sex, politics, and public opinion: what political scientists really … · sex, politics, and...

Sex, Politics, and Public Opinion:What Political Scientists Really LearnedFrom the Clinton-Lewinsky Scandal*Arthur H. Miller, University of Iowa

William Jefferson Clinton is onlythe second president to have

been impeached by the House ofRepresentatives. While he was notremoved from office, roughly half ofthe Senate chamber voted him guiltyon the perjury and obstruction ofjustice charges that the House for-mulated as articles of impeachmentin early February 1999. AlthoughClinton remains in office, numerouspoliticians, political pundits, and me-dia commentators continue to con-demn his extramarital affair withMonica Lewinsky while she was aWhite House intern as immoral andreprehensible.

Many outspoken Republicanshave raised questions about how theAmerican people could continue tohave confidence in their politicalleaders, government institutions, orthe rule of law when a presidentwho lied to the public, lied underoath, and obstructed justice remainsin office. Others ask how they canexplain this immoral behavior andthwarting of the law to their chil-dren, especially in light of all thesalacious details that were madepublic in Special Prosecutor Ken-neth Starr's report on the Clinton-Lewinsky affair.

News of a possible extramaritalaffair between Bill Clinton andMonica Lewinsky during the 1995-96period surfaced in January 1998,coming to light in the course of in-vestigations related to Paula Jones'sexual harassment lawsuit1 against

Arthur H. Miller is a professor of polit-ical science and the director of the IowaSocial Science Institute at the University ofIowa. His studies of trust in governmentstarted in the 1970s. Since then, his workon this topic has expanded to includecomparisons of the United States withcountries in Western Europe and areas ofthe former Soviet Union.

the president. Tapes Linda Trippmade of her conversations withMonica Lewinsky and turned over toStarr's office confirmed the allega-tions. What followed was a year-longexpose as details of the affair gradu-ally tumbled into the public domain.The year ended with the House vot-ing for impeachment on December19, 1998, and the new year beganwith the Senate voting not to re-move Clinton on February 12, 1999.

The purpose here is not to chroni-cle the events of that year but tosearch for potentially broadereffects in the immediate aftermathof the scandal. The goal is to scien-tifically determine to what extent theClinton-Lewinsky affair may haveundermined the American public'strust in government and politicalinstitutions.

I first review the relevant litera-ture on trust in government, thenassess how the public reacted to var-ious aspects of the affair, Starr's in-vestigation, and the impeachmentprocess in the House of Representa-tives. Next, I examine change in thepublic's perception of Clinton's char-acter. Given that the mass mediawere so prominent in keeping thestory before the American people, Ialso consider how well the mediafulfilled their role as watchdogs ofthe public interest. Additionly, I ex-amine the merits of Hillary Clinton'scharge that a right-wing conspiracylay behind the Starr investigationand impeachment from the perspec-tive of ideological divisions withinthe public. Subsequently, I assess theimpact that public reaction to theaffair had on Clinton's ratings, sup-port for Congress and the Republi-can Party, and trust in governmentmore generally. I sketch the broaderimplications of my analysis in a briefconcluding section.

Relevant LiteratureGovernment legitimacy is the

foundation of every democracy.Concerns about the vitality of de-mocracy arise, therefore, when pub-lic trust in political leaders and insti-tutions falls. Generally, Americans'trust in government has been declin-ing since the mid-1960s, with a briefperiod of rising trust in the mid-1980s. Many researchers have calledattention to this long-term trend(see e.g., Citrin 1974; Craig 1993;Miller 1974; Miller and Listhaug1990, 1993, 1998; Nye, Zelikow, andKing 1997). Explanations proposedfor this decline include poor govern-ment performance, public dissatis-faction with government policy out-comes and procedures, governmentdeficits and general economic hard-ship, political ideology, and even therise of post-material values.

When Clinton's affair with Lewin-sky became public, it was immedi-ately treated by the news media as arevelation that might potentiallydrive trust of government evenlower. Very few researchers, how-ever, have examined whether theimmoral behavior of political leadersleads to citizen distrust of govern-ment. For example, none of the au-thors in the most recent compen-dium of explanations for decliningconfidence in American governmentconsiders immoral behavior as acausal factor of political discontent(Nye, Zelikow, and King 1997).

Perhaps this oversight stems fromthe impact David Easton's A SystemsAnalysis of Political Life (1965) hashad on how social scientists thinkabout diffuse public support of gov-ernment. Easton argued that dissat-isfaction with specific policies orspecific leaders would not influencethe more generic and fundamentaltrust that citizens have in govern-ment and political institutions.

PSOnline www.apsanet.org 721

There should be no enduring impactbecause periodic elections provide amechanism for voting the rascalsout, thereby bringing in new leadersand new policies. Such thinking wasechoed by Citrin (1974) when heargued that the growing distrustfound by Miller (1974) in the early1970s was most likely just a tempo-rary phenomena caused by Nixon'srelatively low personal popularity.Miller (1974), in rebuttal, arguedthat unfavorable incumbent behaviorand dissatisfaction with policiescould have an enduring and erodingeffect on trust in government if itpersisted across elections, or over aseries of incumbents, particularly ifthe incumbents were of differentpolitical parties.

As the importance of politicalparties declined during the 1970sand 1980s, the prominence of thepresidency was accentuated, andeventually even Citrin (see Citrinand Green 1986) acknowledged thatpublic dissatisfaction with more pro-found aspects of presidential charac-ter, rather than superficial popular-ity, could have enduring detrimentaleffects on trust in government.Miller and Borelli (1991, 170) simi-larly argued that distrust of govern-ment may arise from negative publicassessments of politically relevantaspects of presidential character(such as strength of leadership, com-passion for others, and morality),but not the mere personal popularityof the incumbent. More recently,Norman Ornstein (1993) argued thatdeclining ethics and honesty amongpoliticians in general have contrib-uted to the growth of political dis-trust.

Given these various theoreticalarguments, it is certainly plausiblethat Bill Clinton's apparent philan-dering affected how the publicjudged him personally and how theyfelt about government and politicalinstitutions more generally. Thepresident is the most visible politicalactor in the country and if peopleuse his behavior as a heuristic tojudge government as a whole, thenhis perceived immorality may be asignificant determinant of broadertrust in government. To test thispossibility empirically, it is first nec-essary to ascertain how the public

reacted to various aspects of the af-fair, the Starr investigation, and theimpeachment process.

Empirical evidence from theHeartland Poll conducted by theIowa Social Science Institute is usedfor examining public reactions to theClinton-Lewinsky affair. This re-gional survey of citizens living inseven Midwestern states has beenconducted annually since 1988.2 Thefall 1998 survey was specifically de-signed to gauge the impact Clinton'saffair with Lewinsky had on the out-come of the congressional electionand broader assessments of govern-ment and political institutions. Allinterviews took place during thethree weeks preceding the Novem-ber 3 election and again during thetwo-week period of November 9-21,1998. While the 1998 Heartlandstudy is the primary source of data,I take data from earlier Heartlandsurveys for longitudinal comparisonsand make occasional references tovarious media polls.

Public Reactions tothe Scandal

Public reaction to the Clinton-Lewinsky affair was often ratherskewed. For example, virtually everynews poll conducted over the courseof the year demonstrated thatroughly two-thirds of the people feltthat Clinton had, at some point, liedabout the affair. Similarly, through-out 1998, in contradiction to the re-peated predictions of virtually everymedia commentator, political pundit,and Republican party leader, Clin-ton's job approval rating rose afterevery new revelation of presumablyyet-more-damning evidence concern-ing the affair.

At the beginning of 1998, Clin-ton's job approval rating hoveredjust above 50%. By the end of theyear, it was approaching 70%,higher than at any previous time inhis presidency. Republicans andnews commentators alike weredumbfounded at this outcome. Aclose look at attitudes towards keyaspects of the affair, the investiga-tion, and the impeachment revealthat the public's response was quiteunderstandable.

The fact that public opinion be-came polarized along party linesvery early, even before many factsabout the affair had been disclosed,goes a long way toward clarifyingwhy people did not become increas-ingly indignant as the scandal playedout. Partisan orientation subse-quently became very strongly corre-lated with attitudes concerning anumber of important aspects of theaffair and how it was handled, suchas whether it was a private or publicmatter and whether the process ofinvestigation was fair and impartial.For example, 30% of Republicans inthe Midwest and 84% of Democratsfelt that the Clinton-Lewinsky affairwas a private matter and ought notto be dealt with as an impeachableoffense. In contrast, a full 63% ofRepublicans saw the affair as a pub-lic issue that deserved investigation.These sharp partisan differenceswere also evident in responses tosurvey questions asking about theperceived fairness of the Starr inves-tigation and the impeachment pro-cess (see Table 1).

More interesting than the largedifference between Democrats andRepublicans is the high degree ofsimilarity between Independents andDemocrats. As Table 1 shows, Re-publicans were uniformly at oddswith nearly everyone else on virtu-ally all aspects of the Clinton-Lewin-sky issue. This is particularly true onthe question of whether the Clinton-Lewinsky encounter was a privateaffair that should have been keptout of the limelight or whether itwas a public issue. On this question,64% of Independents said it was aprivate matter.

In short, public reactions to theClinton-Lewinsky affair were deeplydivided along partisan lines, but notin the traditional style of Democratsand Republicans on opposite sidesand Independents either apatheticor evenly divided. Rather, a cohesivetwo-thirds of Republicans opposedoverwhelming majorities of bothDemocrats and Independents. Thatthis standoff was already evident rel-atively early in 1998 helps to explainwhy the media polls regarding pref-erences for impeachment remainedquite stable throughout the entireyear with roughly two-thirds op-

722 PS December 1999

TABLE 1Attitudes Toward Selected Elements of the Clinton-LewinskyIssue by Partisanship

Party Identification

Preferred Resolution:Clinton Leaves OfficeCongressional CensureDrop the Matter

TotalN

Opinion of the Starr Investigation:ImpartialBoth EquallyPartisan

TotalOpinion of the House ImpeachmentProcess:

ImpartialBoth EquallyPartisan

Total

Republican

64%2115

100%562

585

37100%

481537

100%

Independent

33%2542

100%844

244

72100%

251758

100%

Democrat

12%3157

100%603

102

88100%

121474

100%

Total

36%2539

100%2009

294

67100%

271657

100%

Source: The University of Iowa Social Science Institute (ISSI), 1998 HeartlandPoll (see note 2).

posed and one-third in favor. De-spite every new revelation, the pollsremained firmly fixed, much to thechagrin of Republicans and mediacommentators who continually pre-dicted that public opinion wouldcertainly change in light of the mostrecent revelation. Each time, theywere proven incorrect and each timeClinton's job approval would rise abit more in response to yet anothernew charge, or what most of thepublic perceived as yet another "un-fair" and "partisan" attack or innu-endo.

Clinton's ShiftingCharacter Image

Given that the public's reaction tothe continuous stream of allegationsabout the affair was strongly influ-enced by the partisan lenses theywore, one might expect that Clin-ton's overall image remained rela-tively stable despite the scandal. Theempirical evidence reveals the oppo-site. Certain aspects of public judg-ments regarding Clinton's charactershifted dramatically over the courseof the year, even among Democrats.Considerable media attention fo-cused on Clinton's character

throughout the year-long scandal.However, most of this discussionequated character with morality.News commentators and media pun-dits constantly expressed amazementat how 80% of people asked saidthey believed that Clinton had lied,and nearly the same percentage ap-proved of how Clinton was doing hisjob as president. How, the commen-tators wondered, could the publicsupport a president of questionablecharacter?

A major factor here, which themedia commentators did not seemto realize, is that "character" is mul-tidimensional. Aristotle pointed outcenturies ago that character is com-prised of, at least, good sense, goodwill, and good morals. More con-temporary scholars have similarlydescribed presidential character asan amalgam of leadership, compas-sion for others, and morality (Kind-er 1986; Miller, Wattenberg, andMalanchuk 1986).

One might have expected, as themedia commentators clearly did,that the scandal would underminethe public's judgment of all aspectsof Clinton's character. Yet the em-pirical evidence demonstrates thatthe public is very capable of differ-

entiating how they evaluate the vari-ous aspects of character. As Figure1 vividly demonstrates, the public'sestimation of Clinton's leadershipabilities was not diminished at all bythe scandals of 1998. The percent-age of Midwesterners seeing Clintonas a strong leader in 1998 was virtu-ally the same as when he was firstelected in 1992 and reelected in1996 (see Figure 1).

On the other hand, revelations ofhis affair and proof that he liedabout it did diminish people's regardfor Clinton's compassion for othersand his morality. At the time of his1992 election, Bill Clinton was per-ceived as a caring and compassion-ate person by 75% of Midwestern-ers; by the end of 1998, that figurehad fallen to 62%. While only afraction of that decline can be at-tributed to his affair with Lewinsky,because most of the drop occurredby 1994, the scandal may possiblyaccount for the drop between 1996and 1998.

Much more pronounced is thedecline in the percentage of Mid-westerners who saw Bill Clinton asmoral. Although public judgments ofClinton's morality had alreadysoured significantly by 1996, thechange from 1996 to the end of1998 was dramatic among all parti-san subgroups (see Figure 1).

Relatively few Republicans hadever thought of Clinton as a moralindividual (25% in 1996 and only12% in 1998). More than three-fourths of Democrats (77%), androughly half of Independents (49%),on the other hand, thought of Clin-ton as moral in 1996, but less thanhalf of the members of these groupsjudged him as moral toward the endof 1998 (49% and 26% respectively).Clearly, the revelations of Clinton'sscandal-related behavior provided aninitial disappointment for manyDemocrats while it simply confirmedwhat most Republicans already felt.Despite earlier stories about Clin-ton's involvement with GenniferFlowers, Paula Jones, and evenother women, many Democrats hadcontinued to think of Clinton as amoral individual, thus suggestingthat what constitutes moral behaviorfor these individuals involves some-


FIGURE 1Perceptions of Clinton as a Strong Leader,Compassionate, and Moral

80

Strong Leader

40

201992 1994

Source: ISSI Heartland Poll (see note 2)

thing other than casual sex or mari-tal infidelity.

Realizing that the public con-ceives of character as comprisingtraits other than just morality helpsto clarify and make sense of howindividuals evaluated Clinton. If me-dia commentators had attributedmore sophistication to the publicthan they did, they too would haverealized that viewing the presidentas immoral but a strong and capableleader who cared about others wasnot a contradiction for many.

Evaluating the WatchdogsThe discussion thus far suggests

that the performance of the massmedia in reporting on the Clinton-Lewinsky affair and its aftermathwas less than exemplary. When thestory first broke, a number of newscommentators made predictions thatif Clinton did indeed have an affairwith Monica Lewinsky he wouldmost likely resign soon or else besummarily impeached and removedfrom office. The mass media fre-quently released information suppos-edly leaked from authoritativesources that they subsequently hadto retract or that was proven false.One frequently got the impressionthat the journalists were all Wood-

1996 1998Year

ward and Bernstein wannabes, butthat they did not want to put anyeffort into checking out theirsources.

Public opinion surveys conductedover the course of the scandal con-tinually revealed that a majority ofcitizens held views that ran counterto those expressed by most mediacommentators. Commentators ingeneral favored impeachment andremoval, which was clearly not whatthe public preferred (see the CBSNews/New York Times or CNN/Timepolls conducted over the course ofthe year and available from ORSPublishing). Eventually, therefore,the commentators started to attackthe polls for being misleading andinaccurate. On October 8, J998, forexample, Cokie Roberts claimed onThis Week that the polls were under-estimating the negative sentimentagainst Clinton. Her argument wasthat the national polls were domi-nated by East and West coast peo-ple who are more favorable towardClinton whereas the people in theMidwest, who she claimed werelargely in favor of impeachment,were underrepresented in nationalpolls. Clearly, the Heartland Polldata presented above demonstratethat Ms. Roberts' on-the-air analysiswas quite inaccurate.

Moreover, news media were con-stantly filled with stories about thescandal even though the public over-whelming reported that the affairwas receiving too much coverage.Even 73% of Republican Midwest-erners said there was too much cov-erage.

The conclusion that the media'sorientation was predominately anti-Clinton and pro-impeachment issupported by the differences in howDemocrats and Republicans judgedthe performance of the media incovering the issue. Nearly two-thirds(65%) of the Democrats rated themedia coverage as poor to verypoor, whereas only 44% of Republi-cans rated the coverage negatively(56% said it was good to very good).Given the media's predominantlynegative assessments of the presi-dent, it would be plausible to expectthat the heavy media coverage mayvery well have had some impact onwhether the scandal influenced thepublic's trust in the president, theCongress, the Republicans, and evengovernment more generally. I returnto this topic below.

A Right-Wing Conspiracy?A favorite media pastime was de-

bunking Hillary Clinton's assertionthat the Starr investigation, thePaula Jones lawsuit, and much ofthe gossip surrounding the Clinton-Lewinsky affair were given impetusby a Republican right-wing conspir-acy against her husband.

Excluding the conspiracy notion,the First Lady's statement raises ahypothesis worthy of examination.As demonstrated above, it was pri-marily Republicans who favored aninvestigation into the personal sex-ual behavior of the president andimpeachment for his transgressions.Yet, it may very well be the casethat not all Republicans shared thesame views of the investigation andimpeachment. If Hillary was correct,we should find that it was predomi-nantly right-wing Republicans whowanted Clinton punished.

Indeed, as the data presented inTable 2 reveal, the far right-wingconservative Republicans were mostcritical of Clinton in 1998.3 Mem-bers of the far right wing, which rep-


resents roughly 40% of all Republi-cans, almost uniformly perceivedClinton as immoral, a weak leader,and lacking compassion for others.They also overwhelmingly believedthat Clinton's affair with MonicaLewinsky should be a public issueopen to investigation, that the Starrinquiry and the impeachment pro-cess were free of partisan motiva-tion, and that Clinton should defi-nitely leave office either throughresignation or impeachment (seeTable 2).

The difference between far right-wing views and those of "middle-of-the-road" Republicans (about one-third of the party) is tremendous.The only item on which a majorityof moderate Republicans agreedwith the far right wing of the partywas the question of Clinton's moral-ity. Roughly 60% of moderate Re-publicans felt that Clinton was animmoral individual. Yet, even on themorality question there is a substan-tial difference between the views ofmoderate and far right wing Repub-licans (see Table 2). On all theother items in Table 2 the differencebetween moderate and far rightwing Republicans was even greater.While these huge ideological divi-sions with regard to the scandal oc-curred among the Republicans, lib-eral-conservative ideology was onlyweakly related to attitudes towardthe scandal among Independentsand Democrats.

Hillary Clinton may have beenwrong in suggesting a concertedright-wing conspiracy against herhusband, but it is clear that the farright wing of the Republican Partywas most opposed to Clinton andwas most in favor of removing himfrom office. The data in Table 2 alsoimply that the impeachment in theHouse represented only the wishesof the right wing of the RepublicanParty faithful. Given that impeach-ment proceeded, it is quite clearthat the right wing was controllingthe Republican Party agenda inCongress.

Impact of the ScandalWhile the public's reaction to is-

sues of Clinton's moral behavior, theStarr investigation, and the impeach-

TABLE2Attitudes Toward Clinton and Various Aspects of theLewinsky Scandal Among Republicans byIdeological Orientation

Attitudinal Assessments:Clinton is not at all MoralClinton is a Strong LeaderClinton is CompassionateAffair with Lewinsky is Public IssueStarr Investigation is ImpartialStarr Investigation is PartisanHouse Impeachment Process is

ImpartialHouse Impeachment Process is

PartisanClinton Must Leave Office

Middle ofthe Road

n = 19763654844455240

44

48

ModerateRight-Wing

n = 15274513765563945

39

64

FarRight-Wing

n = 21386341780791264

23

87

Note: Table entries are the percent of each ideological group giving the particu-lar survey response.Source: ISSI Heartland Poll 1998 (see note 2).

ment process are interesting in theirown right, my real purpose is to de-termine if these reactions influencedpolitical behavior or attitudes, par-ticularly trust in political parties,Congress, or government in general,as well as assessment of Clinton'sjob performance. During the year-long scandal, media commentatorsregularly explained Clinton's highjob approval ratings by saying thatthey reflected only people's satisfac-tion with the performance of theeconomy. According to this argu-ment, the booming economy gavepeople a reason to overlook Clin-ton's moral indiscretions. It is im-portant to examine this hypothesizedoutcome because the normative im-plication of the argument is thatAmericans will follow immoral lead-ers as long as they provide economicprosperity. Empirically, what thissuggests is that a multivariate analy-sis of Clinton's job approval ratingshould reflect how the public as-sessed the performance of the na-tional economy more than how itevaluated the events and outcomesof the scandal.

Seventy percent of respondents inthe fall 1998 Heartland survey ap-proved of Clinton's job performance(an approval rating comparable tothat found in national surveys at thetime). A multivariate analysis of his

approval rating demonstrates ratherconvincingly that individual assess-ments of the national economy hadrelatively little impact on how thepublic rated Clinton (see Table 3).Only among Independents did eval-uations of how the economy wasdoing currently relative to a yearearlier have a statistically significantimpact on ratings of Clinton's jobperformance. Similarly, the respon-dents' income had no effect on howthey rated the job Clinton was do-ing. In large part, this absence of aneconomic impact on ratings of Clin-ton's job performance arose becauseRepublicans and Democrats alikeagreed that the national economywas strong.

What did differentiate ratings ofthe president's performance werelargely judgments about the presi-dent's character and reactions to hisaffair with Lewinsky. In general, thepublic's perceptions of Clinton as astrong leader who was caring andcompassionate about others had agreater impact on Clinton's job ap-proval ratings than did evaluationsof his immorality (see Table 3).Among Republicans, however, theirevaluations of Clinton's morality hadthe greatest relative impact of allthe predictor variables. This con-firms earlier analysis suggesting thatRepublicans were far more fixated

PSOnline www.apsanef.org 725

TABLE 3Multivariate Analysis Predicting Clinton's JobApproval Ratings

Prediction Variables

Party IdentificationEvaluation of National EconomyClinton MoralClinton Strong LeaderClinton CompassionateMedia Coverage of ScandalAffair Private/Public IssueStarr Investigation ImpartialImpeachment Process ImpartialResolution of ScandalEducationIncomeAdjusted R2

N

Total

-.07*.04*.09*.23**.18**.02.17**

-.14**.02

-.19**.01.01.611537

Republicans

—.03.31**.11**.16**.01.14**

-.15**.04

-.29**-.04

.01

.64430

Independents

—.08*.12".20**.18**

-.08*.19**.16**.01

-.08*.04

-.02.51645

Democrats

—.01.04.15**.25**

-.06*.11**

-.10**-.05-.26**-.03

.01

.46462

*p <.01**p <.001Note: Table entries are standardized regression coefficients. The dependentvariable ranged from 1 = strongly approve of the job Clinton is doing as presi-dent to 5 = strongly disapprove. Party identification is coded 1 = Republican,2 = Independents, 3 = Democrats. Evaluation of the national economy nowcompared to a year ago is coded 1 = better, 2 = same, 3 = worse. Clintonmoral, strong leader, compassionate are all coded 1 = great deal, 2 = some-what, 3 = a little, 4 = not at all. Media coverage evaluation ranged from 1 =positive to 6 = negative. The affair was a private matter = 1, both = 2 or pub-lic = 3 issue. Perception of Starr investigation and impeachment process arecoded 1 = impartial, 2 = both, 3 = partisan. The preferred resolution of thescandal is coded 1 = Clinton leaves office through resignation or impeachment,2 = censure Clinton, 3 = drop the matter entirely. Education and income runlow to high.Source: ISSI, 1998 Heartland Poll (see note 2).

on the issue of morality than wereDemocrats or Independents.

Turning to the various aspects ofthe scandal, one finds more similar-ity in the equations for the partisansubgroups. For both Republicansand Democrats the preferred resolu-tion of the scandal (basically removeClinton versus dropping the matter)was a major factor explaining presi-dential job approval whereas percep-tions of the impeachment process asimpartial or partisan had no impactamong either partisan group or In-dependents (see Table 3). Perceivingthe Starr investigation as partisanand the affair as a private matter, onthe other hand, contributed tohigher job approval ratings for Clin-ton among all subgroups.

Responses to how the Lewinskyaffair was handled also had a signifi-cant impact on how individuals felt

about the Republican party. Overthe course of 1998 and on into 1999various polls (see ORS Publishingfor results of these polls includingGallup, The Washington Post, TheNew York Times, CBS and ABC,etc.) revealed roughly a 10-pointdecline in general support for theRepublican party (from 48% to37%). A separate multivariate analy-sis of the 1998 Heartland Poll pre-dicting affect toward the Republicanparty reveals that when the publicperceived the Starr investigation andthe House impeachment process aspartisan rather than impartial it un-dermined their support for the Re-publican party.4 The media mayhave reinforced this movement awayfrom the Republican Party: Thosewho evaluated the media coveragenegatively were less supportive ofRepublicans. At the same time, sup-

port for the Republican party wasclearly hurt by how Kenneth Starrhandled the investigation and byhow the Republican Judiciary Com-mittee members handled the im-peachment.

Increases or decreases in supportfor a presidential incumbent or apolitical party, while important tothe current functioning of govern-ment, could have rather short-livedeffects. No enduring damage to pub-lic support for the political regime,therefore, should necessarily be ex-pected to result from a decline insupport for the Republican Party, oreven a decline in Clinton's job ap-proval, if it happens to fall in theaftermath of the scandal. It wouldbe far more profound and disturbingif the Clinton-Lewinsky scandal wasfound to have had an impact on howthe public evaluated such politicalinstitutions as the Judiciary Commit-tee, Congress as a whole, or the gov-ernment in general. The empiricalevidence, however, strongly suggeststhat there appears to be very littleimmediate impact of the scandal onbroader feelings of public trust ingovernment. Those who perceivedthe Starr investigations and theHouse impeachment proceedings asfair and impartial were far moresupportive of Congress than werethose who saw these proceedings asreflecting partisan motivations (seeTable 4). Similarly, evaluations ofthe media coverage of the scandalalso had a significant impact on sup-port for Congress. Respondents whosaw the coverage in a positive lightwere more favorable toward Con-gress. One always needs to be cau-tious when interpreting correlationsinvolving media evaluations and out-come variables such as ratings ofCongress. Nevertheless, in this case,given all the other control variablesentered into the equation (such aspartisanship, education, ratings ofClinton, etc.), it does appear thatthe media were having a direct im-pact on the emerging evaluations ofCongress. The impact of the mediacoverage was strongest among Re-publicans and noticeably weakeramong Independents and Demo-crats. Once again, the variable mea-suring public reactions to Clinton'smorality was far more important


TABLE 4Multivariate Analysis Predicting Support for Congress andTrust in Government

Predictor Variables

Party IdentificationEvaluation of National EconomyClinton MoralClinton Strong LeaderClinton CompassionateMedia CoverageAffair Private/PublicStarr Investigation ImpartialImpeachment Investigation ImpartialResolutionEducationIncomeSupport for CongressAdjusted R2

N

Support forCongress

- .05-.01

.08*

.01

.01-.15**

.03-.09*-.20**

.01-.08*

.01—.15

1529

Trust inGovernment

.01

.03

.08*

.03

.20**

.04

.03

.01

.01

.03-.12**

.03-.28**

.141525

*p <.01**p <.001Note: Table entries are standardized regression coefficients. Support for Con-gress is measured as the average thermometer rating of the Judiciary Commit-tee and Congress. The scale ranges from 0 = most negative, 50 = neutral to100 = most positive. Trust in government combines responses to: "how oftendo you trust the government in Washington to do what is 'right'" and do youagree/disagree with the statement "public officials don't care much about peo-ple like you." The trust scale ranges from 1 = most trusting of government to 6 =most distrusting of government. The predictor variables are defined in Table 3.Source: ISSI, 1998 Heartland Poll (see note 2).

were those who took a more toler-ant stand on his morality.

The analysis presented in Table 4refers only to the end of 1998 and itmay very well be that the scandalhad a cumulative eroding effect ontrust more noticeable later. More-over, Midwesterners as a whole hadbecome more negative towards Con-gress between 1996 and 1998 (themean thermometer dropped from 54to 51). Thus, it is possible that theClinton-Lewinsky affair only indi-rectly undermined trust in govern-ment.

The longitudinal evidence on trustin government among Midwestern-ers, however, strongly refutes thehypothesis that the scandal precipi-tated any generalized distrust ofgovernment. Although the 1997Heartland Poll did not include anyitems for measuring trust in govern-ment, the comparison of 1998 trustlevels with those in 1996 and earlieryears demonstrates vividly that theClinton-Lewinsky affair had not un-dermined political trust (see Figure2). Despite the scandal, trust in gov-ernment actually increased amongDemocrats, Independents, and evenRepublicans between 1994 and theend of 1998.

among Republicans than among In-dependents or Democrats. ForDemocrats, the overriding factor intheir assessment of Congress wastheir perception that the House pro-ceedings were partisan rather thanimpartial and fair (see Table 4).

As the coefficients for the secondmultivariate equation in Table 4 re-veal, public judgments of the Starrinvestigation and the House im-peachment process were not directlyconnected with the extent to whichpeople expressed a general distrustof government. Judgments about theinvestigation and impeachment pro-cess may have had some indirecteffect on trust in government, be-cause ratings of Congress as an insti-tution were significantly related withtrust, but there certainly was not anydirect impact of the Clinton-Lewin-sky scandal on political trust as ofNovember 1998 (see Table 4).

What did have a direct bearing on

trust in government more generallywere various judgments that thepublic formulated regarding Clin-ton's character. Most important forthe public as a whole were judg-ments focused on whether Clintonwas a compassionate and caring in-dividual (see Table 4). That theseconcerns are related with trust ingovernment makes sense because agovernment absent leaders who careabout others would not be a respon-sive government people could trustto act in their best interests.

Judgments about Clinton's moral-ity were also significantly relatedwith attitudes on political trust.Again, however, this measure wasmuch more strongly related withtrust among Republicans than foreither Democrats or Independents.Republicans who perceived Clintonas immoral were substantially lesslikely to trust the government than

Conclusion

Throughout the year of the Clin-ton-Lewinsky scandal and subse-quent impeachment, a large majorityof the American public, while notcondoning his behavior, remainedfirmly fixed in their support of Presi-dent Clinton. This episode and thereaction of the public tell us a greatdeal about the current fault lines ofAmerican politics, the nature ofpublic opinion, and the deprofes-sionalization of the mass media inthe United States.

From the very outset of the scan-dal, a large majority of Americansperceived Clinton's behavior, includ-ing lying to conceal the affair, asimmoral. Yet, that very same major-ity positively evaluated Clinton's jobperformance and opposed his re-moval from office. These reactionswere not simply a reflection of eco-nomic good times or confusion onthe part of an uninformed public.Rather, they represented Americans'


FIGURE 2Trust in Government among Partisan Subgroups

-10

PDI

-601990 1991 1992 1993

Year

1994 1996 1998

Note: The Percentage Difference Index (PDI) reflects the preponderance of distrusting relative totrusting attitudes. It is computed by subtracting the percent distrusting from the percent trustingwhen using the two item trust index defined in Table 4. Trusting is defined as giving zero trustingresponses. The range of the PDI is +100 to - 1 0 0 , where a positive value reflects a preponderanceof trust while negative values reflect a preponderance of distrust.

Source: ISSI Heartland Poll (see note 2)

ability to differentiate a private actfrom a public concern and their re-alization that character is comprisedof more than just morality. Thus,contrary to what John Zaller (1998,188) previously concluded, employ-ing a multidimensional concept ofcharacter is critical if one is to un-derstand public reactions to politicalscandal.

For most, the scandal was notabout the rule of law or punishing apresident who had lied under oathor obstructed justice; it was about azealous special prosecutor who hadspent many of their tax dollars doingthe dirty work of the far right wingof the Republican party. The factthat the majority of the public re-mained steadfast in this beliefthroughout the year, despite the bar-rage of negative press and Republi-can charges against the president, isquite remarkable. Critics of publicopinion research often argue thatpublic opinion is volatile and thatattitudes are frequently uninformedand, hence, easily changeable. Publicreaction to the Lewinsky matter notonly proved these critics wrong, butalso demonstrated the impotence ofthe mass media. While the House

Republicans and even most of theSenate Republicans elected to ig-nore how a majority of the publicwanted the Lewinsky matter han-dled, in the end it was the public'sposition that prevailed.

The credibility of both the Repub-lican Party and the mass media ap-pear to have suffered as a conse-quence of the scandal. While it isnot possible to provide a definitiveassessment of the media's coverageof the scandal without a contentanalysis of that coverage, the surveydata suggests that the public wasquite critical of the role that the me-dia played. For the most part, thepublic perceived the media, espe-cially television, as promoting theremoval of the president. Often,television commentators appearedalmost gleeful at the possibility thatthe president might be removedfrom office. It appears that the pub-lic's negative reaction to the media'scoverage added to the stability ofpublic opinion over the course ofthe year-long ordeal. Given that thepublic came to question the credibil-ity and impartiality of the media, thepotential impact of the media's cov-erage was reduced.

The standing of the Republicanparty among the public also sufferedas a consequence of how they han-dled the affair. Whether the Repub-lican Party will be able to recoupthe respect of American voters re-mains to be seen. The Republicanresponse to the Lewinsky matterreinforced an already emerging im-age of the party as an organizationcontrolled by far right-wing zealots.In addition, the agenda that thepublic has come to associate withthe Republicans is more one of pro-moting a particular version of mo-rality rather than specific policiesthat deal with various areas such associal security, education, and healthcare.

Although the Clinton-Lewinskyaffair may have resulted in a weak-ened presidency (due to the Su-preme Court decision to allow a civilsuit to proceed against a sitting pres-ident), it has not undermined publicconfidence in government more gen-erally. Some may speculate that thisis probably because, in the end, theconstitutional process yielded anoutcome the public preferred. Thiscertainly is not the reason why theaffair had no impact on trust in gov-ernment, at least as of the end of1998. The explanation hinges farmore on the fact that most peoplesaw the investigation into the presi-dent's behavior and the impeach-ment as partisan and as an invasionof the president's right to privacy.Moreover, as previous research hasdemonstrated (see Miller and List-haug 1999), the degree of trust thatcitizens place in their government isa reflection of how the public judgesbroader, more profound aspects ofgovernment performance and notsimply how they feel about a partic-ular official, even the president.

Notes

* My thinking on the subject of this paperand on trust in government over nearly twodecades has been greatly influenced by mycollaboration with Ola Listhaug. I thank himfor collaboratively sharing this common sub-stantive interest over such a long period oftime. Brian McCuen and Karin Anderson de-serve thanks for their help with data analysisand the preparation of the graphs. PeggySwails, Katie Perciach, and Shanan Shaver-


Notz are thanked for their assistance in pre-paring the manuscript. The work on this pa-per is part of a broader project currentlyunderway with my colleague Bruce Gronbeck.This work is partially funded by a grant fromthe Obermann Center for Advanced Studiesat the University of Iowa.

1. On May 27, 1997, the Supreme Courtruled in favor of allowing Paula Jones' attor-neys to proceed with their civil lawsuit againstClinton. The suit arose out of an alleged sex-ual encounter between Clinton and Jones thatsupposedly occurred while Clinton was gover-nor of Arkansas. Because of this ruling, thePaula Jones case went forward. Evidence ofan affair between Clinton and Lewinsky wasmade public when Jones' lawyers tried todemonstrate a pattern of predatory sexualbehavior by Clinton.

2. The seven states include Illinois, Iowa,Minnesota, Missouri, Nebraska, South Da-kota, and Wisconsin, and the survey usually

involves 300 respondents from each state.Each state sample is collected in a fashionthat makes it self-representing. The states arethen properly weighted when the data arecombined to reflect the region. The survey isconducted by phone using random-digit dial-ing to select the phone sample and then theKish method is used for the within-householdrespondent selection. Roughly 2,100 respon-dents are interviewed for about 25 minutes ineach Heartland survey.

There is a sampling error of ±2.2%. Re-sponse rates for each survey fall in the 67-70% range. For more information on theHeartland Poll, please contact me at [email protected].

3. Ideological grouping of Republicans de-pends on individual respondent's self-place-ment on a 7-point scale that ranges from 1 =extremely liberal to 7 = extremely conserva-tive, with 4 = middle of the road. The distri-bution was trichotomized by combining 6 and

7 into a far right wing category, 5 into a mod-erate right wing category, and 1-4 into a mid-dle-of-the-road category. Only 4%' of Repub-licans placed themselves in the 1-3 range ofthe scale. For my purposes, 35% of Republi-cans claimed to be middle-of-the-road, 27%identified themselves as moderate right wing,and 38% claimed to be far right wing.

4. This regression, which is not presentedin a table, utilized the thermometer rating ofthe Republican Party as the dependent vari-able. The thermometer ranges from 0 mostnegative to 100 = most positive. The inde-pendent variables were the same as those in-eluded in Table 3. The amount of explainedvariance was 35%. The predictors having thelargest impact, other than party identification,were evaluations of the media coverage, per-ceptions of the affair as a public or privatematter, and assessments of the Starr investi-gation and House impeachment process aspartisan.

References

Citrin, Jack. 1974. "Comment: The PoliticalRelevance of Trust in Government." Ameri-can Political Science Review 68(Scptember):973-88., and Donald Green. 1986. "PresidentialLeadership and the Resurgence of Trustin Government." British Journal of PoliticalScience 16(June): 431-53.

Craig, Stephen C. 1993. The Malevolent Lead-ers: Popular Discontent in America. Boul-der: Westview.

Easton, David. 1965. A Systems Analysis ofPolitical Life. New York: Wiley.

Kinder, Donald. 1986. "Presidential CharacterRevisited." In Political Cognition, ed.Richard R. Lau and David O. Sears. Hills-dale, NJ: Lawerence Erlbaum Associates.

Miller, Arthur H. 1974. "Political Issues andTrust in Government: 1964-1970." Ameri-can Political Science Review 68(Septem-ber): 951-72.

, and Stephen Borrelli. 1991. "Confidencein Government during the 1980's." Ameri-can Politics Quarterly 19(April): 147-73.., and Ola Listhaug. 1999. "GovernmentPerformance and Political Trust." InGlobal Support for Democratic Govern-ment, ed. Pippa Norris. Oxford: OxfordUniversity Press.. 1993. "Ideology and Political Alien-ation." Scandinavian Political Studies16(March): 120-39.. 1998. "Policy Preferences and PoliticalDistrust: A Comparison of Norway,Sweden, and the United States." Scandi-navian Political Studies 21(March): 161-87.. 1999. "Political Parties and Confidencein Government: A Comparison of Norway,Sweden, and the United States." BritishJournal of Political Science 29(June): 357-86.

Miller, Arthur H., Martin P. Wattenberg, andOksana Malanchuk. 1986. "Schematic As-sessments of Presidential Candidates."American Political Science Review80(June): 521-40.

Nye, Joseph, Philip Zelikow, and David King,eds. 1997. Why People Don't Trust Govern-ment. Cambridge, MA: Harvard UniversityPress.

Ornstein, Norman. 1993. "Less Seems More:What to Do about Contemporary PoliticalCorruption." The Responsive Community4(July): 7-22.

ORS Publishing. "Polling the Nations 1986-1998." Data CD: 4550. Bethesda, MD:ORS Publishing.

Zaller, John. 1998. "Monica Lewinsky'sContribution to Political Science." PS Po-litical Science anil Politics 3l(.lune): 182—89.


sex, politics, and public opinion: what political scientists really … · sex, politics, and...

Documents