figure 5 plts the cefficients based n a series f and civil rights issues that came increasingly t the linear regressins predicting a respndent's partisan identificatin frm the respndent's parent's party 12Each year represents a separate regressin equatin. Partisan self-placement is again measured n a seven-pint scale frm strng emcrat t strng Republican. Bth father and mther's party are cded -1 fr emcrats, fr independents and 1 fr 'OThis is ne f several perids f partisan realignment (Key Republicans. A cmpsite scale labeled parent's party is cn- 1955). structed by adding the tw. This is the variable used in the mdel. Multivariate 1T understand hw a Republican president culd be elected equatins estimated with an rdered respnse mdel shws the same results. while a partisan calitin in favr f the emcrats remained strng, see Green, Palmquist and Schickler (22). 13Fr evidence that cntext effects matter, see Bafumi (23). This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

9 8 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 5 inear regressin predicting party identificatin. The standard cntrls wrk as expected. Females have becme significantly mre emcratic ver time while suthern whites becme mre Republican. Mst imprtantly, here, parental scializatin has weakened as a predictr f partisan identificatin frm the 195s t the dealignment era. Female Grades 9-12 Sme Cllege Cllege r Mre Cathlic c*._ CMu d n cdc O - 6- d " d d d c ~j Jewish Rel-Other Age Age Squared ncme '-" da d a, - cu 66C16 (V) d, C'J OCO 6O "d _d _ O Na Suth White Suth*White Parent's Party 6 6 t - -O._( O 66 ac4 C;i d C C 6 a C; 1 d fre became part f a revised left-right idelgical spectrum at the elite level in American plitics. This is a spectrum which expanded frm the smewhat mre limited ecnmic/bigvernment aspects f New eal liberalism. This affected hw Americans related t the plitical parties and the degree f their partisan allegiances (Carmines, Mcver and Stimsn 1987). Mst visibly and mst imprtant, white sutherners grew increasingly uncmfrtable with the natinal emcratic party, as that party fully accepted the mantle f civil rights and racial equality thrughut the natin. This perid first gave way t the brief partisan dealignment in which party became a less imprtant predictr f the presidential vte, as new generatins came nt the scene and ld partisan lyalties were recnsidered. With the electin f an unabashed idelgical cnservative, President Rnald Reagan, in the 198s, and the realignment that had ccurred in Cngress (cnservative suthern emcrats declined in number and liberal nrthern Republicanism was n the wane as well), the resurgence f partisanship began. The issues that emerged frm the 196s and 197s increasingly divided the tw majr parties, as vters srted themselves anew (Abramwitz and Saunders 1998).14 The issues that wuld further divide the tw parties included abrtin, wmen's rights, the availability f guns, religius values in plitics and gvernment, gay rights, capital punishment, envirnmental prtectin, and ther related matters. Being liberal r cnservative began t take n a mre visible and smewhat new meaning. t became mre clsely assciated with partisanship at the elite level and, as we will examine further, the level f the mass public. Figure 6 tracks respnsesince the 197s t the seven-pint liberal-cnservative self-placement scale. The midpint f the scale (representing mderates) has remained the mdal respnse, but it has been trending dwnward smewhat in a way similar t but nt as striking as the fall-ff pure ndependents. Thse saying liberal (labeled"middle" in the plt 14This is due in n small part t Rnald Reagan's success in redefining the Republican party as the party f cnservatives in 198, an effrt 1964 presidential candidate Barry Gldwater had initiated but with less success. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

10 A NEW PARTSAN VOTER 9 FGURE 6 Plt f respnses t the seven-pint NES idelgical self-placement item. iberal Mderate Cnservative " 6 6 CO CO CO C d SlightlSlightly d d a SlMiddledle C5C... Middle Extremely 6 Extremely since it is in between the respnses slightly and wrk as befre, the effect f idelgical self-placement extremely liberal) have increased smewhat since appears t have increased frm its earliest measurethe 199s. The main change, in tandem with Repub- ment in As new and ld issues srted partisan lican partisanship, has been the increase in thse attachments anew, the public increasingly linked hw calling themselves extremely cnservative s that at they saw themselves idelgically with their partisan the macr level, verall, we see a clear cnnectin identificatin (uskin, Mcver and Carmines 1989). between trends in partisanship and idelgy. s this, Unfrtunately, the idelgical self-placement questhen, reflected at the micr level? tin was nt asked in the NES surveys befre Figure 7 presents the individualevel evidence f There is, hwever, a useful and the increasing imprtance f idelgy in lnger lngitudinal predicting series in the frm f a partisanship.15 Again, a series f linear cmpsite liberal/cnservative regressin "feeling thermmeter" measure. The thermmeter cefficients are estimated ver time. delgical self- scre is based n tw questins in which respndents placement n a seven-pint scale can nw be in- were asked t place liberals and cnservatives a cluded as a predictr, since this measure has been 1-pint scale, depending n their degree f "ht" included in the NES since the 197s. While it is r "cld" affectward each grup.16 This measure tempting t interpret it as fully causal in its effect n can serve as a reasnable prxy fr left/right idepartisanship-which we d nt think it is-fr ur lgical rientatins. When idelgical self-placement purpses, it is sufficient t examine the extent t is replaced with the thermmeter scre (results nt which the tw variables are increasingly intertwined. shwn), this measure is an increasingly strng pre- While the cefficients fr the varius cntrls tend t dictr f partisan identificatin beginning in the 196s. Early in the series, a ten-pint change in the scre results in abut a.4 shift in partisan self- 15As we shw belw, bth partisanship and idelgy are becmplacement. n the 199s, such a change is assciated ing increasingly imprtant in explaining change in attitudes n with as much as twice the shift. dmestic issues and this has als been shwn n freign plicy issues. There is the ptential fr endgeneity here. One culd als Thus, we see that, first, partisanship has taken n argue, f curse, that party predicts idelgy. Panel data frm the a new imprtance in predicting the vte in recent early 199s shw that changes in respndents' attitudes n issues years, and secnd, the data indicate that had a idelgy reciprcal effect n changes in their party identificatin, with a has increasingly infrmed this significant influence in bth directins (see Carsey and partisanship. delgy, ayman (26)). n cntrast, panel data, including bth dmestic and freign plicy issues, frm 2, 22, and 24 shw that the effect f changes f party identificatin and f idelgy n 16The cmpsite thermmeter scre is calculated by NES as issue attitudes verwhelms the reverse effect. This finding is fllws: first, the value fr liberals is subtracted frm 97 and that cnsistent with the view that Bush's idelgical framing f bth difference is added t the value fr cnservatives; this sum is then dmestic and freign issues has effectively plarized the way divided by 2, and.5 is added t the result; finally, the slutin is peple evaluate these issues, whether psitively r negatively, truncated t btain an integer value. The cmpsite scre alng bth partisan and idelgicalines (Snyder, Shapir and crrelates with seven-pint idelgical self-placement at abut Blch-Elkn 27; Veghte, Shaw and Shapir 27)..6 frm 1972 t 22. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

11 1 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 7 inear regressin predicting party identificatin. delgy has grwn as a predictr frm the 197s t present times. Female Grades 9-12 Sme Cllege Cllege r Mre Cathlic 1 cu 1 '-, q O,t -,t U U ) C) Jewish ReW-Other Age Age Squared ncme r d cu ( O -- c; C; O ) 1O j77 C J C: )6 72 ) O Qu c Suth White Suth*White delgy,, d 1, d , as assciated nly with the terms "liberal" and "cn- General Scial Surveys frm 1972 thrugh the mst servative", is in itself nt very infrmative.17 We need recent 26 survey. These data have been used by t knw what the issues are that give idelgical thers t study and track the relatinship between labels meaning t plitical elites and vters alike. partisanship and issue pinins, but we fcus as well These include issues assciated with the New eal n their cnnectins t idelgy and t update and the Great Sciety f the 196s, as well as values- analyses f the GSS data t 26 as we put these based cncerns that are racial, scial, and religius in findings int histrical perspective. nature. T the extent that these issue areas are f liberal and cnservative idelgical thinking, as represented by r reflect idelgy and, ultimately, well as emcratic and Republican partisan attachpartisanship, they are increasingly imprtant in ments, have been increasingly defined by racial (since understanding vting behavir. T what extent, then, the 196s) and especially scial and religius issues has pinin n these issues at the individualevel (since the 197s), then we shuld see these grups' becme increasingly related t idelgy and partisan- stances diverging ver time n these issues. We find ship? That is, t what extent has the public becme that such differences have indeed increased. This divided n these issues in idelgical and partisan takes us int the middle f the nging debate in terms? which plitical scientists have attempted t refute T examine this further, we use the available jurnalists wh verstate such divergences when lngitudinal data frm the NES cumula- examining the plitics f "red versus blue" states tive surveys and the rich data frm the NORC (Gelman et al. 27; Anslabehere, Rdden and Snyder 26; Firina, Abrams and Ppe 26). We agree with Firina, Abrams and Ppe (26) that '7Althugh its increasing relatinship with partisanship (Abra- states are a pr unit f analysis when studying mwitz and Saunders 1998) leads t different inferences abut plitical plarizatin and that a great many Americans the idelgical attentiveness f the public when cmpared t ther measure f take psitins in the idelgical center. idelgical awareness based n individuals' Hwever, pinins n specific issues (Cnverse 1964). what all this understates is that there have been real This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

12 A NEW PARTSAN VOTER 11 changes amng a prtin f the electrate cnsisting strength f liberalism-cnservatism r emcratic/ f liberals and cnservatives as well as emcrats and Republican partisanship since this des nt alter the Republicans wh have cntinued t srt themselves basic results that we reprt.'8 n racial, scial, and religius values issues. These With changes in respect t idelgy, the NES items shw public pinin and the electrate have fairly stable differences acrss the number f ecccurred ver a very extended perid, have fed back nmic welfare issues we examined. t is int party plitics, and shw n sign f interesting t reversing. The nte that fr the mst part mderates and ndependdata fr this deserve full cnsideratin. ents, as wuld be expected, fall in between liberals and cnservatives and partisans, respectively. Als, there is a tendency in sme cases fr mderates and ndependents t The ssues gravitate clser t liberals and emcrats. With respect t idelgical differences n ecnmic welfare issues, there appear t be slight We begin with the cre "big-gvernment" plicy increases in the mean differences in 24 cmpared issues that have divided the parties frm at least the t 2, including spending n welfare, spending n New eal thrugh the Great Sciety t the present. assistance fr the pr, and spending n hmeless- These dmestic, ecnmic, and scial welfare issues ness, but large differences existed r emerged earlier. als serve as a benchmark t study racial issues and There is als a grwing divisin-the largest gaps in the newer scial and religius values issues. Partisan the time series-amng Republicans and emcrats divisins n these issues have nt weakened, and n their attitudes twards gvernment's rle in there is sme evidence fr further plarizatin alng guaranteeing jbs and spending n the hmeless ver them as well. We then turn t values-laden issues. the lng term. Hwever, the differences between the Partisans, and t a lesser extent idelgues, are tw partisan grups n ther items have remained increasingly divided ver abrtin, hmsexuality, fairly cnsistent ver time. and the rle f religin in sciety. This is particularly We als examined ecnmic welfare items frm apparent since the earlier 199s, but it can be seen in the GSS. These data shw simila results and als that the early 198s fr sme issues. What we find fr 24 was a plarizing electin year, as cnservatives issues f race and civil rights is perhaps mst and liberals separated smewhat n attitudes tward surprising. While these issues sharply divided the the gvernment reducing incme differences, imparties in the 198s, if nt much earlier, they have nt prving peple's standard f living, spending n been high n the radar screen in the recent debate cities, spending n assistance t the pr, and abut "culture wars" and partisan plarizatin. spending n the natin's health, but this reversed in t tk the full frce f Hurricane Katrina and the 26. We find the same pattern, but mre dramat- Bush adminstratin's mishandling f aid and recvically and beginning earlier than 24, fr Repubery effrts t bring them again t the fre. But the licans and emcrats, with ndependents behaving rle f race in idelgical and partisan plarizatin much mre like emcrats than Republicans. Overcntinued lng after the height f the American civil all, there is evidence f sustained if nt grwing rights mvement. idelgical and partisan differences n these lngstanding dmestic welfare issues. These findings are Ecnmic Welfare imprtant, since this cntinuing and pssibly grwing surce f partisan cnflict has been under- Ecnmic welfare issues have been lng-standing appreciated in debates abut the "mral values" party "cleavage" issues since the 193s (Page 1978; issues that emerged t divide the parties n expanded Stimsn 1999; Eriksn, MacKuen and Stimsn 22). idelgicalines (anger and Chen 25). They prvide a frame f reference fr examining ther types f issues. While we find sme evidence fr Abrtin increasing party and idelgical divisins n a few ecnmic welfare issues, the predminant stry is ne Whether described in terms f mrality, "family f cnsistency and cntinuity. Sme f the illustrative values", "religius values", r "culture", these issues NES and GSS data are pltted n the Jurnal f Plitics website. We have included three trend lines 18n general, the strng partisans and idelgues differ mre frm in each f ur graphs, including idelgical mder- each ther, and when partisan and idelgical divergence ccurs, ates and ndependents. We d nt distinguish it appears t ccur mre at the extremes. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

13 12 JOSEPH BAFUM AN ROBERT Y. SHAPRO have prduced the mst visible cnflict in dmestic plitics in the United States. Firina, Abrams and Ppe (26) have challenged claims that Americans are plarized n these issues acrss "red and blue states" in the U.S., and they disagree with the interpretatins f thers regarding the magnitudes f these divisins and hw much they have increased (see Carsey and ayman (26), articles in Nivla and Brady (26)). n the cntext f sharply partisan vting and clearer idelgical divides between the parties, we find the public's increasing divisins n these values issues t be impressive. Abrtin has been ne f the mst cntentius and emtinal f these issues. Based n NES data, Figure 8 shws the grwing divisin between cnservatives and liberals n attitudes tward the legality f abrtin (in the first plt) and an even mre striking plarizatin fr Republicans and emcrats (in the secnd plt). At ne time Republicans and emcrats culd nt be differentiated n this issue, a far different picture frm that f 26. GSS data ging back further in time t the early 197s invlved mre cnditinal questins. Figures 9 and 1 track respnses t the GSS battery f questins n whether it shuld be pssible t btain a legal abrtin in a variety f circumstances. They shw a strikingrwth in the differences between the pinins f Republicans and emcrats as well as liberals and cnservatives. Acrss all these questins, idelgical and partisan plarizatin is evident. t is nt surprising that the largest differences ccur ver supprt fr legal abrtins fr reasns unrelated t the health f the mther r birth defects. These differences d nt diminish in the latest 26 survey, and in the case f partisanship the differences between emcrats and Republican reaches an all-time high, with ndependents falling clser t Republicans when the reasn is nt health related. Hmsexuality We find the same divergence, thugh smewhat less striking, fr pinins tward hmsexuality and gay rights. Figure 11 shws data frm the NES and the GSS fr idelgical and partisan subgrups. The first rw shw data frm the NES. There is slight idelgical divergence but mre substantial grwth in the partisan gap, especially cncerning supprt fr gay adptin. The time series is very shrt; it begins in the early 199s and may miss earlier signs f srting. The GSS, hwever, prvides mre extensive data. These data are shwn in the secnd rw f Figure 11. There is a relatively steady mean difference between idelgical grups tward supprt fr allwing hmsexuals t teach in schls, althugh there was a divergence fllwing a shrt perid f cnvergence in the 198s. Generally, bth grups have grwn mre accepting ver time tward hmsexual teachers. Frm 1985 thrugh 26, differences between cnservatives and liberals grew substantially in their feelings regarding the mral acceptability f hmsexual relatins. iberals have been much mre FGURE 8 Mean psitin f cnservatives/mderate/liberals and Republicans/ndependents/emcrats n whether abrtin shuld be legal. Surce: NES Cumulative File. WHEN SHOU ABORTON BE AOWE BY AW f ---- \ Cl) - WHEN SHOU ABORTON BE AOWE BY AW c c < C <R , 19. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

14 A NEW PARTSAN VOTER 13 FGURE 9 Mean psitin f cnservatives, mderates and liberals n abrtin. Surce: GSS Cumulative File. ABORTON F CHANCE OF ABORTON F WANTS NO ABORTON F WOMAN'S BRTH EFECT MORE CHREN HEATH ENANGERE c-c r-r M-c _ c c- c - < c/ cc CC C/C M c~cc(c < /\, M M C M /CC/\ C C C //M %. M M C H O. S \ -M/ C c% / C c 6- M CUc c cc c S /M :-.."- u c ' V" /' \ 4) ;, a) ci -, CU T- CU C / accepting than cnservatives f such relatins. The partisan differences are als apparent. The third and furth plts in rw 2 f Figure 11 shw that Republicans and emcrats have becme mre differentiated since the late 198s in their pinins tward allwing hmsexuals t teach, althugh these partisan differences are less than idelgical nes. Mst striking is the finding that the mean psitins f Republicans and emcrats n the acceptability f hmsexual relatins diverged sharply during the twenty-year perid frm the mid-198s t 24 and grew even wider in 26. Mderates and ndependents usually split the difference between the partisan and idelgical grups. Mral/Family Values Similarly, ther pinins related t religius, mral, r family values issues have becme mre strngly related t idelgy and partisanship. Based n the NES data, Figure 12 shws hw cnservative versus liberal, and Republican versus emcratic views n family values, mral standards, prayer in schls have becme increasingly disparate since the 198s. n fact, the schl prayer stances f Republicans and emcrats were ppsite frm the expected directin befre they grew apart in the late 198s. Mderates favr cnservatives emphasizing family values and allwing schl prayer meanwhile independents side This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

15 14 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 1 Mean psitin f Republicans, ndependents and emcrats n abrtin. Surce: GSS Cumulative File. ABORTON F CHANCE OF ABORTON F WANTS NO ABORTON F WOMAN'S HEATH BRTH EFECT MORE CHREN ENANGERE 5:R RR R-R - R t t: R - R R O RR R - R Q \ (/ R, ar R R S R R O \ R R \ - / / O A, R /_ with emcrats n family values and tlerating different mral standards. The GSS data, shwn in Figure 13, shws a clear increase in the difference between cnservatives and liberals n their cnfidence in rganized religin. Smewhat mre cmplicated (due t sme earlier cnvergence and then wider separatin), but still evident, is the grwing difference in cnservatives' versus liberals' supprt fr prayer in public schls. The grwing differences n this issue are clearer fr emcrats than Republicans, as shwn in the secnd rw f Figure 13. These partisans have als becme less alike in their cnfidence tward rganized religin. As in the NES data, mderateside with cnservatives in supprt fr schl prayer. ndepend- ents are as liberal as emcrats in their degree f cnfidence in rganized religin. ndependents are quite erratic in their supprt fr schl prayer perhaps because they are ambivalent but als wing t small sample sizes. Race and Equality We end ur analysis with issues f race and equality in the United States. Racial issues became increasingly central in 2th-century American partisan plitics after it was clear that the emcratic party, minus its ld suthern wing, had becme the civil rights party. Racial issues have been given scant attentin in the partisan plarizatin debate. These This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

16 A NEW PARTSAN VOTER 15 FGURE 11 Mean psitin f cnservatives/mderates/liberals and Republicans/ndependents/ emcrats n hmsexual relatins. The first rw reprts results frm the NES while the secnd rw reprts results frm the GSS. Surce: NES and GSS Cumulative File. AW PROTECTNG GAYS/ESBANS BE AOWE AW PROTECTNG GAYS/ESBANS BE HOMOSEXUAS TO AOPT HOMOSEXUAS AOWE TO AOPT m c Q ri d O? cq "O c" O 6 "" O 3: C --.. pr--- t12 C an C a M.._.M < O6-? C C 6 TOTECHSXU ETONT TAHmOMSEUETN '~/ Ca) AOW HOMOSEXUA FEENGS ON HOMO AOW HOMOSEXUA FEENGS ON TO TEACH SEXUA REATONS TO TEACH HOMOSEXUA REATONS C ~ FRn TO EAHCEXA c: ETNSTO cu EAH OMEXAREATON Z C HO HMAOHOEFNO c W O c R d - c UF-.,_ M M MC a, U),b cn c E c ks - \ C C( c A ) C t a i c E V: ii 4-- rrrm CCC C5 - U) cu C)ea C. Q Y issues did return t the fre smewhat after the psitin f Republicans and emcrats has substangvernment's mishandling f assistance t the large tially diverged ver time. The ne exceptin-attitudes African-American cmmunity in uisiana during tward affirmative actin-witnessed sme cnver- Hurricane Katrina, but they have been largely ignred gence in the early and mid 199s befre diverging in the plarizatin debate. ike scial and religius thereafter. Mderates tend t side with cnservatives values issues, civil rights and related issues have n these issues but they have grwn clser t liberals helped drive increases in idelgical plarizatin n ensuring schl integratin and finding unequal and, especially, partisan plarizatin. chances unacceptable. On the whle, ndependents Figure 14, tracking NES data, shws clear evi- side with Republicans n racial issues with sme dence f increasing divergence between cnservatives evidence f mvement tward the emcrats ver and liberals n the fllwing attitudes: that cndi- time. tins make it difficult fr blacks in America, that The NORC General Scial Survey data in Figure 16 blacks shuld nt have special favrs, that blacks tell the same basic stry: the mean psitins f shuld try harder, and that we shuld wrry abut idelgical and partisan grups (rw 1 and 2, respecequality in this cuntry. There is less clear divergence tively) have tended t mve smewhat in ppsite in respnses t the ther items, thugh the idelg- directins supprt fr spending mre mney t ical differences are substantial and have nt dimin- imprve the cnditins f blacks and ffering mre ished. n the case f partisanship we might expect gvernment aid t blacks, and there is n sign f that the racial attitudes f Republicans versus em- cnvergence in Mderates and ndependents crats wuld becme mre cnsistently and sharply mve frm cnservative and Republican stances different as ppnents f initiatives tward racial equality fund their desired party hme. Fr all but '19The results remain the same if we analyze subgrups such as ne survey item pltted in Figure 15, the mean whites, sutherners, r nn-sutherners. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

17 16 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 12 Mean psitin f cnservatives/mderates/liberals and Republicans/ndependents/ emcrats n family/mral values issues. Surce: NES Cumulative File. EMPHASS ON TRATONA TOERANCE OF FF SHOU SCHOO FAMY VAUES MORA STANARS PRAYER BE AOWE 9 C N (u c C C--/tC M-M c c Mc EM J J M-M - M C M :3 SM - N FM VM "- EMPHASS ON TRATONA TOERANCE OF FF SHOU SCHOO RY _cc EMPHASS ON TRATONA FAMY VAUES TOERANCE OF FF MORA STANARS SHOUSCHOO PRAYER BE AOWE R R c R >% 4 t-- O - 1 =3 C Q C/)R Rc\j / r - cc) / -- Q c- tward liberal and emcratic stances n these issues ver time. What this analysis f racial issues suggests mst is that in additin t the debate abut mral and religius values issues plarizing American plitics, nt nly has the centrality f party divisins n ecnmic welfare issues cntinued, but als the underlying rle f race as a cntinuing surce f plitical cnflict persists. Nt surprisingly, the issue f race has the ptential t resurface suddenly, as it did after Hurricane Katrina struck. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

18 A NEW PARTSAN VOTER 17 FGURE 13 Mean psitin f cnservatives/mderates/liberals and Republicans/ndependents/ emcrats n religius attitudes. Surce: GSS Cumulative File. CONFENCE N ORGANZE REGON BBE PRAYER N PUBC SCHOOS c O -MC -M CM C c -c C -\ c c c "c-c c cm - - C)C /R /\A CONFENCE N ORGANZE REGON BBE PRAYER N PUBC SCHOOS c~ cr l c )R) R C\i C Cnclusin beginning with its scial and psychlgical bases. This includes the rle f interpersnal cmmunica- With the advent f survey research and the behaviral tin and transmissins (azarsfeld, Berelsn and Gaudet revlutin in plitical science, students f American 1944; Campbell et al. 196), and then plitics have cntinually re-evaluated the individual increasingly-as we have reviewed here-its cnnectin t vters' ther plitical attitudes and preferlevel characteristics f the American vter. Frm the ences. While the apparent statistical effect fund in start, this reflectin lked at alternative explanatins surveys f individuals' partisan attachments n vtr interpretatins f vting behavir and at changes ing lked the same at the start f the current century that might be ccurring in the electrate and in as it did mre than a half-century earlier, the public pinin. Partisanship became central and has partisanship f tday's American vters is different: remained s in thesexplanatins and interpretatin, it is mre cnnected t salient plicy issues and t This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

19 18 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 14 Mean psitin f cnservatives, mderates and liberals n racial issues. Surce: NES Cumulative File. OPPOSE AFFRM GOVT ENSURE FAR SHOU GOVT ENSURE A TO BACKS SCAE ACTON TREATMENT--JOBS SCHOO NTEGRAT < C c C--C QC C C C C C CMM\ m c _ c6 -c c ( M / / C C CE a MM- - / c a -; c- \ S a / a) CON MAKE T FFCUT BACKS SHOUNT HAVE BACKS MUST TRY BACKS GOTTEN ESS FOR BACKS SPECA FAV HARER THAN ESERVE a C S ) _ \ n \/UA E\ / ca (t$ _ CC _ C / cr - / OK F SOME HAVE MORE EQUA CHANCE C a), zu) -/\ ".- a 2_ ~ YaY c)i M- a( WORRY ESS ABOUT EQUATY m - ma /-.2c ) M C C ccc C C -C - liberal-cnservative idelgical identificatin than it and in turn further fueled, an even mre prnunced was at least as far back as the 197s. partisan and idelgical cnflict amng plitical Scializatin-based partisanship f the 195s leaders in bth parties.2 appeared, albeit limited by the available data, t give way in the 196s and early 197s as new issues and 2We have fcused n vters in presidential electins. Fr ffices the internal realigning f the parties n the issue f lwer than the presidency, we have seen increased partisan vting in cngressinal races suggesting that the race frced vters t recnsider their grwing idelgical largely inherited divide is apparenthere as well. This has resulted in bth parties partisan lyalties. New vters and sme existing being equally cmpetitive fr cntrl f the Huse f Representavters rejected any party label when respnding t tive and the Senate (see Bartels (2); Jacbsn (27b)). What has pinin surveys. What fllwed frm the mid-197s happened in the cases f vting fr state and lcal executive and legislative ffices is a subject that requires furtheresearch. t the time in which we are writing (fall 28) have Current research suggests that the relatinship between state level been decades f an increasing cnnectin between partisanship and idelgy has increased substantially since the Carter individuals' expressed partisanship and their self- presidency (Eriksn, Wright and Mcver. 26) as divided gvernment utcmes and split-ticket vting have grwn in state reprted idelgy and a strnger cnnectin between electins (Firina 23). Overall, the relatinship between partibth partisanship and expressed liberal-cnservative sanship and vting in gubernatrial and state legislative electins is idelgy and the pinins f Americans n plicystrng but nt quite as strng as fr the presidency and Cngress. ncumbency cntinues t have a very substantial influence n related issues. This kind f partisan and idelgical vting, but verall there is cnsiderable variatin acrss states that srting and plarizing prcess increasingly reflected, deserves further scrutiny (see Jewell and Mrehuse (21)). This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

20 A NEW PARTSAN VOTER 19 FGURE 15 Mean psitin f Republicans, ndependents and emcrats n racial issues. Surce: NES Cumulative File. OPPOSE AFFRM GOVT ENSURE FAR SHOU GOVT ENSURE A TO BACKS SCAE ACTON TREATMENT--JOBS SCHOO NTEGRAT \ RR R R RR-R / R. RR'R / RR ( R R- R c _ RR 6 R. c;_r / - a) cm - R /1 \ "-, " - R \ 1\ CON MAKE T BACKS SHOUNT BACKS MUST TRY BACKS GOTTEN ESS FFCUTFOR BACKS HAVE SPECA FAV HARER THAN ESERVE a ) R_. ) l ( *i-_ c R, n OK F SOME HAVE MORE EQUA CHANCE WORRY ESS ABOUT EQUATY R RtR ( R-RR, An array f ecnmic, racial, and new scial and and even freign plicy.21 We refer t this as plarreligius values issues have becme aligned mre izatin because it is individuals wh cnsider themvisibly t partisanship and t liberal-cnservative selves strngest in terms f partisanship and idelgy labels and cues, prducing an increasingly issue- wh separate themselves mst clearly n plicy based and idelgically based partisan alignment. preferences and ther plitical attitudes. Self-identified emcrats r Republicans tday have S strng is this cnnectin f partisanship, been as cnsistent partisan vters as their cunter- idelgy, and issue pinins that we must be cautius parts were in the 195s era f party vting defined by in saying that partisan vting has increased since the the New eal ecnmic-based calitin and its 197s, returning t its 195s prminence. t has, in ppnents. Partisans in the first decade f the 21st the simple crrelatinal sense, even when cntrlling century have idelgical beliefs which are mre fr ther demgraphicharacteristics, but nt in the heavily defined by issues beynd ecnmic nes. sense f the psychlgical attachmenthat vters in Ecnmic issues remain very imprtant (still mst imprtant fr vters and partisans, based n sme cmpelling analyses; see Bartels (26); Anslabe- 21As nted earlier (ftntes 1 and 15), partisan plarizatin has here, Rdden and Snyder (26)), but an been tracked elsewhere n idelgi- freign plicy issues (see Shapir and Blch-Elkn (26, 27); Snyder, Shapir and Blch-Elkn cally based partisanship has been increasingly (27)), and it has been mst prnunced in the case f partisan cnnected t racial issues, certain scial values issues, differences in supprt fr the raq war (see Jacbsn(27a, b)). This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

21 2 JOSEPH BAFUM AN ROBERT Y. SHAPRO FGURE 16 Mean psitin f cnservatives/mderates/liberals and Republicans/ndependents/ emcrats n racial pinins. Surce: GSS Cumulative File. MONEY ON MPROVNG CONTONS FOR BACKS SHOU THE GOVERNMENT A BACKS? c c c ) cz -dc cc{ c c: c c c m M M c - M am " C O - m C ) - ON VS C cr Yea (1) cq 4--f C M CON /v C_' F CONTONS FOR BACKS A BACKS? R ~ R R c c\l R R \ a R : - R c R< C-.) C (~c RR R U)C c\j, rp Ya Y er i/ "1- \, a m/c c ( \ cq /1 \ a).,,-"b the 195s had-an attachment which was largely independent f any idelgical r issue pinins they held. What ther implicatins can we draw frm the existence f this new partisan vter? Fr ne, such vters cnstitute the strng base that party leaders can appeal t and are bliged t respnd t, particularly in primary electins. These vters, nce created, may cntribute t the increasingly visible partisan cnflic that ccurs at the elite level. On the ther hand, the large number f vters wh have nt srted themselves int the extremes remain the decisive, stensibly centrist, vters in electins. The level f partisan cnflicthat the cntemprary mass media thrive n and magnify will nly change if the parties put frth candidates t appeal t these mderate vters.22 Mre bradly and nrmatively, these new partisan vters cnstitute new evidence bearing n the questin f the "demcratic cmpetence" f the American vter. The critics wh referred t the apparently mindless, nn-idelgical, nn-issue driven vter that Clumbia and Michigan schlars fund in the 194s and 195s surely must change their tune. ssues and idelgy have becme deeply linked t 22The imprtance f centrist vters was clear in the 26 midterm cngressinal electins when, as shwn in the exit plls, ndependents' negative evaluatins f President Bush and the raq war drve them decisively in favr f emcratic cngressinal candidates (Jacbsn 27b, p.2). This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

22 A NEW PARTSAN VOTER 21 partisanship. Curiusly, this may invlve such a wide their existing attitudes and preferences (e.g., Bafumi range f issues that individuals link t liberal- (24); Eriksn (24); Shapir and Blch-Elkn cnservative idelgy that n ne has yet fund a (26); Wawr (26)). The new partisan vter pses substantial increase in the kind f idelgical cnsis- imprtant nrmative as well as empirical questins tency r "cnstraint" the mass public that Cnverse fr research n plitical behavir. (1964) sught t find. Fr example, Baldassarri and Anther questin that deserves attentin is why Gelman (27) have reprted fr the available NES has partisanship becme mre idelgical in recent data that there has been nly a mdest increase in times? This is a questin that cntinues t engage idelgical cnsistency acrss issues. This has, hw- schlars. Sme believe that the plarizatin f plitical ever, ccurred mre s, as we wuld expect, amng elites has set an example fr the public t fllw strng partisans and especially amng Republicans (Firina, Abrams and Ppe 26). This des nt wh perhaps mre aggressively than therstaked ut explain why elites have plarized. Perhaps there have sharp psitins n the new issues f abrtin and gay been majr party candidates wh have successfully rights (cf. Stimsn (24)). n any case, the imprtance shifted the psitin f their party n majr issues in f issues and idelgy t partisanship wuld seem t hpes f gaining electral advantage. Fr example, speak well t the plitical cmpetence f vters. Barry Gldwater became the state's rights candidate in On the ther hand, there is a pssible dwnside 1964 t attract suthern vters. Althughis campaign that may cme with strng partisanship f this srt, ended unsuccessfully, the Republican Party wuld which requires further study. At the elite level, we eventually becme the state's rights party and lse knw that plitical plarizatin has led t a high level the mantle f civil rights t the emcrats. This helps f visible plitical cnflict, ne that has reached high t explain sme f the idelgical divergence between levels f incivility, as we saw in the debate ver Republican and emcratic legislatrs in the United Clintn's impeachment, the 2 electin results, and States in the 197s. Many schlars regard race as the the raq war, as well as in the 24 and 26 electins. first issue that began t srt ut plitical elites and While incivility and plarizatin in gvernment are then vters in the tw majr parties (Carmines and nt necessarily crrelated, there is evidence that Stimsn 1989; Hetheringtn 21). Of curse, candi- Cngress-bth the Huse and the Senate-had dates have lng sught t explit cleavages that may becme less civil by the 197s, and this decline bring electral advantage (see Stimsn (24)) and yet became mre prnunced as partisan cnflict in- sme attempts are made with greater fervr and creased int the 199s (see Uslaner (2)). The success than thers. Als, the steepest grwth in the brader plicymaking cnsequences f this are nt divergence between Republican and emcratic legisfully clear and warrant furtheresearch (see Firina latrs began in the early 199s. Further, the realignand Abrams (28); Jacbs and Shapir (2)). ment f suthern cnservatives t the Republican At the level f the mass public, what may appear t be increasing cmpetence may have party (fllwing the lead f elites) tk ff in the early negative 199s but increased less befre (see data in McCarty, cnsequences. Strng partisan attitudes may lead t Ple and Rsenthal (26) and Jacbsn (27a)). rigidity f attitudes and pinins in the face f new S what explains the skyrcketing plarizatin in and credible discrepant infrmatin. Nt nly might recent times? Sme have argued fr the grwth f such new infrmatin be avided thrugh selective incme inequalities and the effects f the mass media expsure, but its accuracy and validity might be as explicatrs (McCarty, Ple and Rsenthal 26; denied as a result f "mtivated bias" r flawed Mutz 26; Jacbsn 27a; Prir 27). While these reasning r n reasning at all (see Marcus (1988); explanatins have merit, they are mre likely symp- rd and epper (1979); Taber and dge (26); tms rather than causes f partisan plarizatin. Shapir and Blch-Elkn (26)). Fr example, Parent and Bafumi (28) ffer the argumenthat might sme vters be less engaged in retrspective plarizatin by elites in the United States is largely vting, and therefre less likely t hld incumbents driven by the state f internatinal external threat. accuntable, as they becme mre strngly anchred When threatsubside, as with a uniplar wrld fr a by their partisanship, idelgy, and/r scial, racial, superpwer, elites lse incentives t wrk tgether and religius attitudes (e.g., Bafumi (24))? Will and they gain incentives t cmpete ver the allcaplarized vters be pressured mre greatly than tin f ecnmic and plitical benefits. When extervters in the past by prcesses f attitudinal balance, nal threats increase, dmestic plarizatin shuld cgnitive dissnance, r ratinalizatin as they are decline. n turn, the psturing f elites affects the expsed t new plitical infrmatin that challenges degree f plarizatin in the American electrate. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

23 22 JOSEPH BAFUM AN ROBERT Y. SHAPRO Much f ur data fit well with this explanatin cncerning when candidates can prduce and explit cleavages tward electral gains (Snyder, Shapir and Blch-Elkn 27). A slightly different thery prpsed by Jacbs and Shapir (2) argues that nce parties begin t becme mre hmgenus (perhaps because f party leaders, critical issues, diminished external threat r sme ther reasn), they can eschew public preferences t pursue and achieve plicy gals. Thus, nce the prcess f plarizatin begins, there are mtivatins that lead it t wrsen. This helps t explain sme f what ur data shw. The surces f partisan plarizatin remain an imprtantpic fr furtheresearch. Acknwledgments The authrs thank the editr, the annymus reviewers, Rbert Eriksn, Andrew Gelman, elia Baldassarri, Justin Phillips, and Eric Uslaner fr their cmments and assistance. Rbert Shapir is grateful t the Russell Sage Fundatin fr supprting this research as a 26/27 Visiting Schlar and t Clumbia University's nstitute fr Scial and Ecnmic Research and Plicy. Appendix Al: Variable Appendix A escriptive statistics fr all individual-level variables used fr regressin analysis in this study. Each variable'statistics are reprted with their maximal sample size. Fr a variety f reasns, the actual sample sizes vary thrughut the analysis. Mean Std. ev. Min. Max. N Rep. Pres. Vte Female Educatin Religin Age ncme White Party delgy Parental Party delgy Therm Appendix A2: escriptive statistics fr NES issue variables. Variable Mean Std. ev. Min. Max. N Affirm Actin Fair Treatment-Jbs Schl ntegratin Aid t Blacks ifficult fr Blacks Blacks-Special Favrs Blacks-Try Harder Blacks-ess than eserve Mre than Equal Chance Wrry ess Equality Abrtin Trad. Family Values Tlerant iff Mral Stand Schl Prayer aw Prtect Hmsexuals Gay Adptin Gv't Guar Jbs Spending Sc. Sec Spending Welfare Spending Pr Spending Fd Stamps Spending Hmeless Appendix A3: escriptive statistics fr GSS issue variables. Variable Mean Std. ev. Min. Max. N Abrt-efect Abrt-N Mre Abrt-Wmen's Health Abrt-Can't Affrd Abrt-Nt Married Abrt-Any Reasn Hmsexual Teach Hmsexual Relatins Blacks-mprve Cnd Gv't Aid Blacks Cnfid. Religin Bible Prayer Gv't-Reduce ncme iff Gv't-mprve Stand. iv. Spending Cities Spending Welfare Spending Natin's Health This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

24 A NEW PARTSAN VOTER 23 Manuscript submitted 15 February 27 Manuscript accepted fr publicatin 2 March 28 References Abramwitz, Alan. and Kyle. Saunders "delgical Realignment in the U.S. Electrate." Jurnal f Plitics 6: Adams, Greg "Abrtin: Evidence f ssue Evlutin." American Jurnal f Plitical Science 41: Aldrich, Jhn Why Parties? The Origin and Transfrmatin f Party Plitics in America. Chicag: University f Chicag Press. Anslabehere, Stephen, Jnathan Rdden and James M. Snyder. 26. "Purple America." Jurnal f Ecnmic Perspectives 2: Bafumi, Jseph. 23. "The Regress f delgical Sphisticates." Presented at the 23 Annual Meeting f the American Plitical Science Assciatin, Philadelphia, PA. Bafumi, Jseph. 24. "The Stubbrn American Vter." Presented at the 24 Annual Meeting f the American Plitical Science Assciatin, Chicag,. Baldassarri, elia and Andrew Gelman. 27. "Partisans Withut Cnstraint: Plitical Plarizatin and Trends in American Public Opinin." Unpublished Manuscript. Bartels, arry. 2. "Partisanship and Vting Behavir, " American Jurnal f Plitical Science 44: Bartels, arry. 26. "What's the Matter with What's the Matter with Kansas? " Quarterly Jurnal f Plitical Science 1: Berelsn, Bernard R., Paul F. azarsfeld and William N. McPhee Vting: A Study f Opinin Frmatin in a Presidential Campaign. Chicag: University f Chicag Press. Blch-Elkn, Yaeli and Rbert Y. Shapir. 25. "eep Suspicin: raq, Misperceptin and Partisanship." Public Opinin Prs, Campbell, Angus, Philip E. Cnverse, Warren E. Miller and nald E. Stkes The American Vter. New Yrk: Wiley. Carmines, Edward G. and James A. Stimsn ssue Evlutin: Race and the Transfrmatin f American Plitics. Princetn, NJ: Princetn University Press. Carmines, Edward G., Jhn Mcver and James A. Stimsn "Unrealized Partisanship: A Thery f ealignment." Jurnal f Plitics 49: Carsey, Thmas M. and Geffrey C. ayman. 26. "Changing Sides r Changing Minds? Party dentificatin and Plicy Preferences in the American Electrate." American Jurnal f Plitical Science 5: Cnverse, Philip E The Nature f Belief Systems in Mass Publics. n delgy and iscntent, ed. avid E. Apter. New Yrk: Free Press, pp Cnverse, Philip E. and Gregry B. Markus "Plus ga Change.The NewCPS Electin Study Panel." American Plitical Science Review 73: Eriksn, Rbert S. 24. "Ecnmic Vting: Micr vs. Macr Perspectives." Presented at the Annual Meeting f the Plitical Methdlgy, Stanfrd University. Eriksn, Rbert S., Gerald C. Wright and Jhn P. Mcver. 26. Public Opinin in the States: A Quarter Century f Change and Stability. n Public Opinin in State Plitics, ed. Jeffrey E. Chen. Stanfrd, CA: Stanfrd University Press, pp Eriksn, Rbert S., Michael B. MacKuen and James A. Stimsn. 22. The Macr Plity. New Yrk: Cambridge University Press. Firina, Mrris P Retrspective Vting in American Natinal Electins. New Haven: Yale University Press. Firina, Mrris P. 23. ivided Gvernment. New Yrk: Pearsn ngman. Firina, Mrris P. and Samuel J. Abrams. 28. "Plitical Plarizatin in the American Public." Annual Review f Plitical Science 11: Firina, Mrris P., Samuel J. Abrams and Jeremy C. Ppe. 26. Culture War? The Myth f a Plarized America. 2nd ed. New Yrk: Pearsn ngman. Franklin, Charles H. and Jhn E. Jacksn "The ynamics f Party dentificatin." American Plitical Science Review 77: Gelman, Andrew, Bris Shr, Jseph Bafumi and avid K. Park. 27. "Rich State, Pr State, Red State, Blue State: What's the Matter with Cnnecticut?" Quarterly Jurnal f Plitical Science 2: Green, nald Philip, Bradley Palmquist and Eric Schickler. 22. Partisan Hearts and Minds: Plitical Parties and the Scial dentities f Vters. New Haven: Yale University Press. Hetheringtn, Marc J. 21. "Resurgent Mass Partisanship: The Rle f Elite Plarizatin." American Plitical Science Review 95: Jacbs, awrence R. and Rbert Y. Shapir. 2. Pliticians n't Pander: Plitical Manipulatin and the ss f emcratic Respnsiveness. Chicag: University f Chicag Press. Jacbsn, Gary C. 27a. A ivider, Nt a Uniter: Gerge W. Bush and the American Peple. New Yrk: Pearsn ngman. Jacbsn, Gary C. 27b. "Referendum: The 26 Midterm Cngressinal Electins." Plitical Science Quarterly 122: Jewell, Malclm E. and Sarah M. Mrehuse. 21. Plitical Parties and Electins in American States. Washingtn,.C.: Cngressinal Quarterly. Key, V. O. Jr "A Thery f Critical Electins." Jurnal f Plitics 17: Knight, Kathleen and Rbert S. Eriksn delgy in the 199s. n Understanding Public Opinin, ed. Barbara Nrrander and Clyde Wilcx. Washingtn,.C.: Cngressinal Quarterly Press, pp anger, Gary and Jn Chen. 25. "Vters and Values in the 24 Electin." Public Opinin Quarterly 69: ayman, Geffrey. 21. The Great ivide: Religius and Cultural Cnflict in American Party Plitics. New Yrk: Clumbia University Press. azarfeld, Paul F., Bernard Berelsn and Hazel Gaudet The Peple's Chice: Hw the Vter Makes Up His Mind in a Presidential Campaign. New Yrk: Clumbia University Press. rd, C. G.,. Rss and M. R. epper "Biased assimilatin and attitude plarizatin: The effects f prir theries n subsequently cnsidered evidence." Jurnal f Persnality and Scial Psychlgy 37: uskin, Rbert C., Jhn P. Mcver and Edward G. Carmines "ssues and the Transmissin f Partisanship." American Jurnal f Plitical Science 33: MacKuen, Michael B., Rbert S. Eriksn and James A. Stimsn "Macrpartisanship." American Plitical Science Review 83: Marcus, Gerge E "emcratic Theries and the Study f Public Opinin." Plity 21: Mayer, William G The Changing American Mind: Hw and Why Public Opinin Changed between 196 and Ann Arbr: University f Michigan Press. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins

25 24 JOSEPH BAFUM AN ROBERT Y. SHAPRO McCarty, Nlan, Keith Ple and Hward Rsenthal. 26. Plicy. n Pwer and Superpwer: Glbal eadership and Plarized America: The ance f delgy and Unequal Riches. Exceptinalism in the 21st Century, ed. Peter Rundlet, Mrtn Cambridge, MA: MT Press. H. Halperin, Jeffrey aurenti and Spencer P. Byer. New Miller, Warren E. and Merrill Shanks The New American Yrk: Century Fundatin Press, pp Vter. Cambridge, MA: Harvard University Press. Smith, Eric R. A. N The Unchanging American Vter. Mutz, iana C. 26. Hw the Mass Media ivide Us. n Berkeley: University f Califrnia Press. Characteristics and Causes f America's Plarized Plitics, ed. Snyder, Jack, Rbert Y. Shapir and Yaeli Blch-Elkn. 27. Pietr S. Nivla and avid W. Brady. Washingtn,.C.: Brkings nstitutin Press, pp "Free Hand Abrad, ivide and Rule at Hme: The mestic Plitics f Uniplarity." Presented at the 27 Annual Meet- Nie, Nrman H., Sidney Verba and Jhn R. Petrcik The ing f the American Plitical Science Assciatin, Chicag,. Changing American Vter: Enlarged Editin. Cambridge: Harvard University Press. Stimsn, James A Public Opinin in America: Mds, Cycles, and Niemi, Richard G. and Swings. Bulder, CO: Westview Press. M. Kent Jennings "ssues and nheritance in the Frmatin f Party dentificatin." Amer- Stimsn, James A Public Opinin in America: Mds, ican Jurnal f Plitical Science 35: Cycles, and Swings. 2nd ed. Bulder, CO: Westview Press. Nivla, Pietr S. and avid W. Brady. 26. Red And Blue Stimsn, James A. 24. Tides f Cnsent: Hw Opinin Mve- Natin?: Characteristics And Causes f America's Plarized ments Shape American Plitics. New Yrk and ndn: Cam- Plitics. Washingtn,.C.: Brkings nstitutin Press. bridge University Press. Nivla, Pietr S. and avid W. Brady. 27. Red and Blue Sundquist, James ynamics f the Party System: Align- Natin?: Cnsequences and Crrectin f America's Plarized ment and Realignment f Plitical Parties in the United States. Plitics. Washingtn,.C.: Brkings nstitutin Press. Washingtn,.C.: The Brkings nstitutin. Page, Benjamin Chices and Eches in Presidential Taber, Charles and Miltn dge. 26. "Mtivated Skepticism Electins: Ratinal Man and Electral emcracy. Chicag: in Plitical nfrmatin Prcessing." American Jurnal f University f Chicag Press. Plitical Science 5: Page, Benjamin. and Rbert Y. Shapir The Ratinal Uslaner, Eric M. 2. s the Senate Mre Civil than the Huse. Public: Fifty s f Trends in Americans' Plicy Preferences. n Esteemed Clleagues, ed. Burdette A. mis. Washingtn, Chicag: University f Chicag Press..C.: The Brkings nstitutin, pp Parent, Jseph M. and Jseph Bafumi. 28. "nternatinal Veghte, Benjamin, Plarity and mestic Greg M. Shaw and Rbert Y. Plarizatin: Explaining American Shapir. 27. Scial Plicy Preferences, Natinal efense and Plitical isunity." Wrking Paper. Plarizatin in the United States. n Scial Justice, egitimacy Pmper, Gerald "Frm Cnfusin t Clarity: ssues and and the Welfare State, ed. S. Mau and B. Veghte. Aldersht: American Vters, " American Plitical Science Review 66: Ashgate Publishing td., pp Pmper, Gerald Vters' Chice: Varieties f American Wattenberg, Martin P The ecline f American Plitical Parties. Electral Behavir. New Yrk: dd, Mead and Cambridge: Harvard University Press. Cmpany. Wawr, Prir, Markus. 27. Pst-Bradcast emcracy: Hw Media Greg. 26. "The Ratinalizing Public? " Critical Review 18: Chice ncrease nequality in Plitical nvlvement and Plarize Electins. New Yrk: Cambridge University Press. Wlbrecht, Christina. 2. The Plitics f Wmen's Rights: Parties, Rhde, avid W Parties and eaders in the Psitins, and Pstrefrm Change. Princetn: Princetn University Press. Huse. Chicag: University f Chicag Press. Shapir, Rbert Y. and Yaeli Blch-Elkn. 26. "Plitical Plarizatin and the Ratinal Public." Presented at the Cnference f the American Assciatin fr Public Jseph Bafumis an assistant prfessr f gvern- Opinin Research, Mntreal, Quebec, Canada. ment, artmuth Cllege, Hanver, NH Rbert Shapir, Rbert Y. and Yaeli Blch-Elkn. 27. Y. delgical Shapir is a prfessr f plitical science, Clumbia Partisanship and American Public Opinin tward Freign University, New Yrk, NY 127. This cntent dwnladed frm n Mn, 25 Mar :1:53 PM All use subject t JSTOR Terms and Cnditins