Challenges to the Almond-Lippmann Consensus Mershon Series: Research Programs and Debates Author(S): Ole R

Public Opinion and Foreign Policy: Challenges to the Almond-Lippmann Consensus Mershon Series: Research Programs and Debates Author(s): Ole R. Holsti Reviewed work(s): Source: International Studies Quarterly, Vol. 36, No. 4 (Dec., 1992), pp. 439-466 Published by: Wiley-Blackwell on behalf of The International Studies Association Stable URL: http://www.jstor.org/stable/2600734 . Accessed: 08/11/2012 03:12

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. 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].

Wiley-Blackwell and The International Studies Association are collaborating with JSTOR to digitize, preserve and extend access to International Studies Quarterly.

http://www.jstor.org InternationalStudies Quarterly (1992) 36, 439-466

Public Opiniionand ForeignPolicy: Challenges to theAlmond-Lippmann Consensus MershonSeries: Research Programs and Debates

OLE R. HOLSTI Duke University

This articlesurveys and assesses theoriesand researchon public opinion and foreignpolicy. Most of the evidence is drawnfrom the literatureon the United States. Three twentieth-centurywars have had a significant impact on theoryand scholarship.World War I-the firstpublic relations war-and postwar effortsto create a new internationalorder directedmuch attentionto the nature of public opinion and its impact on foreignaffairs, issues on which realistsand liberals came to quite differentconclusions. The period surroundingWorld War II coincided withthe developmentof scientificpolling. Much of the attentionduring and immediatelyafter the war focused on the extentto which the public might support or oppose an internationalistAmerican role. Extensive researchduring the firsttwo decades afterWorld War II yielded a broad agreement (the "Almond-Lippmannconsensus") on three propositions about public opinion: (1) it is volatile and thus provides inadequate foundationsfor stable and effectiveforeign policies, (2) it lacks coherence or structure,but (3) in the final analysis,it has littleif any impact on foreignpolicy. The VietnamWar and its aftermathstimulated a new outburstof researchactivity on public opinion and foreignpolicy, much of which has challenged each of these three propositions.The article concludes with suggestionsfor furtherresearch efforts,including: (1) case studiesemploying archival sources to assessmore directlythe impact of public opinion, (2) cross-nationalstudies, (3) developmentof standard questionsin order to encourage bettercumulation of surveyresults, and (4) researchthat will enable us to distinguishfindings that are time- and context-boundfrom those thattranscend the Cold War period.

Introduction Many questions about the role of public opinion in foreign policy are at the center of persisting debates between the liberal-democratic and realist approaches to foreign affairs.Is public opinion a force for enlightenment-indeed, a necessary if not sufficientcondition for sound foreign policy-as celebrated by the Wilsonians

Author'snote: For helpfulcomments and suggestionson an earlier draftof this paper I am gratefulto Stephen Earl Bennett,William Chittick, Thomas Graham,Jon Hurwitz,Benjamin Page, Mark Peffley,Philip Powlick,Bruce Russett,Eugene Wittkopf,and twoanonymous reviewers for International Studies Quarterly.

?) 1992 InternationalStudies Association 440 Public Opinionand ForeignPolicy and otherliberals? There is a long,liberal-democratic tradition, dating back at least to Kant and Bentham,that foreign policies of democraciesare more peaceful,at least in part because the public can play a constructiverole in constrainingpolicy makers;only accountability to the public can restrainthe war-makingproclivities of leaders.1 Alternatively,are Hans Morgenthauand othersof the realistschool correctin describingpublic opinion as a barrierto thoughtfuland coherentdiplomacy, hin- deringefforts to promotenational interests that transcend the moods and passions of the moment?The realisttradition is intenselyskeptical of the public's contribu- tion to effectiveforeign policy. At the veryminimum, most realistswould distin- guish between foreignpolicy and other public policyissues; the public mightbe sufficientlyinformed to deal withlocal issues thatimpinge on theirdaily lives, but foreignaffairs are too remote fromtheir experience, and in any case theyhave littleinclination to become more informedabout such complex and remoteissues. Finally,the effectiveconduct of diplomacyrequires secrecy,flexibility, and other qualities thatwould be seriouslyjeopardized were the public to have a significant impacton foreignpolicy. Thus, to permitthe public a strongvoice in policywould be to place the democracies,if not the stabilityof the internationalsystem itself, at a distinctdisadvantage. Moreover, it would permitthe emotional to govern the rational.Hans Morgenthausummarized the case againstan activerole for public opinion in words that would gain the support of most if not all realists:"The rationalrequirements of good foreignpolicy cannot fromthe outsetcount upon the support of a public opinion whose preferencesare emotional rather than rational"(Morgenthau, 1978:558). The long-standingdebate betweenliberals and realistswas intensifiedby World War I, whichmight be describedas the firstpublic relationswar. From its inception both the Allied and Central Powers tried to win over "worldopinion" in various ways,including publication by many governmentsof highlyselective document collections-the so-calledcolor books-all of whichwere intendedto absolve them and to place the blame for the war on their adversaries.The propaganda war duringthe conflictwas almostas intenseas thaton the battlefield. PresidentWilson's hopes fora new postwarworld order depended significantly on democratizingforeign affairsand diplomacy. Elihu Root, a distinguished Republican and formerSecretary of State,effectively summarized the positionof those who welcomed an increasingrole for the public in the conduct of foreign affairs. When foreignoffices were ruled by autocracies or oligarchiesthe danger of war was in sinisterpurpose. When foreignaffairs are ruled by democraciesthe danger of war will be in mistakenbeliefs. The world will be gainer by the change, for, while thereis no human wayto preventa kingfrom having a bad heart,there is a human way to prevent a people from having an erroneous opinion. (Root, 1922:5) By more effectiveinternational education, "the people themselveswill have the means to testmisinformation and appeals to prejudiceand passion based on error" (Root, 1922:5). But not all observersjoined Wilson and Root in applauding the prospect of popular diplomacy. During the postwar era, the journalist Walter Lippmann published two trenchantcritiques of the central premisesof classical liberalism (Lippmann, 1922, 1925). Accordingto Lippmann, the common man is too fully engaged in the requirementsof earning a livingand otherwiseattending to his

1 Recent research has found that the foreignpolicies of democracies are indeed differentfrom those of other politiesand, further,that democracies do n-otengage in war againsteach other. OLE R. HOLSTI 441 most immediateneeds to have the time or inclinationto satisfythe heroic, but clearlyunrealistic, assumptions about the informedand engaged citizencelebrated in classicaldemocratic theory. The chasm between theoryand realityis especially wide on foreignaffairs, which are typicallyfar removed from the directexperiences of the mass public. Consequently,the "picturesin the head" of the average citizen are unlikelyto have much correspondencewith the real world of international affairs.Moreover, journalist Lippmann became increasinglyskeptical about the abilityof his own professioneffectively to fillin the gap betweenthe real worldand the averagecitizen's stereotypes. His studyof the Russianrevolution, as depicted in the pages of the NewYork Times during the period 1917-20, did nothingto assuage his pessimism about the ability of the media to serve as a source of valid informationabout the world for the public (Lippmann and Metz, 1920). The events of the 1930s and the outbreak of World War II, which seemed to raise serious questions about Wilsonian assumptions while apparently providing compelling empirical confirmationfor the realist approach to international relations,further tipped the balance of the debate on public opinion and foreign policyin favorof the skeptics. The period encompassingWorld War II and its immediateaftermath coincided withthe inceptionof scientificpublic opinion polling. Despite the wounds to the reputationof polling and pollstersinflicted by the LiteraiyDigest debacle in the Roosevelt-Landonelection, we can date the era of scientificsurveys from the establishmentof the Gallup poll in 1936 or of the Public Opinion Quarterlytwo years later.It maybe worthnoting that President Roosevelt was a pioneer in the use of a public opinion consultant-Hadley Cantril,one of the foundingfathers of the new science-for guidance on policy. Policy makersand manyothers who feltthat an irresponsibleAmerican isola- tionismafter 1919 had contributedto the outbreakof World War II worriedthat the public mood might trace out a pattern resemblingthe experience of the period afterWorld War I: wartimeidealism and internationalism,followed soon thereafterby cynicismand disenchantmentwith active American leadership in effortsto create a more stable internationalorder.2 This concern was reflectedin the frequencywith which polling organizations during World War II asked respondents general questionsabout the United Statestaking an activerole in, or staying out of, world affairs,and specific queries about support for or opposition to Americanmembership in the UnitedNations. These two features-an empirical approach that relied heavilyon systematic polling data and a normativeconcern that an isolationistpublic mightlead the United Statesto repeat the failedisolationist policies of the interwarperiod-may be found in two of the pioneeringworks on public opinion and foreignpolicy: Thomas Bailey's The Man in theStreet (1948) and Gabriel Almond's TheAmerican Peopleand Foreign Policy (1950).

The Post-WorldWar II Consensus The availabilityafter World War II of growing sets of polling data and the institutionof systematicstudies of votingbehavior, combined withthe assumption

2Justbefore President Roosevelt left for the Yalta Conferencein 1945, Hadley Cantrilgave the presidenta memo which stated: "Although the overwhelmingmajority of the American people now favor a strong international organizationnecessarily dominated by the big powers,it is unrealisticto assume thatAmericans are international- minded. Their policy is ratherone of expediency,which, at the moment,takes the formof internationalism.The presentinternationalism rests on a ratherunstable foundation: it is recent,it is not rooted in anybroad or long-range conceptionof self-interest,it has littleintellectual basis" (Cantril,1967:76). 442 Public Opinionand ForeignPolicy of a leadership role in world affairsby the United States,served to stimulatea growthindustry in analysesof public opinion. The consensusview that developed duringthis period of some fifteenor twentyyears after the end of World War II and just priorto theVietnam escalation centered on threemajor propositions:

* Public opinion is highlyvolatile and thus it providesvery dubious foundations for a soundforeign policy. * Public attitudeson foreignaffairs are so lacking in structureand coherencethat they mightbest be describedas "non-attitudes." * At the end of the day, however,public opinion has a verylimited impact on the conductofforeign policy.

Let us examine each of these propositions,and the evidence upon which they rested,in more detail.

PublicOpinion Is Volatile As noted earlier,Walter Lippmann's books of the interwarperiod described the mass public as neithersufficiently interested nor informedto play the pivotalrole assigned to it by classicaldemocratic theory. At the heightof the Cold War thirty yearslater, Lippmann had become even more alarmed,depicting the mass public as not merelyuninterested and uninformed,but as a powerfulforce that was so out of synchwith reality as to constitutea massiveand potentiallyfatal threat to effec- tivegovernment and policies. The unhappy truthis that the prevailingpublic opinion has been destructively wrongat the criticaljunctures. The people have impresseda criticalveto upon the judgments of informed and responsible officials.They have compelled the government,which usually knew what would have been wiser,or was necessary,or whatwas more expedient,to be too late withtoo little,or too long withtoo much, too pacifistin peace and too bellicose in war, too neutralistor appeasing in negotiationsor too intransigent.Mass opinion has acquired mountingpower in this country.It has shown itselfto be a dangerous masterof decision when the stakesare lifeand death. (Lippmann, 1955:20) Similarly pessimistic conclusions and dire warnings were emerging from disparateother quartersas well. Drawingon a growingbody of polling data and fearing that the American public might relapse into a mindless isolationism, because onlya thinveneer of postwarinternationalism covered a thickbedrock of indifferenceto the world,Gabriel Almond depicted public opinion as a volatile and mood-drivenconstraint upon foreignpolicy: "The undertowof withdrawalis stillvery powerful. Deeply ingrained habits do not die easy deaths. The world outside is stillvery remote for most Americans;and the tragiclessons of the last decades have not been fullydigested" (Almond, 1950:85). Consequently,"Perhaps the gravestgeneral problem confrontingpolicy-makers is thatof the instabilityof mass moods, the cyclicalfluctuations which stand in the way of policy stability" (Almond,1950:239).3 Six yearslater, Almond restatedhis thesisin Lippmannesque language, citingnot only the instabilityof public moods, but other deficienciesas well. He told an audience at the National War College, "For persons responsible for the makingof securitypolicy these mood impactsof the public have a highly

3 Almond's use of the term "mood" differsfrom that of Frank Klingberg.Almond refersto sudden shiftsof interestand preferences,whereas Klingberghas used the term to explain American foreign policy in termsof generation-longsocietal swingsbetween introversionand extraversion.For the latterusage, see Klingberg(1952, 1979, 1983) and Holmes (1985). OLE R. HOLSTI 443 irrationaleffect. Often the public is apatheticwhen it should be concerned,and panickywhen it should be calm" (Almond,1956:372, 376). Otherswhose writingsprovided support for the main outlinesof the pessimistic Almond-Lippmannthesis included a distinguishedlist: the dean of American diplomatic historians(Thomas A. Bailey); the foremostproponent of a realist approach to internationalaffairs (Hans J. Morgenthau); and the diplomat-histo- rian who has oftenbeen depicted as the intellectualfather of the Americanpolicy of containment,George F. Kennan. For the latter,the metaphor of a dinosaur vividlydepicted the problemsof democraticforeign policy. But I sometimeswonder whether in thisrespect a democracyis not uncomfortably similarto one of those prehistoricmonsters with a body as long as thisroom and a brain the size of a pin: he lies there in his comfortableprimeval mud and pays littleattention to his environment;he is slow to wrath-in fact,you practically have to whackhis tail offto make him aware thathis interestsare being disturbed; but, once he graspsthis, he laysabout him withsuch blind determinationthat he not only destroyshis adversarybut largelywrecks his native habitat. (Kennan, 1951:59) Furthersupport for the criticsand skepticsemerged fromthe growingbody of polling data whichyielded ample evidence of the public's limitedstore of factual knowledgeabout foreignaffairs. Innumerable surveys revealed such stunninggaps in informationas: X percent of the Americanpublic are unaware that there is a communistgovernment in China, Y percent believe that the Soviet Union is a memberof NATO, or Z percentcannot identifya single nation borderingon the PacificOcean. Such data reinforcedthe case of the criticsand led some of themto propose measuresto reduce the influenceof the public. Thus, Lippmann (1955) called for strongerexecutive prerogatives in foreignaffairs, and Bailey (1948:13) wonderedwhether the requirementsof an effectiveforeign policy might make it necessaryfor the executivedeliberately to misleadthe public.

PublicOpinion Lacks Structure and Coherence A growingvolume of data on public opinion and voting behavior, as well as increasinglysophisticated methodologies, enabled analystsnot only to describe aggregateresults and trends,but also to delve into the structureof politicalbeliefs. Owing to immediatepolicy concerns about the U.S. role in the postwarera, many of the earlystudies were largely descriptive, focusing on such issuesas participation in internationalorganizations and alliances, the deploymentof troops abroad, securitycommitments, foreign aid, trade and protectionism,and the like. The underlyingpremise was that a single internationalist-isolationistdimension would serveto structureforeign policy beliefs, much in the waythat a liberal-conservative dimensionwas assumed to providecoherence to preferenceson domesticissues. In a classic studybased on data from the late 1950s and early 1960s, Philip Converse (1964) concluded thatthe politicalbeliefs of the mass public lack a real structureor coherence. Comparingresponses across severaldomestic and foreign policyissues, he found littleif any "constraint"or underlyingideological structure thatmight provide some coherence to politicalthinking. In contrast,his analyses of elites-congressional candidates-revealed substantiallyhigher correlations among responsesto variousissues. Moreover,Converse found thatboth mass and elite attitudeson a given issue had a short half-life.Responses in 1956 only modestlypredicted responses two years later, much less in 1960. These findingsled him to conclude that mass political beliefsare best described as "non-attitudes." AlthoughConverse's findings were laterto become the centerof an activedebate, it should be emphasized thathis was not a lone voice in the wilderness.His data were drawnfrom the National Election Studies at the Universityof Michigan,and his findingswere only the mostwidely quoted of a seriesof studiesfrom the NES 444 Public Opinionand ForeignPolicy that came to essentiallythe same conclusion about the absence of structure, coherence,or persistencein the politicalbeliefs of the mass public-especially on foreignaffairs (Campbell, Converse, Miller, and Stokes,1964).

PublicOpinion Has LimitedImpact on Foreign Policy A significantreason forinterest in public opinion on foreignaffairs arises from the assumptionthat in some waysand at least some of the time it has an impact,for betteror worse,on the conduct of the nation's externalpolicy. Certainly it is easy to find public expressionsby policy makers avowing the importanceof public opinion. Duringhis debates withStephen Douglas, AbrahamLincoln assertedthat, "withpublic sentimenton its side, everythingsucceeds; with public sentiment against it, nothingsucceeds"; and in 1936 Secretaryof State Cordell Hull stated that,"since the timewhen ThomasJeffersoninsisted upon a 'decent respectto the opinions of mankind,'public opinion has controlledforeign policy in all democracies"(New YorkTimes, Dec. 6, 1936). Althoughsuch hyperbolicstatements are unlikelyto withstandserious empirical scrutiny,the drivingforce behind much of the post-WorldWar II attentionto public opinion on foreignpolicy issues was the fear that an ill-informedand emotional mass public would serve as a powerfulconstraint on the conduct of American diplomacy, establishing unwise limits on policy makers, creating unrealisticexpectations about whatwas feasiblein foreignaffairs, otherwise doing seriousmischief to Americandiplomacy and, giventhe American role in the world, perhaps even to internationalstability. As BernardCohen (1973) demonstratedin a criticalsurvey of the literature,however, the constrainingrole of public opinion was oftenasserted but rarelydemonstrated-or even put to a systematictest. By the middle of the 1960s a consensus in fact seemed to emerge on a third point: Public opinion has little if any real impact on policy. Or, as the point was made most pithilyby one State Department official:"To hell with public opinion.... We should lead, and not follow" (quoted in Cohen, 1973:62). The weight of research evidence cast doubt on the potency of public opinion as a drivingforce behind, or even a significantconstraint upon, foreignpolicy-making. For example, a classic studyof the public-legislatorrelationship revealed that constituents'attitudes on foreignpolicy had less impacton membersof the House of Representativesthan did their views on domestic issues (Miller and Stokes, 1963). Cohen's research on the foreignpolicy bureaucracyindicated that State Departmentofficials had a rathermodest interestin public opinion, and to the extent that they even thought about the public, it was as an entityto be "educated" rather than a lodestar by which to be guided (Cohen, 1973). The propositionthat the presidenthas "almosta freehand" in the conduct of foreign affairsreceived support from other analysts,including Lipset (1966), LaFeber (1977), Levering(1978), Paterson (1979), and Graebner (1983). This period also witnesseda proliferationof case studiesof keyforeign policy decisions; with rare exceptions, however, they make no reference to public opinion. But it is not alwaysclear whetherthat is because: (1) public opinion was irrelevantas an explanation to the decisions under consideration,(2) decision- makersquietly anticipate public opinion withoutconsciously doing so, (3) it was excluded from the research design and thus no effortwas made to assess its impact,or (4) disproportionateresearch attention to internationalcrises-events thatare usuallycharacterized by shortdecision time-tended to exclude episodes in whichdecisions are the culminationof a long politicalprocess; all other things being equal, the more protractedthe decision process,the more likelyare policy makersto be subjected to the impact of public opinion throughthe activitiesof Congress,pressure groups, the media, and opinion leaders. OLE R. HOLSTI 445

Although these studies did not answer all questions about the impact of the public-for example, the realitiesof researchaccess required Cohen to focus on preciselythose personswho are most shelteredfrom the effectsof elections-the weightof the evidence would at least have assuaged those who shared fearsthat mass public opinion "has shown itselfto be a dangerous masterof decision when the stakesare lifeand death" (Lippmann, 1955:20).

The Renaissance of Interestin Public Opinion and ForeignPolicy Justas World War II and fearsof postwarisolationism among the mass public gave rise to concern about public opinion and its impacton foreignpolicy, the war in Vietnamwas the impetusfor a renewedinterest in the subject.It was a majorcatalyst in stimulatinga reexaminationof the consensusthat had emergedduring the two decades afterWorld War II. The VietnamWar had at least twodirect effects. Most broadly,many of thosewho had believedthat a strongerexecutive hand on the tiller of public policy,relatively free from the whimsand vagariesof public moods, best servesboth national interestsand global stability,came to reexaminetheir views in the lightof the VietnamWar. Indeed, the widelyread columnistWalter Lippmann, who only a littlemore than a decade earlier had despaired of the tyrannyof a fecklesspublic and had called for a strongerexecutive to counteractthe mass public, became a leading criticof the Johnson administration'sVietnam policy; eventuallyhe came to regardthe public,which had become increasinglyskeptical of thewar effort, as more enlightenedthan the administration. At a narrowerlevel, some criticsof the war became increasinglypersuaded that the Gallup, Harris,and other commercialpolls distortedpublic attitudestoward the war by posing excessivelyrestrictive and simplisticquestions. For example, among the most widelyasked questions was whetherrespondents supported or opposed currentAmerican policy in Vietnam; deeper probes that might have offeredrespondents an opportunityto express their views about other policy optionswere farless commonlyemployed by these polling organizations.Thus, in addition to secondaryanalyses of surveydata relatingto the war (e.g., Mueller, 1973), the conflictin SoutheastAsia also stimulatedindependent surveys designed specificallyto assess foreignpolicy in greaterdepth and breadth than the typical surveyconducted by Gallup and the othermajor polling organizations. The firstof these studies, the Verba-Stanfordsurveys, focused on American policy in Vietnam (Verba, Brody,Parker, Nie, Polsby,Ekman, and Black, 1967; Verba and Brody,1970). Verba and his colleagues in factfound supportfor the administration'sVietnam policy, but they also unearthed approval for such alternativepolicies as negotiatingan end to the conflict.The period followingthe Verba-Stanfordpolls has witnesseda proliferationof studieswith a foreignaffairs focus,including surveys of both the mass public (Rielly,1975, 1979, 1983, 1987, 1991; Hurwitz and Peffley,1987, 1990; Americans Talk Security,1987-1991; Americans Talk Issues, 1991) and of opinion leaders (Barton, 1974-75, 1980; Rielly, 1975, 1979, 1983, 1987, 1991; Russett and Hanson, 1975; Holsti and Rosenau, 1984, 1988, 1990; Chittickand Billingsley,1989; Koopman, Snyder,and Jervis,1990a, 1990b, 1991). As a consequence, we are no longer totallydependent on evidence generated by the major polling organizations. Moreover, these independent surveysare oftendesigned withpolicy and/or theoreticalconcerns thatcan onlyimperfectly be probed by secondaryanalyses of the Gallup and other, more generalpublic opinion polls. Thus, armed with growingcentral archives of data generated by the major polling organizationsas well as evidence produced by the independent surveys, duringthe past twodecades analystshave begun to challenge importantaspects of the consensusdescribed above. 446 Public Opinionand ForeignPolicy

Challenge#1: Is PublicOpinion Really So Volatile? WilliamCaspary presented the firstsystematic challenge to the Almond thesisthat public opinion on foreignaffairs is best characterizedby volatilemoods.4 He took issue withAlmond's heavyreliance on a singlequestion in whichrespondents were asked to identify"the most importantissue today."Caspary's analysis of a broader set of questions led him to conclude that "Americanpublic opinion is characterized by a strongand stable permissivemood" towardactive international involvement (Caspary,1970:546). Mueller's (1973) studyof public opinion towardthe Korean and Vietnamwars posed another challenge to the thesisof mindlesschanges in public attitudes.To be sure,public supportfor the U.S. war effortin both conflictseventually changed, but in waysthat seemed explicable and rational,rather than randomand mindless. More specifically,he foundthat increasing public oppositionto the conflictstraced out a patternthat fit a curveof risingbattle deaths, suggesting that the public used an understandable,if simple, heuristic to assessAmerican policy.5 The most comprehensive challenge to the Almond-Lippmann thesis has emergedfrom studies conducted by BenjaminPage and RobertShapiro. Their evidence includes all questionsthat have been posed by major polling organizations since the inception of systematicsurveys in the 1930s. Of the more than 6000 questions,almost 20 percent have been asked at least twice,providing Page and Shapiro witha large data set to assess the degree of stabilityand change in mass public attitudes.Employing a cutoffpoint of a differenceof 6 percentfrom one surveyto another to distinguishbetween continuityand change, theyfound that mass opinion in the aggregateis in factcharacterized by a good deal of stability, and that this is no less true of foreignpolicy than on domesticissues (Page and Shapiro, 1988). More important,when attitudeshifts take place, theyseem to be neither random nor 180 degrees removed from the true state of world affairs. Rather, changes appear to be "reasonable, event driven" reactions to the real world,even ifthe informationupon whichthey are based is marginallyadequate at best.They concluded that, virtuallyall the rapid shifts[in public opinion] we found were related to political and economic circumstancesor to significantevents which sensible citizens would take into account. In particular,most abrupt foreign policy changes took place in connection with wars, confrontations,or crises in which major changes in the actions of the United States or other nations quite naturallyaffect preferences about whatpolicies to pursue.6 Because theiranalyses are based on aggregateresponses rather than on panel studies in which the same respondents are interviewedrepeatedly, Page and Shapiro cannot address definitivelyone aspect of the volatilityquestion: Precisely whatproportion of individualshave in factchanged theirminds? For an issue on which the public divided 50 percent-50percent in each of two time periods,it is theoreticallypossible thatall respondentsdid so. However,volatility approaching

4 By 1960, Almond himselfwas backing away from his most pessimisticdiagnoses. See his new preface to a reprintingof TheAmerican People and Foreign1Policy. 5 "During the summer of 1965, as the Johnson Administrationwas moving towardfateful decisions regarding Vietnam,George Ball warned:'We can't win,' he said, his deep voice dominatingthe Cabinet Room. 'The warwill be long and protracted,with heavy casualties. The most we can hope for is a messyconclusion. We must measure this long-termprice againstthe short-termloss thatwill result from withdrawal.' Producing a chartthat correlated public opinion withAmerican casualties in Korea, Ball predictedthat the American public would not supporta long and inconclusivewar" (Clifford,1991:412). 6 Page and Shapiro (1982:34); see also Page and Shapiro (1983, 1984, 1988, 1992); Page, Shapiro, anidDempsey (1987); Shapiro and Page (1988). OLE R. HOLSTI 447 this magnitudeseems highlyunlikely because, as Page and Shapiro have shown, opinion changestend to be in directionsthat "make sense" in termsof events. The volatilitythesis can be tested most directlyby individual-levelrather than aggregate analysisof opinion data. Using differentmethods for correctingfor measurementerror, several studies have shownconvincingly that at the individual level mass foreignpolicy attitudesare everybit as stable as domestic attitudes (Achen, 1975; Peffleyand Hurwitz,1992a). These studiesrevealed an impressive level of stabilityduring times of constancyin the internationalenvironment. A panel studyby Peffleyand Hurwitz(1992a) also found verysubstantial stability in policyattitudes and internationalimages even during a period-the late 1980s- thatwitnessed rapid and dramaticchanges in Soviet-Americanrelations and other aspectsof internationalaffairs. Similar conclusions,supporting Page and Shapiro and casting doubt on the Almond-Lippmannthesis, have also emerged fromother studies.Jentleson (1992) found thatduring the post-Vietnamera, variationsin public supportfor the use of force are best explained by differencesbetween force to coerce foreignpolicy restraintby others,and force to influenceor impose internalpolitical changes withinanother state; the formergoal has receivedmuch strongersupport than the latter.7 An interestingvariant of the "rationalpublic" thesis stipulatesthat the public attemptsto moderateAmerican behavior towardthe USSR by expressingprefer- ences for a conciliatorystance from hawkishadministrations while supporting more assertivepolicies from dovish ones (Nincic, 1988). To the extentthat one can generalize fromthis studyfocusing on the Carterand Reagan administrationsto otherperiods or otheraspects of foreignpolicy, it furtherchallenges the Almond- Lippmann thesis-indeed, it turnsthat propositionon its head-for it identifies the public as a source of moderationand continuityrather than of instabilityand unpredictability. It is important to emphasize that none of these challenges to the Almond-Lippmannthesis is based on some newlyfound evidence that the public is in factwell informedabout foreignaffairs. Not onlydo polls repeatedlyreveal that the mass public has a very thin veneer of factual knowledge about politics, economics, and geography;they also reveal that it is poorly informedabout the specificsof conflicts,treaties, negotiations with other nations, characteristicsof weapons systems,foreign leaders, and the like. Because the modest factualbasis upon which the mass public reacts to international affairs remains an unchallenged-and unchallengeable-fact, we are faced with a puzzle: If a generallypoorly informed mass public does indeed react to internationalaffairs in an events-driven,rational manner, what are the means that permitit to do so? Recall thata not-insignificantbody of researchevidence indicatedthat mass public attitudeslack the kind of ideological structurethat would providesome coherence acrossspecific issues and persistencethrough time.

Challenge#2: Do PublicAttitudes Lack Structureand Coherence? Philip Converse's (1964) chapteron mass belief systemsis one of the mostwidely cited studiesin the Americanpolitical science literature.In recentyears it has also spawned a vast literaturewhich has, on the one hand, vigorouslychallenged his findingsand, on the other,supported the main thrustof Converse's conclusions

7 For additional evidence about the "rationalpublic," the stabilityof policy preferences,and issue voting,see Bennett (1972:742), Free and Watts(1980:50), Wittkopf(1986, 1990), Graham (1986, 1988, 1989), Krosnick(1988a, 1988b, 1990, 1991), Russett(1990), and Peffleyand Hurwitz(1992a, 1992b). 448 Public Opinionand ForeignPolicy that mass public attitudeslack ideological structure,whereas those of leaders are characterizedby far greater coherence. Part of the debate is methodological, centeringon the manner in which questionsare framed,the clarityof questions, the degree to whichthe unsure are prodded to statea position,and similarissues of research procedures. Did the evidentiarybase include enough questions to support the conclusions? Did the analytical methods deal adequately with problems of measurementerror? Does an analysis that examines correlations across specificpublic policy issues exhaust the possible structuresthat mightbe used to lend coherence to political thinking?Studies that have raised significant methodologicalquestions about the Conversefindings include Achen (1975) and Sullivan,Piereson, and Marcus (1978). Anotherpart of the controversyfocuses on trends,specifically on the durability of findingsthat, to a large degree,drew from evidence generated during the 1950s. This was a period of American economic, political,and militarydominance in foreignaffairs-the shock of Sputnikin 1957 notwithstanding-withthe 1956 and 1960 elections takingplace afterthe Korean War and before escalation of the Vietnamconflict. Domestically, the Eisenhoweryears were markedby relativelylow inflation and unemploymentand, despite the Montgomerybus boycott and Greensborosit-ins, the fullimpact of the civilrights movement had yetto be felt. According to the critics,this period, both celebrated and criticizedfor marking "the end of ideology,"is insufficientlyrepresentative for assessingthe degree of ideological consistencyamong the general public. In support of that view, a number of analystsfound that,beginning with the Johnson-Goldwaterelection campaign of 1964, ideological consistencyamong the public did in factincrease (Nie and Anderson, 1974; Nie, Verba, and Petrocik,1976). Some corroborating evidence also appeared to emerge from Hero's (1969) assessment of public opinion polls on domesticand foreignpolicy issues from the late 1930s to 1967. In general,he found a veryweak relationshipbetween the two,with some indications of a strengtheningduring the post-Eisenhoweryears. Those who claim to have found a greaterideological consistencyamong the general public duringthe turbulentera of the 1960s and 1970s have also encountered criticism.Are the claims of greaterissue consistencyrooted in increasing ideological consciousness?Alternatively, are theymerely the resultof parrotingof ideological rhetoric,or of some methodologicalartifact? This is not the place to providea blow-by-blowaccount of the manyand variedanswers to these and other questionson the issue; forexcellent and detailed summariesof the vastliterature, see Kinder (1983), Kinderand Sears (1985), and Snidermanand Tetlock (1986). It will sufficeto say that there appears to be an emergingconsensus that public responses to issues are not adequately captured by the most familiarbipolar dimensions: liberal-to-conservativeor internationalist-to-isolationist.If these dimensionsconstitute the standard,then mass public attitudesdo indeed appear to lack structure.Given that tentativeconclusion, does the literatureon foreign policy attitudesreveal anythingelse about organizingconcepts that mightlend some coherence to mass public attitudeson internationalaffairs? Althoughthe more recentresearch literature has yetto createa consensuson all aspects of the question, there does appear to be a considerable convergenceof findingson twogeneral points relating to beliefstructures:

1. Even though the general public may be ratherpoorly informed,attitudes about foreignaffairs are in fact structuredin at least moderatelycoherent ways. Indeed, low informationand an ambiguous foreignpolicy environment are actually likely to motivate rather than preclude some type of attitudestructure. 2. A single isolationist-to-internationalistdimension inadequately describes the main dimensionsof public opinion on internationalaffairs. OLE R. HOLSTI 449

An early study,based on the firstof the quadrennial Chicago Council on ForeignRelations (CCFR) surveys,employed factor analysis and other methodsto uncover three foreign policy outlooks: "liberal internationalism,""conservative internationalism,"and "non-internationalism"(Mandelbaum and Schneider, 1979). A comparable trichotomy("three-headed eagle") emerged from early analysesof the data on opinion leaders generatedby the ForeignPolicy Leadership Project (FPLP) (Holsti,1979; Holstiand Rosenau, 1979, 1984). Others have questioned the divisionof foreignpolicy attitudes into three types ratherthan dimensions,and theyhave offeredcompelling evidence in supportof theircritiques. Chittick and Billingsley(1989) have undertakenboth originaland secondaryanalyses which indicated the need for three dimensions,including one that taps unilateralist-multilateralistsentiments, not three types,to describe adequatelythe foreignpolicy beliefs of both the mass public and leaders. (See also Bardes and Oldendick,1978; Chittick,Billingsley, and Travis,1990). A major set of contributionsto the debate about how best to describe foreign policy attitudeshas come fromWittkopf's exemplary secondary analyses of the CCFR surveysof both the general public and leaders (Wittkopf,1986, 1990). His results,developed inductivelyfrom the firstfour CCFR surveys,revealed that with a singleexception, two dimensions are necessaryto describeforeign policy attitudes: "support-opposemilitant internationalism" (MI) and "support-opposecooperative internationalism"(CI). Dichotomizingand crossingthese dimensionsyields four types, with the quadrants labeled as hard-liners(support MI, oppose CI), internationalists(support MI, supportCI), isolationists(oppose MI, oppose CI), and accommodationists(oppose MI, supportCI). Support for Wittkopf'sMI/Cl scheme also emerges from a reanalysisof the FPLP data on Americanopinion leaders (Holsti and Rosenau, 1990). That study put the MI/Cl scheme to a demandingtest because of three major differencesin the data sets: (1) The CCFR surveyswere undertakenin 1974, 1978, 1982, and 1986, whereasthe fourFPLP studiesfollowed two years later in each case; (2) the two sets of surveyshave only a fewquestionnaire items in common; and (3) the MI/Cl scheme was developed largelyfrom data on the mass public, whereas the FPLP surveysfocused solely on opinion leaders. It may be worth noting that although the origins of the MI/Cl scheme are strictlyinductive, the militantinternationalism and cooperativeinternationalism dimensionscorrespond closely to the mostvenerable approaches to international relations: realism and liberalism.Realism views conflictbetween nations as a natural state of affairsrather than an aberrationthat is subject to permanent amelioration.Such realistconcepts as securitydilemma, relative capabilities, and "zero sum" viewof conflictare also basic to the MI dimension.There are similarly intimatelinks between liberalism and the cooperativeinternationalism dimension. Liberalism denies that conflictis an immutable element of relations between nations.It definessecurity in termsthat are broader than the geopolitical-military dimensions, and it emphasizes the cooperative aspects of relations between nations. Institutionbuilding, improved international education and communication,and tradeare but a fewof the waysin whichnations may jointly gain and thus mitigate,if not eliminate,the harshestfeatures of internationalrelations in an anarchic system.In short,the CL dimension shares importantelements with the liberal school of internationalrelations theory. These MI and CL dimensionsalso seem clearlyrelated to other conceptualizationsof Americanthought on foreign affairs.For example,Hughes's (1980:50) distinctionbetween the "securityculture" and "equityculture" in Americanforeign policy, or Billington's(1987:632) cate- gories of "realist-conservatives"and "idealist-liberals"appear to parallel, if not matchexactly, the MI and CL dimensions. Even if one accepts the necessityof tapping attitudeson both militantand cooperativeinternationalism, however, there is also some evidence that theyare 450 Public Opinionand ForeignPolicy not sufficientto describe all contoursof contemporaryforeign policy attitudes. A furtherdistinction between unilateralism and multilateralismhas been suggestedby a numberof studies (Wittkopf,1986; Hinckley,1988, 1992; Chittickand Billingsley, 1989; Russett,1990). It is not reasonable to demand that any belief structure should encompassall possible aspectsof foreignaffairs, and thereis indeed rather persuasiveevidence thatattitudes toward some ratherimportant issues cut across the main dimensionsidentified above. Trade and protectionism,an issue that is likelyto become more ratherthan less contentiousduring the 1990s, is one such example; questions revolvingaround Israel and American policy toward that nation appear to formanother cluster of attitudesthat does not fitneatly into the MI/Cl scheme. A somewhatdifferent approach towardattitude structures emerges from several importantstudies of the mass public conducted byJon Hurwitzand Mark Peffley (1987, 1990; Peffleyand Hurwitz,1985, 1992a, 1992b). In contrastto Converse's search for "horizontal"coherence that relies on correlationsamong attitudes towardvarious issues, Hurwitzand Peffleyproposed and tested a hierarchically organized foreignpolicy belief structurein which specificpolicy preferences are derivedfrom postures (militarism, anticommunism, and isolationism)that, in turn, are assumed to be constrainedby a set of core values (moralityof warfare,eth- nocentrism)about the internationalcommunity. They found thatsuch structures did in factexist among respondentsto theirsurveys. Thus, a fewrather general beliefs-for example,attitudes toward militarism, or a generalpreference toward a "tough-minded"approach toward foreignaffairs-appear to serve as heuristics whichenable one to respond in a reasonablycoherent manner to a broad range of issues, including defense spending, nuclear arms policy, militaryinvolvement abroad, Sovietpolicy, and internationaltrade (Hurwitzand Peffley,1987). It is importantto state once again that none of these studies challenges the overwhelmingevidence that the American public is poorly informed about internationalaffairs; indeed, even the Persian Gulf War, the firstconflict to be telecastin real time,increased the normallylow level of informationamong the generalpublic byonly a verymodest amount (Bennett,1992). Rather,these studies appear to suggestthat, even in the absence of much factualknowledge, members of the mass public use some simple-perhaps even simplistic-heuristicsin order to make some sense of an increasinglycomplex world;a fewsalient criteria rather than complete informationmay serve as the bases ofjudgment. Stated differently, although lacking a deep reservoirof factualinformation, members of the mass public mayoperate as "cognitivemisers," employing a fewsuperordinate beliefs to guide theirthinking on a broad range of issues.For furtherevidence thatpeople organize theirpolitical worlds in richerand more diverseways than indicated by Converseand his colleagues,see Conoverand Feldman (1984). Clearlythe recentresearch has yet to produce a consensuson manyimportant issues relatingto the structureand foreignpolicy beliefs among the mass public. Nevertheless,it is evidentthat the earlier consensus depictingpublic attitudesas lackingany real coherence has been challengedfrom various quarters. As a result of substantialempirical research, there is now a good deal of credible evidence suggestingthat the public does use variousheuristics-although not necessarilythe traditional liberal-to-conservativeor internationalist-isolationistblueprints-for organizingpolitical thinking.

Challenge#3: Is PublicOpinion Really Impotent? Among the mostimportant questions about public opinion are: To whatextent, on what kinds of issues, under what circumstances,and in what typesof political OLE R. HOLSTI 451 systems,if any, does it have an impacton public policy?If it has an influence,what are the means bywhich public attitudesmake theirimpact felt by decision-makers? These are also the most difficultquestions, for our abilityto answerthem is not materiallyenhanced by the manytechnical improvements that have characterized public opinion research during the past half century:better sampling designs, greaterattention to constructionof questions,more sophisticatedstatistical models to analyzethe data, and the like. Not surprisingly,then, we have a good deal more systematicevidence describingthe state of, or trendsin, public opinion than on how it has affectedthe actual conductof foreignaffairs. As noted earlier,most of the evidence throughthe 1960s pointed towardthe conclusion of public impotencein the foreignpolicy-making process. Even when thereappeared to be some correspondencebetween public sentimentsand foreign policy,not all analystswere prepared to accept the inferencethat the formerhad any independent impact on the latter.According to several of them, including revisionisthistorians, any evidence of a correlationbetween public opinion and foreignpolicy merelyserves to underscore the effectivenessof effortsby policy makers,aided and abettedby pliant print and electronicmedia, to manipulatethe masspublic intoacceptance of the rulingelites' class-basedinterests. There is no shortageof evidence thatmost post-WorldWar II presidentshave followedTheodore Rooseveltin thinkingthat the WhiteHouse is a "bullypulpit," whetherit is used to "scarethe hell out of them"in order to gain supportfor aid to Greece and Turkey, to warn against the dangers of "unwarrantedinfluence, whethersought or unsought,by the military-industrialcomplex," or to drum up supportfor assistance to the contrasin Nicaragua.It is also evidentthat such efforts have not been equally successful.At least one noted public opinion analysthas asserted that the relationshipbetween leaders and the public has changed- "farewellto 'the Presidentknows best,"' as he put it-but it remainsto be demon- stratedthat the equation has been permanentlychanged (Yankelovich,1979). The more difficultquestion concernsthe influencein the otherdirection. How much did public impatiencelead the Carteradministration to embarkon the ill- fated effortto free American hostages in Iran, or the Reagan administrationto withdrawU.S. Marines fromLebanon? Did PresidentKennedy genuinely believe thathe would be impeached should he failto forceremoval of Sovietmissiles from Cuba, as he told his brother,or was he merelyseeking to bolsterdecisions arrived at forreasons thathad nothingto do withpublic opinion? Perhaps a more telling example from the Caribbean confrontationemerges from recentlypublished transcriptsof a crucial White House meetingon October 27, the day before the crisiswas resolvedpeacefully. It appears that Kennedywas prepared to accept a compromise solution-withdrawalof Soviet missiles in Cuba in exchange for removalof Americanmissiles in Turkey-ratherthan initiatea furtherescalation of the confrontation,and thathe would have done so in large part becauseit would havebeen hard to explain to the public why such a seeminglyequitable arrangement had been rejected(Bundy and Blight,1987). Some other anecdotal evidence may also be suggestive,but it hardlyoffers irrefutableanswers to thisquestion. Franklin D. Rooseveltwas the firstpresident to make extensive use of public opinion data, but during recent decades every administrationhas employedpublic opinion professionals.We have relativelyfew detailed accountsfrom these analystsabout how theirexpertise and data wereused in the policyprocess, but those that exist (e.g., Cantril,1967; Beal and Hinckley, 1984; Hinckley,1992) suggestthat the mass public is not viewed merelyas an essentiallyshapeless, malleable lump that can readilybe molded throughpublic relations activitiesand compliant media to meet the immediate needs of the administration.Hadley Cantril, who undertook public opinion analyses for 452 Public Opinionand ForeignPolicy

PresidentsRoosevelt, Eisenhower, and Kennedy,summarized his experiencein this manner:"I wantto emphasizethat no claim is made here thatthe [public opinion] data and suggestionsLambert and I provided the President [Roosevelt] were crucial to his decisions. But actions taken were certainlyvery often completely consistentwith our recommendations"(Cantril, 1967:42). Althoughit bears repeatingthat the evidence describingpublic opinion stillfar outstrips,both in quantityand quality,that on the causal links between public opinion and foreignpolicy, research in recentyears has begun to cast doubt on the earlierconsensus about public impotence.In addition to anecdotal evidence,two classes of studies have contributedto challengingthe thesis that foreignpolicy processes are imperviousto impact from the public: quantitative/correlational analysesand case studies. Severalrecent quantitative studies have challengedsome importantfoundations of the theorythat, at least on foreignand defense issues, the public is virtually impotent.One elementof thatthesis is thatpolicy makers are relativelyfree agents on foreignpolicy questions because these issues pose few dangers of electoral retributionby voters; elections are said to be decided by domestic questions, especiallythose sometimesdescribed as "pocketbook"or "bread and butter"issues. However, a systematicstudy of presidentialcampaigns between 1952 and 1984 revealed thatin fiveof the nine electionsduring the period, foreignpolicy issues had "largeeffects." Or, as the authorsput it,when presidentialcandidates devote campaign time and other resourcesto foreignpolicy issues, theyare not merely "waltzingbefore a blind audience" (Aldrich,Sullivan, and Borgida,1989). Recent research on voting behavior has also emphasized the importanceof retrospectiveevaluations on performanceon voter choice among candidates, especiallywhen one of them is an incumbent(Fiorina, 1981; Abramson,Aldrich, and Rhode, 1990). Because votersare perceived as punishingincumbent candidates or partiesfor foreign policy failures (for example, the Iran hostageepisode) or rewardingthem for successes (for example, the invasionof Panama to capture General Noriega), decisionsby foreignpolicy leaders maybe made in anticipation of public reactionsand the probabilitiesof successor failure. The electoralretribution hypothesis received a differentkind of testin a study ofAmerican policy toward China duringthe threedecades followingestablishment of Mao Tse-tung'sgovernment in 1949. Kusnitzfound that,with few exceptions, the correspondencebetween public preferencesand U.S. policywas remarkably high.At timespolicy led opinion and at othertimes opinion led policy,but on the whole the two remained in harmony.These findingsare explained by issue visibility,partisan differences,and the nonrandom changes of opinion, which combined to create the belief among leaders that the possibilityof electoral retributionrequired them to pay close attentionto public opinion on the China issue (Kusnitz,1984:173, 176). Anotherrecent study seems to cast some doubt on the universalvalidity of the classic Miller-Stokes(1963) findingthat, compared to domesticissues, public attitudes on foreignpolicy questions have farless impacton membersof Congress.A careful analysisof voting on Pentagon appropriationsat the beginning of the Reagan administration's defense buildup revealed that ". . . public opinion was a powerfulforce for policy change in the realm of defensespending in the firstyear of the Reagan administration.Moreover, the impact of constituencyopinion appears to have been remarkablybroad-based, influencing all sortsof representa- tivesacross a wide spectrumof specificdefense spending issues" (Bartels, 1991:467). Finally,two major studieshave measured the congruence between changes in public preferencesand a broad range of policies over extended periods.The first, a studyof public opinion and policy outcomes spanning the years 1960-1974, revealed that in almost two thirdsof 222 cases, policyoutcomes correspondedto OLE R. HOLSTI 453 public preferences.The consistencywas especiallyhigh (92%) on foreignpolicy issues. Monroe (1979:11) offersthree possible explanations for his findings: Foreign policyissues permitmore decision-makingby the executive,are likelyto be the object of relativelyless interestand influenceby organized interestgroups, and are especiallysusceptible to elite manipulation.The second studycovered an even longerspan-1935 to 1979-which included 357 significantchanges of public preferences(Page and Shapiro, 1983). Of the 231 instancesof subsequent policy changes,153 (66%) were congruentwith changes in public preferences.There was littledifference in the level of congruencefor domestic (70%) and foreignpolicy (62%) issues. Although anecdotal evidence and correlational analyses can make useful contributionstoward understanding the opinion-policyrelationship, they are not an entirelysatisfactory substitute for intensivecase studiesthat could shed more directlight on how, if at all, public opinion influencesforeign policy-making. It is not whollysufficient to describethe stateof or trendsin public opinion on an issue immediatelypreceding or during foreignpolicy decisions.A findingthat major decisionsseemed to be correlatedwith public preferencesdoes not,by itself,estab- lish a causal link;for example, policy-makers might be respondingto pressuresand constraintsfrom the internationalsystem-as realist theoristsinsist that they should-without any significantattention to public sentimentson the issue,even if those attitudes are highly congruent with those of the decision-makers. Alternatively,the actual directionof causalitymight run frompolicy-makers to the public ratherthan vice versa,as depicted by criticswho describethe public as the malleable targetsof public relationsefforts by American elites (Ginsburg,1988; Margolisand Mauser, 1989). When opinion change precedes policychange, this interpretationloses potency.However, we could not rule out stillanother possibility:the administrationmanipulates events; the events,now part of the information available to the public, resultin opinion change, followedby policychanges that are congruentwith opinions. A somewhatrelated variant of this sequence is the "rally'round the president"hypothesis, according to whichthe executivemay manipulatethe public indirectlyby undertaking external initiatives and responding to eventsabroad in a manner calculated to increase his popularitywith domestic constituents(Brody and Shapiro,1989; Marra,Ostrom, and Simon, 1990). Among the more difficultcases are those dealing with public opinion as a possible constrainton action. During the 1980s, the Reagan administration undertook a massive public relations campaign of dubious legalityto generate public support for assistance to the "contra" rebels in Nicaragua (Parry and Kornbluth,1988), but a carefulanalysis of surveyson the issue revealed that a majorityof the public opposed Americanmilitary involvement in CentralAmerica (Sobel, 1989; see also Hinckley,1992). Would the Reagan administrationhave intervenedmore directlyor massivelyin Nicaragua or El Salvadorin the absence of such attitudes?Solid evidence about contemporarynon-events is, to understatethe case, ratherhard to come by. Case studiesseem to be the onlyway to address such questions,although even this approach is not whollyfree of potentialproblems. Does an absence of documentaryreferences to public opinion indicate a lack of interestby decision-makers?Alternatively, was attention to public attitudesso deeply ingrainedin theirworking habits that it was unnecessaryto make constant referencesto it? Are frequentreferences to public opinion an indication of a significantimpact on decisions-or of a desire on the part of officialsto be "on record"as havingpaid attentionto public sentiments? These examples do not imply that we are limited to simple one-directional models of the links between the public and policy makers; examples of more complex alternativesmay be found in Rosenau (1961), Graber (1968), Russett (1990), Hughes (1978), and Hinckley (1992). Moreover, a full analysisof the 454 Public Opinionand ForeignPolicy opinion-policylinks would often require explorationsinto many aspects of the domesticpolitical process, including the role of partiesand candidates in raising issues, the impact of interestgroups, the role of the media, the level of elite competitionon specificissues, and the like. The literatureon each of these topics, even if strictlylimited to foreignpolicy issues in the United States,is enormous, and space limitationspreclude their inclusion in this essay; Bauer, Pool, and Dexter (1963) and Hughes (1978) are usefulin thisrespect. In order to selectbetween competing hypotheses about opinion-policylinkages, there are no satisfactoryalternatives to carefullycrafted case studies employing interviewsand/or archivalresearch designed to uncover how, if at all, decision- makers perceive public opinion; feel themselvesmotivated or constrainedby it; factorit into theiranalyses of policy options; and otherwisetake it into account when selectinga course of action,including a decisionnot to takeexternal action. Althoughthe literatureaddressing these questionsis not large,especially when compared to the number of studies that describe the state of public opinion, several examples illustratethis type of research. The availabilityof substantial collections of documents relatingto the 1914 European crisisenabled Richard Fagen (1960) to studythe uses and assessmentsof public opinion duringthe weeks leading up to World War I. Doris Graber undertook an intensivestudy of four decisionsduring the earlyperiod of Americanhistory-Adams's decision to renew negotiationswith France in 1800, the Louisiana Purchase, Madison's policies leading up to the War of 1812, and enunciationof the Monroe Doctrine. Despite personality,ideological, and other differencesamong the four presidents,she found that in each case public opinion was "an importantfactor in decision making,but byno means the mostimportant single factor" (Graber, 1968:318). A studyof foreignpolicy-making about a centurylater came to a ratherdifferent conclusion. RobertHilderbrand was unable to discoverthat public opinion had a significant impact on foreign policy-making during the quarter century (1897-1921) encompassingthe McKinley,Roosevelt, Taft, and Wilson administrations; to the extentthat public opinion entered into executivediscussions, it was onlyafter policy decisions had been made (Hilderbrand,1981:202). Stilldifferent findingsemerged from a studyof public opinion and foreignpolicy from the period leading up to World War II through PresidentTruman's 1947 speech requestingaid to Greece and Turkey.Michael Leigh (1976) testedtwo approaches to the foreignpolicy process: the traditionalmodel that the public constrains Americanpolicy makers, versus the radical model thatmanipulation of the public in favorof predeterminedpolicy choices not only takes place but also invariably succeeds. His findingsvalidated neither model. Strikingevidence thatpublic opinion has a significantimpact on policyemerges froma studyof fourcases of Americanarms controlpolicy-international control of atomicenergy, the LimitedTest Ban Treaty,the SALT I/ABM Treaty,and SALT II-spanning every administrationbetween Presidents Truman and Reagan. Graham (1989) used a researchdesign thatincluded an analysisof over 500 public opinion surveysand an examinationof primarysource materialsto determineif correlationsbetween public opinion and policydecisions were causal or spurious. The evidence revealed thatpublic opinion had an importantimpact on decisions at all stagesof the policyprocess, from getting on the agenda throughnegotiation, ratification,and implementation.Moreover, its impact varied, depending on whether public support for a policy reached the level of majority(50-59%), consensus(60-69%), preponderant(70-79%) or virtuallyunanimous (80%+). Studies of the opinion-policylinks are not limitedto cases in which sufficient time has passed to permit full examination of the relevant archives. Philip Powlick'sanalysis of the role of public opinion in U.S. decisions on the Lebanon OLE R. HOLSTI 455 interventionduring the firstReagan administrationrelied almost wholly on interviewswith policy makers. Whereas public opinion influencedmany mid-level officialsand a few higher ones-for example, Caspar Weinberger8-it had little impact on others, including Ronald Reagan, Robert McFarlane, and George Shultz. Foreign policy officialstended to regard congressional moods as the relevantmanifestation of public opinion.9Powlick concluded thatpublic opinion formedthe basis of severalrecommendations to pull the Marinesout of Lebanon; it helped ensure thatthe decision to withdrawwould be warmlyreceived by most officialsand members of Congress. However, President Reagan's decision to withdrawwas apparentlyless influencedby public opinion than by the kinds of external sources that realistswould advocate following.Public opinion was thus onlyone of severalfactors that came togetherto bringabout the evacuationfrom Beirutin February,1984 (Powlick,forthcoming). Taken as a whole, these case studieswould seem to suggestthat the impact of public opinion has increasedduring recent decades. This tentativeconclusion also receives some support from interviewsof foreignpolicy officials.Although the bureaucratsinterviewed by Powlick(1991) were not notablymore sanguine about the public than were those takingpart in similarresearch by Cohen (1973) two decades earlier,they were more inclinedto accept the legitimacyof a public contri- bution to the policyprocess. Consequently,these officialstended to avoid policies thatwere seen as likelyto engenderpublic opposition.In contrast,Graham (1989) found littlechange over timebecause the public opinion has played an important role in arms controldecisions since the Truman years.This, then, is one of the areas in whichthe need foradditional research is mostapparent.

OtherRecent Research Developments

OpinionLeaders Until recently,one of the glaringgaps in public opinion research has been the neglectof opinion leaders. Virtuallyall approaches to government-fromtheories that view the United States as a pluralistdemocracy to those that depict it as pseudo-democracyruled by self-perpetuatingelites who seek to use foreignpolicy as an instrumentof narrow,class-based interests-recognize the disproportionate influence of some citizens (Devine, 1970). Moreover, at least since Almond's seminalstudy of TheAmerican People and Foreign Policy (1950), it has been customary to distinguishbetween various strata of the public. Typicallya distinctionhas been drawnbetween opinion leaders,the informedpublic, and the mass or uninformed public,although the precise termsand the shape of the distributionamong strata mayvary from study to study(compare Kreisberg,1949; Almond, 1950; Rosenau, 1961; Genco, 1984; and Neuman, 1986). But onlyrarely have systematicstudies of respondents in the top strata been undertaken. On occasion the Gallup Organizationhas surveyedsamples of persons listedin Who's Who in America-for example, a July1953 poll on tariffs-butnot frequentlyenough to be of value in

8Among six conditionsthat should be met before the United Statescommits combat forcesabroad, according to Weinberger,"there must be some reasonable assurance we will have the supportof the Americanpeople" (New York Times,November 26, 1984, A5:1). There is also evidence that General Schwarzkopfwas reluctantto fightin the PersianGulf without appropriate public support(Hinckley, 1992:20). 9 If congressionalmoods are more generallyviewed by the executiveas expressionsof public opinion, then it opens up another large body of evidence on interveningvariables between public opinion and foreignpolicy. For recentstudies that explore these linkageson the StrategicDefense Initiative,weapons procurement,and sanctionson South Africa,see Lindsay(1990, 1991) and Hill (1992). 456 Public Opinionand ForeignPolicy assessing trends. Until the firstof the Chicago Council on Foreign Relations surveysin 1974 (Rielly,1975) -followed by similarstudies in 1978, 1982, 1986, and 1990 (Rielly,1979, 1983, 1987, 1991)-and the Foreign PolicyLeadership surveys institutedtwo years later (1976, 1980, 1984, 1988, and 1992), therewas relatively littlesystematic information about leadershipviews on foreignaffairs. In addition to the CCFR and FPLP surveys,during recent years there have also been a number of one-time studies of opinion leaders (Barton, 1974-75, 1980; Russett and Hanson, 1975; Sussman,1976; Chittickand Billingsley,1989). A continuingseries of surveys,focused on national securityexperts, has also been initiatedrecently (Koopman, Snyder,andJervis, 1990a, 1990b, 1991). The prior question of identifying"opinion leaders" is among the possible barriersand sources of controversyin studiesof this type.There is, of course, an extensivedebate on the precise definitionof opinion leaders in the United States, and if consensus on thatquestion were a prerequisiteto elite surveys,they would never be undertaken.Those designingleadership studieshave typicallybypassed the broader question and selected subjectsin one of twoways: (1) identifyingkey roles and then surveyinga sample of those fillingthem, or (2) identifyinggroups thoughtto be logical sources of opinion leaders and then drawingsamples from directoriesor rostersof such persons-for example, biographiesin Who'sWho in America, Who's Who in American Politics, and similar sources; subscribers to the journal InternationalSecurity; or studentsat the NationalWar College. A somewhatnarrower area of disagreementrevolves around the abilityto use data from national probabilitysamples for studyingelites. To oversimplify somewhat,the contendingviews can be reduced to twopositions on the adequacy of respondents'education levels as a surrogatemeasure of leadershipstatus. The affirmativeview was presented in a prognosisof how public opinion research is likelyto develop duringthe nexthalf century. The mathematicsand economics of surveysmake them most cost-effectivefor assessinglarge undifferentiatedpopulations, i.e. mass publics. And untilyou get to a handful of individualsat the very top, you don't learn very much from studying"elites"-they seem to be just like bettereducated people in the general population. These two generalizationssuggest no major shiftin our attention. The same is true in totalitariancountries (and, I suspect,poor countries). (Davis, 1987:S178-S179) Although the question is far from settled,at this point the proponents of the opposite viewpointwould appear to have at least a plausible case. A number of studieshave found that,by itself,the level of educational attainmentis an inadequate yardstickfor identifying opinion leaders.Extensive analyses of the 1968, 1980, and 1984 National Election Studiesand the firstthree Chicago Council surveysby Krosnickand Carnot (1988) indicatethat education is an insufficientindicator of the attentiveforeign policy public, much less of foreignpolicy opinion leadership. Their findingsreveal little support for the hypothesisthat the "foreignpolicy attentivepublic" is composed simplyof highlyeducated personswho are concerned withall aspectsof public policy.An earlierstudy also castdoubt on the sufficiencyof educationas a measureof leadership(Rogers, Stuhler, and Koenig,1967). Several studies have compared the views of opinion leaders and the general public (Luttbeg, 1968; Rielly,1975, 1979, 1983, 1987, 1991; Holsti, 1987, 1991; Chitticket al., 1990; Wittkopf,1990; Schneider, 1992). Much of the evidence indicatesgreater similarities than differencesin the waysthat respondents in the two groups structuretheir political beliefs (Luttbeg, 1968; Wittkopf,1990), but Chittickand his colleagues (1990) found some differenceswhen foreignpolicy beliefstructures are traced over time.On the otherhand, thereexist some rather substantialdifferences in the substanceof their policy opinions. A few examples fromthe mostrecent (1990) Chicago Council surveywill be sufficientto illustrate OLE R. HOLSTI 457 some of the more persistentgaps betweenopinion leaders and the general public (Rielly,1991). Leaders overwhelmingly(97%) supported an active U.S. role in world affairs,whereas the general public was more tepid in its enthusiasm(59%). Both groups have shownlittle variation in thisrespect over a period of nearlytwo decades. Almostthree fourths of the leaders agreed thatthe Cold War has ended, but only about a third of the general public expressed that view. Whereas 90 percentof the leaders favoredforeign economic aid, especiallyto EasternEurope, onlyhalf that proportion of the general public did so; moreover,among the latter, fightingdrugs in Latin Americawas a favoritetarget of assistanceprograms. When asked to identifythe most importantforeign policy problems, the top threeitems forleaders included issueswith an overwhelminglyinternational character: spread of nuclear weapons, arms control,and improvementof the internationalenvironment.The comparablelist for the general public included issueswith a significant domestic dimension: protectionof Americanjobs, protectionof U.S. business interestsabroad, and adequate energysupplies. The momentouschanges in the formerSoviet Union and elsewherethat have characterizedthe period since 1988 suggestthe possibility-butcertainly not the inevitability-ofa "great debate" on the appropriate values, goals, roles, and strategiesthat the United States should pursue in the post-Cold War era. If one assumes that top officialsand opinion leaders will have overwhelminginfluence, whereasthe general public willhave littleor none, then the evidence suggeststhat the United Stateswill continue to pursue an internationalistforeign policy, broadly defined. In that case, the debates are likelyto center on how the United States participates(hard-liners vs. accommodationists).If, on the other hand, public preferencesplay a significantrole in shaping at least the broad contours of Americanforeign policy, then the policydebates are likelyto focus on whetherthe United Statesshould playan activeinternational role or focusmore on issues that have a directdomestic impact (internationalists vs. isolationists).

TheSources of Foreign Policy Opinions During the early post-WorldWar II years,bipartisan cooperation between the White House and Congress made possible such initiativesas aid to Greece and Turkey,the Marshall Plan, and NATO. Each of these strikingdepartures from traditionalforeign policies had rathersolid public support;for further evidence on these and other undertakings,see Levering (1978), Foster (1983), and Wittkopf (1990:166-193). Agreementamong prominentleaders of the twomajor partiesno doubt contributedto the fact that, among the general public, Democrats and Republicansdiffered little with respect to these and other internationalistforeign policyundertakings. A 1946 Gallup Poll revealedthat 72 percentof respondentsin both politicalparties favored an "active"international role for the United States, and the 1947 programof aid to Greece and Turkeyreceived approval from56 percent of both Democrats and Republicans. Nor were there notable partisan differencesin oppositionto sendingU.S. troopsto Indo-Chinain 1953 and 1954. Since the VietnamWar, sharp partisandifferences, reinforced by even deeper ideological cleavages,have characterizedthe foreignpolicy beliefs of both opinion leaders and the general public (Wittkopf,1990; Holsti and Rosenau, 1984, 1990). These differenceshave persisted into the 1990s, although President Reagan's second-termturnaround on such issues as arms control and conciliationof the SovietUnion did tend to reduce Republican and conservativeopposition to such policies,thereby closing somewhat the partisanand ideologicalgaps (Holsti,1991). The ideological-partisancleavages on domesticissues tend to reinforcethose on foreignpolicy. This was distinctlynot the case duringthe period betweenthe end of World War II and escalation of the conflictin Vietnam.In theirstudy of the 458 Public Opinionand ForeignPolicy

1948 election, Berelson and his colleagues (1954) found a limited correlation between domesticeconomic issues and either civil rightsor internationalissues. 'To know,for example, that someone supportedthe New Deal on economic issues provided no indication of his internationalor civil rightsposition" (Berelson, Lazarsfeld,and McPhee, 1954:197-198). For furtherevidence of the lack of correlationbetween domesticand foreignpolicy issues during the decade and a halfafter World War II, see Campbell et al. (1964:113) and Key (1961:158). During the past twodecades the lines of cleavageon domesticand foreignpolicy issues have increasinglycome to overlap (Russettand Hanson, 1975; Wittkopf, 1990). The evidence is especiallystrong at the level of opinion leadership (Holsti and Rosenau, 1988). Consequently,the putative,moderating influence of cross-cut- tingcleavages has oftenbeen absentin the stridentpolitics of the 1970sand 1980s. Two other "gaps" have received considerable attentionfrom opinion analysts. The VietnamWar era spawned interestin generationaltheories of politics.The writingof distinguishedsocial scientists,from Karl Mannheimto Sam Huntington, seemed to offercogent hypothesesto explain the divisionsof American society duringthe Vietnamtrauma, pitting, according to some, the "Munichgeneration," which had witnessedthe bitterconsequences of effortsto appease expansionist dictators,against the 'Vietnamgeneration," which argued thatit was experiencing the poisoned legacy of an ill-informedeffort to apply the "lessonsof Munich" to the jungles of SoutheastAsia (Allison, 1970-71; Roskin,1974). This explanation has encountered mixed success when confrontedwith hard evidence. Wittkopf (1990) found evidence of generational differencesamong the general public. However,at the level of opinion leaders the primarylines of cleavage appear to lie withinrather than between generations(Holsti and Rosenau, 1980, 1990). The generationalhypothesis appears to farebest when applied to specificgroups such as politicalactivists (Converse, 1987) or protesters(Jennings, 1987). The term "gender gap" has also become a prominentpart of our political vocabulary.There is some evidence of a systematicgender gap. For example, Converse (1987:61) found substantialgender differencesand that "'Rambo' themes come very disproportionatelyfrom males." According to Baxter and Lansing (1983), war/peaceconcerns are an exceptionto the generalrule thatmen and women agree on most issues. Wittkopf(1990) also found gender-based differencesalong both the militantinternationalist and cooperativeinternation- alist dimensions, and another study identified gender as among the most importantdemographic predictors of foreignpolicy attitudes(Fite, Genest, and Wilcox, 1990). A broad examination of surveydata led Shapiro and Mahajan (1986) to the conclusion that systematicgender differencesalso emerge from "compassion" issues and those involvingregulation and protection. However, Conover (1988) argues that "feminism,"an ideological predispositiontoward a certain stance on women's issues, overshadowsgender in explainingbeliefs and values. A number of recentAmericans Talk Securitysurveys also cast some doubt on stereotypicalviews of gender differences.Women consistentlyexpressed more skepticalopinions about the USSR than did men. Althoughwomen were more inclined to describe themselvesas doves (ATS-9:119), men offeredmore support on all fivearms controlitems by an average marginof 6 percent;expressed more trustin the USSR on six of seven items (6 percent); assessed Gorbachev more favorablyon eight items (9 percent); had more benign viewsof Sovietmotives on six of nine items (3 percent); assessed relationsbetween the superpowersmore favorablyon ten of thirteenitems (5 percent); and expressed a more optimistic view on the likelihood of nuclear war (10 percent). In the only exception to this pattern,women were more criticalon threeitems relating to the impactof defense spending (6 percent). OLE R. HOLSTI 459

At the level of opinion leaders, evidence of systematicgender differencesis harderto come by,except on such issuesas tradeand the environment;women are more protectionistand pro-environment.Four FPLP surveysspanning a dozen years indicate that women in leadership positions hold attitudes,including on war/peace issues,that are generallyquite similarto those of men in comparable roles. Moreover,in theirpolitical attitudes,liberal women are likelyto resemble liberalmen, Republicanwomen tend to resembleRepublican men, and so on. Surveyresearchers have also produced evidence on other backgroundfactors, includingoccupation, education, military service, travel abroad, region,and so on, but the mutuallyreinforcing partisan and ideological cleavages have clearlybeen dominant in recent years. The public opinion data generally support the conclusion of three perceptive observers: "For two decades, the making of Americanforeign policy has been growingfar more political-or more precisely, farmore partisanand ideological" (Destler,Gelb, and Lake, 1984:13). Whetherthe end of the Cold War will contributeto softeningand blurringthe partisan- ideologicalchasm remains to be demonstrated.

Conclusion The consensus of the mid-1960son the nature, structure,and impact of public opinion has clearly come under vigorous challenge during the past quarter century.The VietnamWar, while not the sole causal factorin the reexaminationof the conventionalwisdom, was certainlya catalyst.If a new consensus has yet to emerge on all of the issues discussed above, at least it seems safe to statethat the fieldis markedby innovative research and activedebates on the implicationsof the results.Nevertheless, there are at least four areas that seem to call for additional effort:case studiesemploying archival sources thatwill provide the most directly relevant evidence on the impact of public opinion, cross-nationalresearch, developmentof standard questions in order to encourage cumulationof survey results,and research that will enable us to distinguishresults that are time-and context-boundfrom those thattranscend the Cold War period. As indicatedearlier, by farthe leastwell developed of the areas discussedin this essayhas been the opinion-policylink. We have impressivecorrelational evidence that policy changes are in fact predominantlyin the directionfavored by the public, but our confidencewould be enhanced by more substantialevidence of a causal nature.The typeof researchdesign employedin the previouslycited study by Thomas Graham (1989), combining analyses of surveydata with archival researchin order to assessthe causal impactof public opinion on decision making, is a good model forfuture studies. More researchof thistype should go a long way towardanswering some of the most importantquestions about the opinion-policy relationship.It should also providefurther insight into otherimportant questions: What are the relevantindicators of public opinion? How much do polls matter? What about expressionsof congressionalpreferences? What about the impact of the media or interest groups? How and why do policy makers (and administrations)differ in their sensitivityto differentindicators of public sentiments? It willno doubt have occurred to readersby now thatthis essay is almostwholly confined to evidence of American public opinion. But questions about public opinion and foreignpolicy are obviouslyof much broader concern,especially in an era of expanding democracy,and a good deal of futureeffort should be directed towardcomparative analysis. Some examples of innovativecomparative research include Eichenberg (1989) and Risse-Kappen (1991). The latter found that 460 Public Opinionand ForeignPolicy althoughpublic opinion was importantin each of four countries-France,Japan, West Germany,and the United States-its impact was significantlyaffected by domesticstructures and coalition-buildingprocesses among elites. Distinctionsof thiskind clearlytake us a long waytoward a fullerunderstanding of opinion-policy linkages.Dramatic changes in EasternEurope and the formerSoviet Union should open up possibilitiesfor a range of comparativestudies that would have been quite impossiblejust a fewyears ago. Mueller (1973:265) has appropriatelypointed to an importantlimitation on opinion surveysarising from the factthat "the poll interviewis a ratherprimitive stimulus-responsesocial situationin whichpoorly thought out answersare casually fittedto questionsthat are overlyingenuous." At minimum,that warning reminds us to be cautiousabout makinginferences from a singledatum on public opinion. It is permissibleto be more venturesomein drawingtrends from longitudinal data, but doing so assumes that the questions posed have remained constant.Whereas commercialpolling organizationsask certain standardquestions at quite regular intervals-for example, on rating presidential performance-there is nothing comparable with a strictlyforeign policy focus. During and immediatelyafter World War II, respondentswere regularlyasked about the desirabilityof an active U.S. role in the world,but interestin that question appears to have died in the mid-1950s;it was only revivedin the wake of the war in Vietnam when global activismcame under increasingcriticism. In othercases, longitudinalanalyses may be rendered suspect, as in 1956 when Gallup made a "minor"alteration to its standard question on foreignaid by adding the phrase, "to preventtheir [the recipientsof aid] going communistic"at the end of the question. Among the usefulfeatures of the Chicago Council quadrennial studies (Rielly, 1975-1991) has been a carryoverof certainquestions from survey to survey.Their question asking respondentsto rate the importanceof a series of foreignpolicy goals has been especiallyuseful for those with an interestin trendanalyses. With a few exceptions, however, the independent surveysthat have been undertakenin recentyears appear to have taken littlenote of questions in other studiesthat mightprovide the basis for comparativeanalysis. In one sense thisis understandable;the whole rationale for an independent surveyis to undertake probes that have been overlookedby others.But it is also regrettablethat there appears to be ratherlimited communication at the planning stage between those who are designingsurveys. The developmentof even a handfulof standardforeign policy questions thatwould be included in all such surveyswould go a long way towardimproving a less-than-outstandingrecord of cumulativefindings. Finally,so much of the evidence cited above has emerged froma four-decade- long period in whichforeign affairs were dominatedby the Cold War thatwe need to address questions about whetherand how the end of that confrontationmay affector even render obsolete what we have learned about public opinion and foreignpolicy. At the most obvious level, there has been a sea change in public attitudestoward virtually all of the issuesthat dominated the Cold War era. Indeed, one could make a plausible case that in manyrespects changing public attitudes precededrather than followedthose at the pinnacles of governmenton such issues as the appropriate level of defense spending, the primarythreats to American national security,assessments of Mikhail Gorbachev's goals, and the motivations underlyingSoviet foreignpolicy (ATS surveys;Holsti, 1991). For example, well before the demolitionof the Berlin Wall or the final collapse of the USSR, the public ranked the danger to Americannational securityfrom the SovietUnion in seventhplace, just behind the greenhouseeffect (ATS-9:51-54). The end of the Cold War also raises some questions about the structuringof foreign policy attitudes.Substantial evidence indicates that assessmentsof the SovietUnion have playeda keyrole in foreignpolicy belief structures; for example, OLE R. HOLSTI 461 it is a centralelement in the hierarchicalmodel developed by Hurwitzand Peffley (1990) as well as in Wittkopf's(1986, 1990) militantinternationalism dimension. Some interestingresearch questions emerge from the collapse of the SovietUnion. Will it resultfor manyin a loss of structureand consequent disorientationabout foreignaffairs? In a replacementof the Sovietsby another adversary such as Japan? Are there persons who, if deprived of one enemy, must search for another? Alternatively,are the key concepts that structurebeliefs about foreign affairs sufficientlygeneric thatthey will survive the dramaticinternational changes of the past fewyears? It is my guess that such dimensionsas militantinternationalism, cooperative internationalism,and unilateralism-multilateralismwill continue to structureforeign policy attitudes, but the changes we have witnessedsince the late 1980s are of such unprecedented magnitude that this must be treated as a hypothesisthat requires systematic testing. At the broadest level, if we are indeed enteringinto a period of fewermajor power confrontations,and greaterattention to such nonmilitaryissues as trade, immigration,the environment,and the like-there is ample surveydata thatmuch of the Americanpublic believesthis to be the case-it mayalso be an era in which public opinion playsa more autonomous role. Even those who do not fullyaccept the "manipulatedpublic" thesiswould acknowledgethat crises and confrontations abroad providea settingin whichopportunities and temptationsfor manipulating the public are far greater than on nonstrategicissues. Not only are the latter typicallyresolved over a longer time period-providing greateropportunities for the public, pressuregroups, the media, and Congress to play a significantrole- but they also tend to be more resistantto claims that the needs for secrecy, flexibility,and speed of action make it both necessaryand legitimatefor the executiveto have a relativelyfree hand. In short,we may be movinginto a period in which the relationshipbetween public opinion and foreign policy takes on added rather than diminished significance,but we should also be waryof assumptionsthat the theories,evidence, and linkages emergingfrom research during the Cold War era will necessarily travelintact into an era of strikinglydifferent circumstances.

References

ABRAMSON, P., J. H. ALDRICH, AND J. RHODE (1990) Change and Continuityin the 1988 Election. Washington,DC: CongressionalQuarterly. AcHEN, C. H. (1975) Mass PoliticalAttitudes and the SurveyResponse. AmericanPolitical Science Review 69:1218-1231. ALDRICH,J. H., J. L. SULLIVAN, AND E. BORGIDA (1989) ForeignAffairs and Issue Voting:Do Presidential Candidates "Waltzbefore a Blind Audience?"American Political Science Review 83:123-141. ALLISON, G. T. (1970-71) Cool It: The ForeignPolicy of Young Americans.Foreign Policy 1:144-160. ALMOND, G. (1950) 7he AmericanPeople and ForeignPolicy. New York:Praeger. Reprintedin 1960 witha new introduction. ALMOND, G. (1956) Public Opinion and National Security.Public Opinion Quarterly 20:371-378. AMERICANS TALK ISSUES (1991) Serial National Surveys of Americans on Public Policy Issues. Washington,DC: AmericansTalk Issues Foundation. AMERICANS TALK SECURITY (1987-91) A Series of FifteenSurveys Concerning National SecurityIssues. Winchester,MA: AmericansTalk Security. BAILEY, T. A. (1948) 7'heMan in theStreet: 7he Impactof American Public Opinion on ForeignPolicy. New York:Macmillan. BARDES, B. A., AND R. OLDENDICK (1978) Beyond Internationalism:The Case for MultipleDimensions in ForeignPolicy Attitudes. Social Science Quarterly 59:496-508. BARTELS, L. M. (1991) ConstituencyOpinion and CongressionalPolicy Making: The Reagan Defense Buildup. AmericanPolitical Science Reviezv 85:457-474. 462 Public Opinionand ForeignPolicy

BARTON,A. H. (1974-75) Conflictand Consensus among American Leaders. Public OpinionQuarterly 38:507-530. BARTON,A. H. (1980) Fault Lines in AmericanElite Consensus.Daedalus 109:1-24. BAUER, R. A., I. DES. POOL, AND L. A. DEXTER (1963) AmericanBusiness and PublicPolicy. New York: AthertonPress. BAXTER,S., AND M. LANSING (1983) Womenand PoliticsThe VisibleMajority. Ann Arbor: Universityof MichiganPress. BEAL,R. S., AND R. H. HINCKLEY (1984) PresidentialDecision Making and Opinion Polls. Annals 472:72-84. BENNETr, S. E. (1972) Attitude Structuresand Foreign Policy Opinions. Social ScienceQuarterly 55:732-742. BENNETT,S. E. (1992) The Persian Gulf War's Impact on Americans' Political Information. Washington,DC: NES Conferenceon "The PoliticalConsequences of War." BERELSON,B., P. F. LAZARSFELD,AND W. N. MCPHEE (1954) Voting.Chicago: Universityof Chicago Press. BILLINGTON,J.(1987) Realismand Vision in ForeignPolicy. Foreign Affairs 65:630-652. BRODY,R., AND C. SHAPIRO (1989) "A Reconsiderationof the RallyPhenomenon in Public Opinion." In PoliticalBehavior Annual, edited byS. Long. Boulder,CO: WestviewPress. BUNDY, M. [TRANSCRIBER],AND J. G. BLIGHT [EDITOR] (1987) "October 27, 1962: Transcriptof the Meetings of the ExComm." InternationalSecurity 12:30-92. CAMPBELL, A., P. E. CONVERSE, W. E. MILLER, AND D. E. STOKES (1964) The AmericanVoter. New York: Wiley. CANTRIL,H. (1967) The Human Dimension:Experiences in PolicyResearch. New Brunswick,NJ: Rutgers University Press. CASPARY,W. R. (1970) The Mood Theory: A Studyof Public Opinion and Foreign Policy. Ame7ican PoliticalScience Review 64:536-547. CHITTICK, W., AND K. R. BILLINGSLEY(1989) The Structureof Elite Foreign Policy Beliefs. Western PoliticalQuarterly 42:201-224. CHITTICK, W., K R. BILLINGSLEY, AND R. TRAvIS (1990) Persistenceand Change in Elite and Mass Attitudestoward U.S. ForeignPolicy. Political Psychology 11:385-402. CLIFFORD,C., WITHR. HOLBROOK (1991) Counselto the President: A Memoir.New York:Random House. COHEN, B. C. (1973) ThePublic's Impact on Foreign Policy. Boston: Little,Brown. CONOVER,P.J. (1988) Feministsand the Gender Gap. Journalof Politics 50:985-1010. CONOVER,P. J.,AND S. FELDMAN(1984) How People Organize the PoliticalWorld: A SchematicModel. AmenicanJournal of Political Science 28:95-126. CONVERSE,P. E. (1964) "The Natureof BeliefSystems in Mass Publics."In Ideologyand Discontent,edited byD. E. Apter.New York:Free Press. CONVERSE,P. E. (1987) "The EnduringImpact of the VietnamWar on AmericanPublic Opinion." In Afterthe Storm: American Society a Decade after the Vietnam War. Taipei, R.O.C.: Academic Sinica. DAvIS,J. A. (1987) The Future Study of Public Opinion: A Symposium.Public Opinion Quarterly 51:S178-S179. DESTLER, I. M., L. H. GELB, ANDA. LAKE (1984) Our Own WorstEnemy. New York:Simon and Schuster. DEVINE,D. J. (1970) TheAttentive Public. Chicago: Rand McNally. EICHENBERG,R. (1989) Public Opinion and National Securityin WesternEurope. Ithaca, NY: Cornell UniversityPress. FAGEN,R. (1960) Some Assessmentsand Uses of Public Opinion in Diplomacy.Public Opinion Quarterly 24:448-457. FIORINA,M. (1981) RetrospectiveVoting in AmericanNational Elections. New Haven, CT: Yale UniversityPress. FITE, D., M. GENEST, AND C. WILCOX (1990) Gender Differencesin Foreign PolicyAttitudes. American PoliticsQuarterly 18:492-513. FOSTER,H. S. (1983) ActivismReplaces Internationalism. Washington, DC: Foxhall Press. FREE, L., ANDW. WATTS (1980) InternationalismComes of Age ... Again. PublicOpinion 3 (April-May): 46-50. GENCO, S. J. (1984) The AttentivePublic and American Foreign Policy. Ph.D. dissertation,Stanford University,Stanford, CA. GINSBURG,B. (1988) TheCaptive Public: How Mass OpinionPromotes State Power. New York:Basic Books. GRABER,D. A. (1968) PublicOpinion, the President, and ForeignPolicy: Four Case Studiesfrom the Formative Years. New York: Holt, Rinehart and Winston. GRAEBNER,N. A. (1983) "Public Opinion and Foreign Policy:A PragmaticView." In Interaction:Foreign Policyand PublicPolicy, edited by D. C. Piper and R. J. Terchek. Washington,DC: American Enterprise Institute. OLE R. HOLSTI 463

GRAHAM,T. W. (1986) PublicAttitudes toward Active Defense: ABM and StarWars, 1945-1985. Cambridge, MA: Centerfor International Studies, M.I.T. GRAHAM, T. W. (1988) The Patternand Importanceof Public Knowledgein the Nuclear Age.Journal of ConflictResolution 32:319-334. GRAHAM, T. W. (1989) The Politicsof Failure: StrategicNuclear Arms Control,Public Opinion, and Domestic Politicsin the United States-1945-1980. Ph.D. dissertation,M.I.T., Cambridge,MA. HERO, A. O., JR. (1969) Liberalism-ConservatismRevisited: Foreign vs. Domestic Federal Policies, 1937-1967. PublicOpinion Quarterly 33:399-408. HILDERBRAND,R. C. (1981) Powerand thePeople: Executive Management of Public Opinion in ForeignAffairs, 1897-1921. Chapel Hill: Universityof NorthCarolina Press. HILL, K. A. (1992) The Domestic Sources of Foreign Policymaking:American Mass Attitudesand CongressionalVoting on South AfricanSanctions. Paper presentedat the annual meetingof the InternationalStudies Association, Atlanta, GA. HINCKLEY, R. H. (1988) Public Attitudestoward Key Foreign PolicyEvents. Journal of ConflictResolution 32:295-318. HINCKLEY,R. H. (1992) People,Polls, and Policy-Makers.New York:Lexington Books. HOLMES,J. E. (1985) TheMood/Interest Theory of American Foreign Policy. Lexington: University Press of Kentucky. HOLSTI,0. R. (1979) The Three-Headed Eagle: The United States and SystemChange. International StudiesQuarterly 23:339-359. HOLSTI,0. R. (1987) "PublicOpinion and Containment."In Containingthe Soviet Union, edited byT. L. Deibel andJ. L. Gaddis. Washington,DC: Pergamon-Brassey's. HOLSTI,0. R. (1991) "AmericanReactions to the USSR: Public Opinion." In Soviet-AmericanRelations afterthe Cold War, edited byR. Jervisand S. Bialer. Durham,NC: Duke UniversityPress. HOLSTI,0. R., AND J. N. ROSENAU (1979) Vietnam,Consensus, and the Belief Systemsof American Leaders. WorldPolitics 32:1-56. HOLSTI,0. R., AND J.N. ROSENAU(1980) Does Where You Stand Depend on When You Were Born? The Impact of Generation on Post-VietnamForeign Policy Beliefs. Public Opinion Quarterly 44:1-22. HOLSTI, 0. R., AND J. N. ROSENAU (1984) AmericanLeadership in WorldAffairs: Vietnam and theBreakdown ofConsensus. London: Allen and Unwin. HOLSTI, 0. R., AND J. N. ROSENAU (1988) The Domestic and Foreign Policy Beliefs of American Leaders.Journal of Conflict Resolution 32:248-294. HOLSTI, 0. R., AND J. N. ROSENAU (1990) The Structureof Foreign PolicyAttitudes among American Leaders.Journal of Politics 52:94-125. HUGHES, B. B. (1978) TheDomestic Context of American Foreign Policy. San Francisco:Freeman. HUGHES, T. L. (1980) The Crack-Up.Foreign Policy 40:33-60. HULL, C. (1936) Speech at Pan AmericanConference, Buenos Aires,December 5, 1936. Text in New YorkTimes [Dec. 6, 1936], p. 47. HURWITZ, J., AND M. PEFFLEY (1987) How Are Foreign Policy AttitudesStructured? A Hierarchical Model. AmericanPolitical Science Review 81:1099-1120. HURwITz,J., AND M. PEFFLEY (1990) Public Images of the SovietUnion.Journal of Politics 52:3-28. JENNINGS, M. K. (1987) Residue of a Movement:The Aging of the American Protest Generation. AmericanPolitical Science Review 81:367-382. JENTLESON, B. W. (1992) The PrettyPrudent Public: Post-PostVietnam American Opinion on the Use of MilitaryForce. InternationalStudies Quarterly 36:49-73. KENNAN, G. F. (1951) AmericanDiplomacy, 1900-1 950. Chicago: Universityof Chicago Press. KEY,V. O., JR.(1961) PublicOpinion and AmericanDemocracy. New York:Knopf. KINDER, D. R. (1983) "Diversityand Complexityin AmericanPublic Opinion." In PoliticalScience: The State of theDiscipline, edited by A. W. Finifter.Washington, DC: American Political Science Association. KINDER, D. R., AND D. 0. SEARs (1985) "Public Opinion and Political Action." In Handbookof Social Psychology,edited byE. Aronsonand G. Lindzey.New York:Random House. KLINGBERG, F. L. (1952) The HistoricalAlternation of Moods in AmericanForeign Policy.World Politics 4:239-273. KLINGBERG, F. L. (1979) "Cyclical Trends in American Foreign Policy Moods and Their Policy Implications."In Challengesto America:United States Foreign Policy in the 1980s, edited by C. W. Kegley,Jr., and P. J. McGowan.Beverly Hills, CA: Sage. KLINGBERG, F. L. (1983) CyclicalTrends in AmericanForeign Policy Moods: The Unfoldingof America's World Role.Lanham, MD: UniversityPress of America. 464 Public Opinionand ForeignPolicy

KOOPMAN,C.,J. SNYDER, AND R.JERvIs(1990a) AmericanElite Viewsof Relationswith the SovietUnion. Journalof Social Issues 45:119-138. KOOPMAN,C., WITH J. SNYDER AND R. JERVIS (1990b) Theory-Drivenvs. Data-DrivenAssessments in a Crisis.Journal of Conflict Resolution 34:694-722. KOOPMAN,C., J. SNYDER, AND R.JERVIS (1991) U.S. Elite Views about Change and Changes in Viewson Soviet-U.S. Relations. Paper prepared for the annual meeting of the InternationalSociety of PoliticalPsychology, Helsinki. KREISBERG, M. (1949) "DarkAreas of Ignorance."In PublicOpinion and ForeignPolicy, edited by L. Markel.New York:Harper. KROSNICK, J. A. (1988a) AttitudeImportance and AttitudeChange. Journal of Experimental Social Psychology24:240-255. KROSNICK, J. A. (1988b) The Role of AttitudeImportance in Social Evaluation: A Study of Policy Preferences,Presidential Candidate Evaluations,and Voting Behavior.Journal of Personality and SocialPsychology 55:196-210. KROSNICK, J. A. (1990) Americans' Perceptionsof PresidentialCandidates: A Test of the Projection Hypothesis.Journal of Social Issues 46:159-182. KROSNICK, J. A. (1991) The Stability of Political Preferences: Comparisons of Symbolic and NonsymbolicAttitudes. American Journal of Political Science 35:547-576. KROSNICK, J. A., AND C. CARNOT(1988) Identifyingthe Foreign AffairsAttentive Public in the U.S.: A Comparisonof CompetingTheories. Mimeographed.Ohio State University,Columbus, OH. KUSNITZ,L. A. (1984) PublicOpinion and Foreign Policy: America's China Policy. Westport, CT: Greenwood Press. LAFEBER, W. (1977) "AmericanPolicy-Makers, Public Opinion and the Outbreak of the Cold War, 1945-1950." In The Originsof theCold War in Asia, edited by Y. Nagai and A. Iriye.New York: Columbia UniversityPress. LEIGH, M. (1976) MobilizingConsent: Public Opinion and AmericanForeign Policy. Westport, CT: GreenwoodPress. LEVERING,R. B. (1978) ThePublic and American Foreign Policy, 1918-1978. New York: Morrow. LINDSAY, J. (1990) Parochialism, Policy, and ConstituencyConstraints: Congressional Voting on StrategicWeapons Systems.American Journal of Political Science 34:936-960. LINDSAY, J. (1991) Testing the Parochial Hypothesis:Congress and the StrategicDefense Initiative. Journalof Politics 53:860-876. LIPPMANN,W. (1922) PublicOpinion. New York:Macmillan. LIPPMANN,W. (1925) ThePhantom Public. New York:Harcourt, Brace. LIPPMANN,W. (1955)Essays in thePublic Philosophy. Boston: Little, Brown. LIPPMANN,W., AND C. METZ(1920) A Test of the News. Special Supplement.New Republic 23:1-42. LIPSET,S. M. (1966) The President,the Polls, and Vietnam.Transaction (September/October):19-24. LUTTBEG, N. R. (1968) The Structureof Beliefsamong Leaders and the Public. PublicOpinion Quarterly 32:398-409. MANDELBAUM, M., AND W. SCHNEIDER (1979) "The New Internationalisms."In Eagle Entangled: U.S. ForeignPolicy in a ComplexWorld, edited byK. A. Oye etal. New York:Longman. MARGOLIS, M., AND G. A. MAUSER, EDS. (1989) ManipulatingPublic Opinion. Pacific Grove, CA: Brooks/Cole. MARRA, R. A., C. W. OSTROM, AND D. M. SIMON (1990) Foreign Policy and PresidentialPopularity. Journalof Conflict Resolution 34:588-623. MILLER, W. E., AND D. E. STOKES (1963) ConstituencyInfluence in Congress.American Political Science Review57:45-46. MONROE, A. D. (1979) Consistencybetween Public Preferencesand National PolicyDecision. American PoliticsQuarterly 7:3-19. MORGENTHAU, H.J. (1978) Politicsamong Nations, 5th ed., rev. New York: Knopf. MUELLER,J.E. (1973) War,Presidents and Public Opinion. New York: Wiley. NEUMAN,W. R. (1986) TheParadox of Mass Politics: Knowledge and Opinionin theAmerican Electorate. Cambridge,MA: HarvardUniversity Press. NIE, N., AND K. ANDERSON (1974) Mass Belief SystemsRevisited: Political Change and Attitude Structure.Journal of Politics 36:540-591. NIE, N., S. VERBA, AND J. R. PETROCIK (1976) TheChanging American Voter. Cambridge, MA: Harvard UniversityPress. OLE R. HOLSTI 465

NINCIC,M. (1988) The United States,the Soviet Union, and the Politics of Opposites. WorldPolitics 40:452-475. PAGE,B. I., AND R. Y. SHAPIRO (1982) Changes in Americans' Policy Preferences,1935-1979. Public OpinionQuarterly 46:24-42. PAGE,B. I., AND R. Y. S-APIRO (1983) Effectsof Public Opinion on Policy. AmericanPolitical Science Review77:175-190. PAGE,B. I., AND R. Y. SHAPIRO(1984) Presidentsas Opinion Leaders: Some New Evidence. PolicyStudies Journal12:649-661. PAGE,B. I., AND R. Y SHAPIRO(1988) Foreign Policy and the Rational Public. Journalof Conflict Resolution32:211-247. PAGE, B. I., AND R. Y. SHAPIRO (1992) The Rational Public:Fifty Years of Trendsin Americans'Policy Preferences.Chicago: Universityof Chicago Press. PAGE, B. I., R. Y. SHAPIRO,AND G. R. DEMPSEY (1987) What Moves Public Opinion? AmericanPolitical ScienceReview 81:23-43. PARRY, R., AND P. KORNBLUTH(1988) Iran-Contra'sUntold Story.Foreign Policy 72:3-30. PATERSON,T. G. (1979) PresidentialForeign Policy,Public Opinion, and Congress:The Truman Years. DiplomaticHistory 3:1-18. PEFFLEY,M. A., ANDJ.HURWITZ (1985) A HierarchicalModel of AttitudeConstraint. AmericanJournal of PoliticalScience 29:871-890. PEFFLEY, M., ANDJ. HURWITZ (1922a) InternationalEvents and Foreign PolicyBeliefs: Public Responses to ChangingSoviet-American Relations. American Journal of Political Science 36:431-461. PEFFLEY,M., ANDJ. HURWITZ (1922b) Models of AttitudeConstraint in ForeignAffairs. Political Behavior. POWLICK,P. J. (1991) The AttitudinalBases for Responsivenessto Public Opinion among Foreign PolicyOfficials. Journal of Conflict Resolution 35:611-641. POWLICK,P. J. (forthcoming)"Public Opinion and the Lebanon Intervention."In Mass Communication, Democratization,and thePolitical Process. Westport, CT: Greenwood. RIELLY, J. E., ED. (1975) AmericanPublic Opinion and U.S. ForeignPolicy 1975. Chicago: Chicago Council on ForeignRelations. RIELLY,J. E., ED. (1979) AmericanPublic Opinion and U.S. ForeignPolicy 1979. Chicago: Chicago Council on ForeignRelations. RIELLY,J. E., ED. (1983) AmericanPublic Opinion and U.S. ForeignPolicy 1983. Chicago: Chicago Council on ForeignRelations. RIELLY,J. E., ED. (1987) AmericanPublic Opinion and U.S. ForeignPolicy 1987. Chicago: Chicago Council on ForeignRelations. RIELLY, J. E., ED. (1991) AmericanPublic Opinion and U.S. ForeignPolicy 1991. Chicago: Chicago Council on ForeignRelations. RiSSE-KAPPEN,T. (1991) Public Opinion, Domestic Structure, and Foreign Policy in Liberal Democracies. WorldPolitics 43:479-512. ROGERS,W. C., B. STUHLER,AND D. KOENIG(1967) A Comparison of Informedand General Public Opinion on U.S. ForeignPolicy. Public Opinion Quarterly 31:242-252. RoOT,E. (1922) A Requisitefor the Success of Popular Diplomacy.Foreign Affairs 1:3-10. RoSENAU,J.N. (1961) PublicOpinion and ForeignPolicy. New York:Random House. ROSKIN,M. (1974) From Pearl Harbor to Vietnam: ShiftingGenerational Paradigms and Foreign Policy.Political Science Quarterly 89:563-588. RUSSETT,B. (1990) Controllingthe Sword. Cambridge, MA: HarvardUniversity Press. RUSSETT, B. M., AND E. C. HANSON (1975) Interestand Ideology:The ForeignPolicy Beliefs of American Businessmen.San Francisco:Freeman. SCHNEIDER, W. (1992) "The Old Politicsand the New World Order." In Eagle in a New World,edited by K. A. Oye, R.J. Lieber, and D. Rothchild.New York:HarperCollins. SHAPIRO,R. Y., AND H. MAHAJAN(1986) Gender Differencesin Policy Preferences:A Summaryof Trends fromthe 1960s to the 1980s. PublicOpinion Quarterly 50:42-61. SHAPIRO, R. Y, AND B. I. PAGE (1988) Foreign Policy and the Rational Public. Journalof Conflict Resolution32:211-247. SNIDERMAN, P. M., AND P. E. TETLOCK (1986) "Inter-relationshipof Political Ideology and Public Opinion." In PoliticalPsychology, edited byM. G. Hermann. San Francisco:Jossey-Bass. SOBEL, R. (1989) Public Opinion about United StatesIntervention in El Salvador and Nicaragua. Public OpinionQuarterly 53:114-128. 466 Public Opinionand ForeignPolicy

SULLIVAN, J. L., J. E. PIERESON, AND G. E. MARcus(1978) Ideological Constraintin the Mass Public: A MethodologicalCritique and Some New Findings.American Journal of Political Science 22:234-249. SUSSMAN,B. (1976) Elitesin America.Washington, DC: WashingtonPost. VERBA, S., AND R. A. BRODY (1970) Participation,Policy Preferences,and the War in Vietnam. Public OpinionQuarterly 34:325-332. VERBA, S., R. A. BRODY, E. B. PARKER, N. H. NIE, N. W. POLSBY, P. EKMAN, AND G. BLACK (1967) Public Opinion and the War in Vietnam.American Political Science Review 61:317-333. WITTKOPF, E. R. (1986) On the Foreign Policy Beliefsof the American People: A Critique and Some Evidence. InternationalStudies Quarterly 30:425-445. WITTKOPF, E. R. (1990) Faces ofInternationalism: Public Opinionand ForeignPolicy. Durham, NC: Duke UniversityPress. YANKELOVICH,D. (1979) Farewellto "PresidentKnows Best." Foreign Affairs 57:670-693.