Rhetorical convergence and issue knowledge in the 2000 presidential election.

Waldman, Paul ; Jamieson, Kathleen Hall

Both of us use similar language to reach an exactly opposite outcome.

--Al Gore (presidential debate, October 17, 2000)

Citizens often say that all politicians sound the same. While occasionally used as an excuse for apathy, this complaint captures, at least in part, a critical characteristic of contemporary political discourse. This was seldom more true than in the presidential campaign of 2000. Although the candidates' actual policy positions differed substantially on dozens of issues, at times it seemed to many that they were making similar arguments, using similar terms to describe their policies, and in general converging toward a single set of goals. This convergence influenced what voters knew and thought about George W. Bush and Al Gore. It also represented a successful rhetorical strategy on Bush's part and a failure by the Gore campaign.

The subject of rhetorical ambiguity as a campaign strategy has been addressed before, with mixed results. While Downs (1957) argued that ambiguity was advantageous to candidates, a view supported by Page (1976), Shepsle (1972) attempted to demonstrate that theoretical conditions exist under which ambiguity results in fewer votes, a view supported by Alvarez (1998). Campbell (1983) showed that ambiguity is most effective when the candidate's position is at odds with majority opinion and when opinion is widely dispersed. Unsurprisingly, candidates have the greatest incentive to be ambiguous when specificity would alienate significant numbers of voters.

When candidates are vague, one must surmise their true positions. Often, voters simply "project" their own views onto their candidates of choice; instead of looking at their positions and deciding whom they like, they decide whom they like and then infer that the candidate's positions mirror their own. The stronger and more apparent a candidate's ideology, the less likely it is that voters will project their own views on to that candidate (Conover and Feldman 1989). Unlike some other information shortcuts, such as the use of party identification to infer issue positions, projection is not inherently likely to be accurate, although in practice the accuracy of projection is related to an individual's political sophistication (Krosnick 1990).

But ambiguity as explored by these scholars is not precisely the same as the rhetorical convergence that could be observed in the campaign of 2000. Bush and Gore did not simply decline to state where they stood or issue vague pronouncements potentially compatible with any policy position. Nor did they adopt a few selected positions associated with the other party to cloud perceptions of their general ideologies, a strategy Norpoth and Buchanan (1992) suggest may not be effective. Rather, they pledged support to a common set of ends--maintaining a strong military, providing prescription drug coverage, strengthening social security, and increasing accountability in education--but differed only on the means by which these ends could be achieved.

Nonetheless, the most common interpretation of the 2000 presidential election is not that of common issue positioning. Instead, the election is seen as one in which one candidate had an advantage on issues, while the other had an advantage on personality. Consequently, Bush's victory may be explained by, as The Baltimore Sun put it, "negative attitudes toward [Gore] personally, which seem to be negating his perceived advantage on issues" (West 2000, A1). Although there were many factors contributing to the election's final outcome, the premise of the explanation--that Al Gore had the advantage on issues--has not been challenged. In contrast, we will argue that Gore's "advantage" on issues was substantially mitigated by the public's lack of knowledge and misperception, to the point of being no advantage at all. First, many and in some cases most voters were not aware of where the candidates stood. Second, partly because Gore's positions were in fact more popular, when people made mistakes in attributing issue positions to the candidates, the results helped Bush and hurt Gore. The unnoticed factor in the 2000 election was thus not uninformed opinion but misinformed opinion. As others have noted, the two are conceptually distinct and have differing kinds of consequences, although both can be substantial (Dalager 1996; Gilens 2001; Kuklinski et al. 2000).

The rhetoric of the two candidates did not serve to reduce voter confusion. One explanation for candidates' rhetorical choices can be found in the spatial theory of voting originated by Downs (1957), which in its simplest form states that each voter selects the candidate closest to his or her favored policy positions, and therefore the rational decision for the candidate (depending on what the opponent does) is to place himself or herself at the median point of voter opinion, thereby maximizing the number of votes he or she will receive. Bill Clinton's "triangulation" strategy for the 1996 election, in which he positioned himself between his own party and the Republicans, is a near-perfect example of the spatial theory in action (King 1997).

Although the spatial theory and associated rational choice analyses have been extensively critiqued in recent years, particularly for their failure to make accurate empirical predictions (see Green and Shapiro 1994), in 2000 Al Gore and George W. Bush appeared to conduct themselves--rhetorically if not in policy--as though they were following the prescriptions of the spatial theory. In many cases, they placed themselves on the popular side of the issue in question; in others, they cloaked discussion of unpopular positions in terms that might have convinced many paying less than perfect attention that their positions were something they were not. This strategy was particularly attractive to George W. Bush, who appeared to implement it successfully.

Rhetorical choices make it possible for candidates to cloud understanding of unpopular positions through the use of what Jacobs and Shapiro (2000) call "crafted talk." They argue that a key element of policy making is the effort by political actors to move public opinion by attaching popular words and symbols to policies that may begin with little popular support, resulting in a politics that appears far more responsive to public opinion than it actually is. Although Jacobs and Shapiro are concerned more with policy debates than with the positioning of candidates during campaigns, their analysis goes a substantial way toward explaining both the campaign rhetoric of 2000 and the shape of voter awareness of Bush and Gore.

Most political professionals believe that each presidential campaign revolves around no more than three issues. If one candidate can shift the agenda to issues on which his position or record is more popular than that of his opponent, he stands a good chance of winning the election. Changing people's minds on an issue is difficult, as political actors are well aware (Jacobs and Shapiro 2000); it is far easier to shift public attention and prime a certain set of issues (Iyengar and Kinder 1987). Campaigns thus attempt to heighten the salience of some issues and reduce that of others. For example, the famous 1992 Clinton campaign mantra, "It's the economy, stupid," was an instruction for the candidate and his aides to keep the focus on an area in which voters felt he would do a better job than his opponent. The Bush campaign in turn attempted to heighten the salience of candidate character, on which they felt they had an advantage.

Although candidates spend much of their time competing to set the agenda, many citizens believe that politicians adopt positions that they believe are popular but that do not reflect conviction. Bald dishonesty of this sort is actually uncommon. More often, candidates simply shift attention or change the terms they use to describe their positions. The complicating factor is that because they poll relentlessly, both campaigns know the public's priorities. As Jacobs and Shapiro (2000) put it, "The irony of contemporary politics is that politicians both slavishly track public opinion and, contrary to the myth of `pandering,' studiously avoid simply conforming policy to what the public wants" (pp. 7-8).

In 2000, as in years past, the candidates denied that polls mattered to them. When asked about the use of polls, Gore said that "I don't think they have all that much meaning." Bush concurred, "I don't base my decisions on what polls and focus groups say" (NBC Nightly News, October 6, 2000). Of course, this professed lack of faith in polls did not stop either candidate from mounting an extensive polling operation.

In 2000, to a large extent the campaigns did not compete to determine the issue agenda but instead simply took their cues from what their polls told them were the issues of greatest concern to the public. In the 2000 National Annenberg Election Survey (NAES) (1)--the study on which the data in this article are based--as well as those conducted by other organizations such as Gallup, education consistently topped the list of issues Americans found most pressing. While social security and health care ranked somewhat lower, they were of intense concern to the age group with the highest voter turnout: senior citizens. These issues became the primary focus of both candidates.

A look at the candidates' advertising shows the extent to which they had nearly identical agendas. To analyze these agendas, we divided each ad into a series of claims and then recorded the issue being discussed in each. Education, prescription drugs for seniors, and social security made up three of the top four issues for each candidate. These three accounted for more than 40 percent of the claims in Gore's ads and more than 50 percent in Bush's.

Had the candidates offered clear distinctions between their goals and plans on these issues, the rhetorical convergence in their advertising might have been beneficial to voters. Unfortunately, most of the ads were not specific enough to allow viewers to make comparative judgments. Witness the following two ads about education. If one simply reads the transcripts, it is difficult to tell which candidate is speaking:

Candidate A: Strengthening education begins with us, not with government. Parents simply have to get involved and take responsibility to make sure that their children study and learn. But government has to take responsibility for what it can do. Fix failing schools, reduce the class size, set high standards for students and for teachers. We are in an information age. We have a responsibility to make sure that our schools are the very best in the entire world.

Candidate B: Our parents and our schools and our citizens ought to have high expectations for every child.

Announcer: [Candidate B]'s education plan is hailed as remarkable for its bold conservatism.

Candidate B: The best way to have the best school system is to have programs emanate from the local level.

Announcer: [Candidate B]'s plan calls for teaching character in the classroom. His plan holds schools accountable, blows the whistle on failure, gives parents more choices, and focuses on teaching all children to read.

Candidate B: Every parent wants their child to get the very best.

Candidate A asserts that strengthening education does not begin with government, while Candidate B wants programs that emanate from the local level. Candidate A wants high standards, while Candidate B favors accountability. Candidate A wants to fix failing schools, while Candidate B wants to blow the whistle on failure. Although the use of the phrase "bold conservatism" should tip off the reader that the first candidate is Al Gore and the second George Bush, it is difficult for a viewer to see precisely what one would do differently than the other.

One might view the fact that the candidates settled on a common agenda as a positive development: while they might have advanced their own agendas, the candidates addressed the public agenda. If they had been able to do so in a way that allowed most voters to understand the distinctions between them, the case for that conclusion might be compelling. But by November 7, many voters remained unsure of precisely how George Bush's education or Medicare plans differed from Al Gore's. Although Bush favored school vouchers, the word hardly ever passed his lips, perhaps because the word may be less popular than the idea. Nearly half of those we interviewed immediately after the campaign were unaware that Bush supported vouchers. The previous Bush education ad says that he favors "giving parents more choices," but this could forecast any number of policies, some looked on more favorably by voters than others. Indeed, it could mean support for choice within the public school system, a position favored by Gore.

In the final debate, moderator Jim Lehrer asked Bush directly, "Governor, what is your position on vouchers?" Bush responded,

 First of all, vouchers are up to states. If you want to do a
 voucher program in Missouri, fine. See, I strongly believe in
 local control of schools.... When we find children trapped in
 schools that will not change and will not teach, instead of saying,
 "Oh, this is okay in America--just to shuffle poor kids through
 schools"--there has to be a consequence. And the consequence is
 that federal portion of federal money will go to the parent so the
 parent can go to a tutoring program or another public school or
 another private--or a private school. (October 17, 2000)

Bush says that "vouchers are up to states" and then describes a federal voucher program he would support. This was also one of the only times during the campaign that Bush mentioned his support of vouchers for private schools. His Web site articulated the position this way:

 Failing schools will be given a finite period to change; if they
 fail to do so, children of low-income parents will have the option
 of transferring to another public school or using their share of
 federal funds to pay for another option (tutoring, charter school,
 etc.).

The "etc." refers to private schools, which lie at the heart of the voucher controversy. (2)

Senior Citizens: Target 1

Although senior citizens are often courted in campaigns, Bush and Gore pleaded for their votes to an unusual degree for a presidential campaign. In their advertising, one-fifth of the claims made by Bush and one-quarter of the claims made by Gore concerned prescription drug coverage for seniors, Medicare, or social security. One Bush ad that excerpted his convention speech could easily have been offered by Hubert Humphrey or Walter Mondale:

Bush: We will strengthen social security and Medicare for the greatest generation and for generations to come. I believe great decisions are made with care, made with conviction. We will make prescription drugs available and affordable for every senior who needs them. You earned your benefits. You made your plans. And President George W. Bush will keep the promise of social security, no changes, no reductions, no way.

Health care was an area in which the candidates' differences were particularly hard to discern. For instance, one of Bush's ads stated that "while Washington deadlocked, he delivered a patient's bill of rights that's a model for America." Although it took credit for a bill Bush had actually opposed, he made the claim in the first debate as well. Sounding a similar note, Gore said he was "leading the fight for a patient's bill of rights." Whereas Gore in his ads promised to "give control back to patients and doctors, not HMO bureaucrats," Bush accused Gore of forcing seniors into a "government HMO" (i.e., Medicare). Both said they favored a patient's bill of rights; both criticized HMOs. The fact that their plans differed radically was virtually impossible to discern from their advertising. By the race's end, the voters had a slightly better understanding of Gore's position than Bush's: about half of those we interviewed knew that Gore favored a Medicare benefit, while around a third correctly identified Bush's position. Only slightly more than a quarter of Americans correctly identified both positions, meaning that three-quarters were somewhat confused about the candidates' prescription drug proposals. A Pew Research Center (2000) poll in October showed that 59 percent of the public thought there were important differences between Bush's and Gore's prescription drug plans. According to the NAES, however, most people did not quite know what those differences were.

Although Bush may have engaged in the greater amount of crafted talk, at times Al Gore mimicked the Texas governor's language as well. For instance, anyone reading the following excerpt from an October 24 speech would conclude that it must have been Bush talking:

 I'm opposed to big government.... I'm for a smaller, smarter
 government, one that serves people better, but offers real change
 and gives more choices to our families.... I don't believe there's
 a government solution to every problem. I don't believe any
 government program can replace the responsibility of parents, the
 hard work of families or the innovation of industry.

In fact, the speaker was Al Gore.

Since Democrats tend to be more supportive of government and Republicans less so, the language they use to advocate their candidacies usually differs. One analysis of a half-century's worth of presidential campaign speeches showed that as a general rule, Democrats talk about policy while Republicans talk about principles (Jarvis 1999). Like many of his Republican predecessors, George W. Bush spent a good deal of time criticizing the federal government in general terms, despite his plans for expanded spending on a number of programs. As the above example illustrates, when the Gore campaign surmised that Bush's mantra, "He trusts government, I trust the people," was beginning to stick, Gore distanced himself from "big government" without dropping any of his proposals for government spending.

As the campaign neared its end, the question of who was more opposed to government became more and more muddled. In their first debate, Gore adopted Bush's formulation on the issue of abortion: "He trusts the government to order a woman to do what it thinks she ought to do. I trust women to make the decisions that affect their lives, their destinies, and their bodies." In their final debate, Gore said, "During the last five years, Texas' government has gone up in size. The federal government has gone down. Texas' government has gone up. Now, my plan for the future, I see a time when we have smaller, smarter government." Gore's general philosophy about the size and scope of government was roughly the same as most in his party, but his rhetoric might have led some to believe that he was in favor of substantial reductions in the role of government. The voter was left with two competing claims: Bush said Gore loves government; Gore said Bush loves government. Instead of a debate on the appropriate size of government, voters were offered competing accusations. Each candidate claimed that the other was insufficiently devoted to a goal that was lauded by both.

Despite its paradoxical (some might say hypocritical) character, the stance that both candidates adopted--being against government but for government programs--mirrored the contradiction in American public opinion. For many years, surveys have found that if asked whether government should be bigger or smaller, large majorities will respond "smaller." But when asked about specific programs, similarly large majorities will say that the government should spend more money, whether on the environment, the military, health care, or any of a wide array of programs. The exception is programs that offer benefits to poor Americans, where majorities generally say the government should spend less (Cantril and Cantril 1999). Consequently, railing against big government while offering to spend more government money is an effective strategy. Although Bush spent more time criticizing Washington and Gore spent more time talking about government programs, both candidates embraced the contradiction to a certain extent.

This is not to say that the candidates never added specific details to their plans-they did, in position papers and speeches. But the specifics never reached the majority of American voters. When a candidate gives a thirty-minute speech, complete with policy details, reporters select one or two quotations to feature in their stories. The quotations selected are seldom those that provide the most lucid explanation of a policy position. Instead, they are likely to be the harshest attacks the candidate offered that day, regardless of whether criticism of his opponent composed 99 percent or 1 percent of what he had to say. Our research on the past twenty years of elections has shown that the news consistently overrepresents the amount of attack present in candidate discourse, whether the campaign is predominantly made up of attack, contrast, or advocacy. In addition, the dominant motif of strategy or "horse race" news involves the truncation of issue argumentation in ways that make issue distinctions more difficult for voters to discern (Jamieson and Waldman 1997).

Even when a reporter makes a good faith effort to find quotations that accurately represent a candidate's speech, he or she will select statements that synopsize the candidate's arguments. In 2000, the candidates took many identical positions on the broad goals of policy. The distinctions, though substantial, lay in the details. Synopsizing quotations tend to describe the broader goals. When the details failed to pass through the news filter, voters learned only that both candidates wanted to protect social security, offer prescription drug coverage to seniors, or support a patient's bill of rights. The means by which they intended to accomplish these goals differed radically--but it was difficult for voters to understand how.

As a consequence, despite the closeness of the 2000 race, the distinctions were less clear than they had been in previous elections. In 1996, we conducted a survey in which people were asked about Bill Clinton's and Bob Dole's positions on various issues. A comparison of those data with data from 2000 shows that despite the common belief that 1996 was a "boring" election to which voters paid little attention, distinctions between Clinton and Dole on many issues were substantially clearer than distinctions between Gore and Bush.

For instance, on the issue of abortion, immediately following the 1996 debates fully 80 percent of those we surveyed accurately judged that Bob Dole favored making it harder for women to obtain abortions, while less than 14 percent thought that Bill Clinton favored such restrictions. At a similar time in 2000, only 64.1 percent knew that George W. Bush favored restricting abortion, while 25.5 percent erroneously thought that Al Gore agreed. (3)

One area in which the distinctions were clearer was the death penalty. While a similarly small number were aware of Clinton and Gore's support for the death penalty-36 percent and 41.5 percent, respectively-many more were aware of Bush's support in 2000 (78.3 percent) than Dole's in 1996 (53 percent). Bush, of course, is a strong advocate of the death penalty whose state executed more prisoners than under any governor since the death penalty was reinstated in 1976. During the campaign, many news stories were written about the case of Gary Graham, an inmate for whom Bush refused to grant a stay of execution despite questions about his guilt.

When exploring the effects of candidate rhetoric on public impressions, one must keep in mind that there is not a single entity called "the public" whose members receive an identical body of information. A small number of people are familiar with the entire scope of the candidates' records and issue positions, but most pick up their information in drips and drabs as the campaign progresses. For many, the simplified messages of ads and sound bites are the only exposure they get. In addition, many of the issue positions that consistently distinguish the two major parties--and that elites take for granted--are not understood by a significant proportion of the electorate. Even the relatively simple question asked by the National Election Studies (NES), "Would you say that either one of the parties is more conservative than the other at the national level? Which party is more conservative?" has never been answered correctly by more than 60 percent of respondents. Ideally, presidential campaigns should provide a quadrennial instruction in these important issue distinctions.

Most voters are exposed to the visible rhetoric--not the policy papers on the candidates' Web sites or the text of their economic plans but the content of their ads, what they say in debates, and the mantras they repeat on the stump. In 2000, this rhetoric made it particularly difficult to distinguish between the candidates.

Since 1952, the NES have asked voters, "Do you think there are any important differences in what the Republicans and Democrats stand for?" Prior to 2000, the years in which the most people, 63 percent, answered in the affirmative were 1984 and 1996, elections in which a strong economy propelled an incumbent to a relatively easy victory. The number seeing a difference between the parties has been about 60 percent in presidential election years but falls below 50 percent in off-year congressional elections, indicating that for a significant number of people, a perception of party differences is not a consistent part of their understanding of politics but is rather something that is clarified to a greater or lesser extent by presidential campaigns. In 2000, the number of NES respondents answering that there were important differences between Democrats and Republicans reached a new high of 70 percent. Similarly, polling by the Pew Research Center (2000) found that 61 percent of voters felt that "George W. Bush and Al Gore take different positions on the issues." In an era in which partisan conflict in Washington is both a reality and a consistent storyline of news--and in the midst of a hard-fought presidential campaign--the fact that most people would perceive a difference between the parties is unsurprising. Less encouraging is the degree to which they were able to discern the content of that difference.

The Gore campaign's pollsters asked a related question about the candidates themselves. They asked a national sample the following question: "Both Al Gore and George W. Bush proposed plans on education, a patient's bill of rights, and prescription drugs for seniors. On these, which [plans are better]?" Exactly half of the respondents chose that either Gore's plans were better or Bush's were better. The other half could not decide between them, with 21 percent of the total sample admitting that "it is too confusing to figure out whose proposals are better" (Jamieson and Waldman 2001).

At any given point in the campaign, a similar number to that found in Gore's polling--between 15 percent and 20 percent of the people the NAES interviewed--placed the candidates at the same spot on an ideological scale. About half of those rated both candidates as "moderate," while a third of them rated both as "conservative." Seeing candidates as the same has a strong negative relationship with other variables associated with political interest and knowledge. For instance, as the campaign came to a close, 23.5 percent of our respondents who did not finish high school placed both candidates in the same ideological position, compared to just 9.5 percent of college graduates.

During the campaign, George W. Bush spent time and effort addressing issues that in the past have been "owned" by the Democrats, with two results. The first was that given the lack of depth with which most issues were discussed, the candidates looked more alike--both wanted to improve education, protect social security, and preserve Medicare. The second effect was that Bush blunted what might have been some of Gore's strongest advantages on issues. Polling done by The New York Times and CBS late in the campaign showed that while more people trusted Gore to do a better job than Bush on education and social security, the gap was extremely small. Both campaigns were well aware of the progress Bush had made on these issues. During the Annenberg postelection debriefing, Gore pollster Stan Greenberg reported the struggles the Gore campaign had:

 On questions of social insurance, like HMO reform, prescription
 drugs, we won those issues in the end, though we had to battle
 them. We paid a high price for that, but not on the issue itself.
 We beat George Bush on prescription drugs and on Social Security,
 but Bush looked like he wanted to do something. He didn't look like
 a man with bad intentions. It also put us on seniors' issues for a
 much longer period than we wanted to be.... On education, we won by
 four, which was clearly unacceptable as a Democrat. (Jamieson and
 Waldman 2001, 95-96)

Bush advertising consultant Mark McKinnon agreed:

 When you looked at the issue matrix on this election, all the
 issues that people cared about were typically Democratic issues:
 education, Social Security, health care. So, we knew that while we
 probably couldn't win on those issues, we had to at least keep them
 close. Fortunately we had a candidate who had been talking about
 those issues, not just in this campaign, but for years as governor
 in Texas. So there was a platform there, and a history. Our
 strategy was to stay close on those issues. (Jamieson and Waldman
 2001, 145-46)

The result was an election in which most voters perceived a difference between the candidates and the parties they represented, but when it came to issues, they were not sure exactly what those differences entailed. Whereas in previous elections each candidate carved out an area of specificity on his strongest issues and remained vague on those where his opponent was strong, in 2000 both candidates focused on the same agenda. When distinctions on the issues on that agenda were blurred, means of contrast became harder for voters to find.

Lack of Knowledge Benefited Bush More than Gore

Many postelection journalistic analyses concluded that while Gore had an advantage on the issues, voters found Bush to be the more appealing personality. Although our data support the second half of this conclusion, they also show that Gore's advantage on issues was substantially undercut by voter misperception. Despite the fact that Gore's positions on a variety of issues were more popular than Bush's, many voters were not aware that they agreed with Gore.

The NAES asked respondents to indicate where the candidates stood on a variety of policies not only to test their knowledge of the candidates but also to discover how closely their perceptions matched reality. By combining the answers to these questions with those we asked about the respondents' own beliefs, we can assess not only how many voters agreed with each candidate's actual positions but how many voters agreed with what they believed to be each candidate's positions. When we do so, we discover not only that many people misunderstood where the candidates stood but that in most cases this misperception worked to Bush's benefit and Gore's detriment.

Figures 1 and 2 show first the percentage of voters who placed themselves and the candidate on the same side of the issue in question, regardless of whether they were correct about the candidate; this is termed perceived agreement. Second, they display the percentage whose opinion matched the actual position that the candidate took during the campaign. To focus on what voters had learned by the end of the campaign, these figures show only data from the final stage of the campaign, after the last debate on October 17.

[FIGURES 1-2 OMITTED]

On five of the nine issues, the percentage of voters who believed they agreed with Bush exceeded the percentage who actually agreed with him. The discrepancy was more than thirty points on the question of requiring a license to purchase a handgun and more than twenty-five points on the issue of providing health insurance to all children. The only issues where substantially more people actually agreed with Bush than perceived they agreed with him were two of his more popular positions, support for the death penalty and allowing workers to invest a portion of their social security contributions in the stock market. In both cases, however, the discrepancy did not exceed fifteen percentage points.

For Gore, on the other hand, the picture is far different. On all nine issues, more people actually agreed with Gore than knew or believed they did. (4) In a few cases, the differences are substantial. While 73 percent of voters shared Gore's support for the death penalty, only 38 percent thought they agreed with him (and some of these were death penalty opponents who thought Gore shared their view). Only 43 percent of voters thought they agreed with Gore on the question of whether citizens should have the ability to sue their HMOs, although 75 percent did in fact agree with him. On these nine issues, perceived agreement with Gore lagged behind actual agreement by an average of seventeen percentage points.

The net result was that Gore's "advantage" on the issues, if understood not as the popularity of his positions but as the proportion of voters who believed they agreed with him, was virtually nonexistent. On these nine issues, the popularity of Gore's position exceeded the popularity of Bush's by an average of twenty-four points (61 percent to 37 percent) during this final period of the campaign. But on perceived agreement, the candidates were almost equal, with an average of 42 percent of voters believing they agreed with Bush on any given issue, compared to 44 percent who believed they agreed with Gore. While in many of these cases an alternate question wording might have generated slightly different results, the overall pattern is unmistakable. The accepted interpretation of the campaign--that Gore led on the issues but Bush was better liked personally--turns out to be false. Although it was true that voters found Bush's personality more appealing, Gore did not hold any appreciable advantage on issues. These data demonstrate that the story of political knowledge in the 2000 election was not so much about uninformed opinion but about misinformed opinion.

The question is not so much why Americans did not have a better understanding of Gore's positions but why they did not have a better understanding of both candidates' positions. There were, of course, some issues that were better understood than others. For instance, 77 percent knew that Bush favored the death penalty. As we noted, the application of the death penalty in Texas was the topic of many news stories and late-night jokes during the campaign. In contrast, only 37 percent knew where Bush stood on gun licenses, which was not a frequent topic of discussion. Overall, the number of respondents who answered that they did not know where Gore stood was identical to the number who said the same about Bush: an average of 23.5 percent responded that they did not know where Gore stood on a given issue, compared to 23.4 percent for Bush. Furthermore, as many people gave the wrong response when citing Bush's position (25.4 percent) as they did when citing Gore's (23.5 percent). But when these errors were made, they were more likely to result in an increase in perceived agreement for Bush and a decrease in perceived agreement for Gore (see Figure 3).

[FIGURE 3 OMITTED]

To assess the informational environment that could have produced such errors, we performed a content analysis of every story about the presidential campaign that appeared on the three broadcast network news shows between July 1, 2000, and election day. Each story was divided into a series of statements by or about the candidates, and among other variables, we recorded whether the statement mentioned any issue. Fewer than a third of the statements in television news mentioned an issue at all. Included in this figure are any mentions of an issue, no matter how brief. Although the networks did air occasional in-depth explorations of campaign issues, these stories made up a small portion of the total news about the campaign. More typical was coverage similar to the following, from NBC Nightly News on October 9:

Tom Brokaw: NBC News in-depth tonight, presidential politics. And with the election now just a month away, the Gore campaign is going on the offensive, perhaps because of several days of attacks by the Bush camp. The latest polls show this race tightening to a dead heat. The MSNBC/Reuters daily tracking poll had Gore with a five-point lead over George W. Bush less than a week ago. Today Gore's lead is down to one point now. That's well inside the margin of error.

For the Gore campaign, there's nothing fuzzy about this kind of math. NBC's Claire Shipman begins tonight's in-depth reporting with the Gore campaign's new emphasis on what it calls Governor Bush's mistakes and missteps.

Claire Shipman: Al Gore picking up some light reading material during a break from his debate preparations. But while he stays sunny and above the fray, his campaign unleashes a torrent of negative firepower against George W. Bush today, pointing up his campaign bloopers, as Gore staffers call them, and attacking his record as Texas governor on everything from health care to the environment.

Joe Andrew (Democratic National Committee [DNC]): He says he wants to be judged by his own record; we're going to make sure he's judged by his own record.

Shipman: The Democratic National Committee says Joe Lieberman will head to Texas after Wednesday's debate for a, quote, "failed leadership tour," and it's spending several million dollars on three biting new ads in battleground states.

Excerpt from Gore ad: Now take a deep breath and imagine Seattle with Bush's Texas-style environmental regulation. George Bush: before he talks about cleaning up Washington, maybe he should clean up Texas.

Shipman: Why the sudden assault when the public says in poll after poll it's not in the mood for negativity? Because the Gore team believes it has to counter Bush's relentless attacks on the vice president's credibility, which it fears are working.

George W. Bush: The voters are going to have to draw their own conclusion about somebody who seems to have a consistent pattern of exaggerating.

Shipman: And the same message on the Bush Web site, "The Gore Files: Anything to Get Elected." The DNC counters with "Bush Light: Less Leadership, Less Experience, More Right-Wing Flavor." Will any of this work?

Larry Sabato (University of Virginia): Negative campaigning almost always works to some degree. It's bound to add a point or two to the side that's most--most effective at it.

Shipman: The Gore team so far is trying to have it both ways. Candidate Gore is staying positive, while his running mate, his aides, and his ads do the dirty work. And Bush doing some of the jabbing himself clearly hopes his attacks will hurt his opponent more than they hurt him. Claire Shipman, NBC News, Sarasota.

Political discourse often involves inductive argument, moving from specific facts to a general conclusion. In this case, the Gore campaign cites specific environmental problems in Texas, argues that they constitute a failure on George Bush's part, then concludes that he is therefore unprepared to assume the presidency. However, the NBC News report ignores the specifics cited by both sides to focus on the fact that they are attacking each other and to explain the strategy behind the attacks. We learn where the attacks are taking place (Web sites, advertisements), and a political science professor explains the efficacy of attacking. But a citizen looking for information to assist him or her in casting her vote would have trouble finding anything useful in this report. One study comparing network news coverage of the past four presidential elections found that 71 percent of stories in 2000 were primarily concerned with the "horse race" as opposed to issues, compared to 48 percent in 1996 and 58 percent in both 1992 and 1988 (Lichter 2001).

When an issue is relatively uncomplicated and the positions of the candidates unambiguous, only a minimum of explanation and repetition on the part of the news is necessary for most of the electorate to appreciate the differences between the candidates. Gore was pro-choice, and Bush was pro-life; even though abortion was not a topic of frequent discussion, each candidate's position was understood by approximately two-thirds of voters. But when their proposals are complex, the differences subtle, the details arcane, and the candidates pledging support to the same goals, news has an extra responsibility to confront the issue in depth and repeatedly. In 2000, the news--particularly television news--largely failed to do so.

For issue positions to be recalled by voters, they have to be repeated by the candidates and the news. Just as surveys such as ours show that most voters do not know most of a candidate's issue positions at the end of the campaign, experiments have demonstrated that voters quickly forget issue information unless they are exposed to it repeatedly (Lodge, Steenbergen, and Brau 1995). The reluctance of journalists to repeat information they have already reported--in other words, the need for the news to be new--hampers the ability of voters to make choices based on issues. Whether Gore would have won the election had he been more aggressive in highlighting the issue positions on which he aligned with a majority of the public is difficult to say, but his rhetoric does not appear to have been successful.

Ideological Perceptions

Ideology provides clues to how the candidate will act in the future and what positions he will take on issues that have yet to emerge. As a consequence, ideological assessments provide a window into the more general conclusions people make about candidates. We asked our survey respondents to rate the candidates on an ideological scale running from very liberal to very conservative. The results indicate that many people--particularly Republicans--viewed the candidates through their own ideological lenses. While Democrats and Republicans both placed Bush at nearly the same point on the ideological scale (with Republicans actually rating him as slightly more conservative), there were stark differences in how people rated Gore. Forty percent of Democrats thought Gore was moderate, while only 6 percent described him as very liberal. Republicans, on the other hand, viewed Gore as much farther to the left. Only 18 percent of Republicans rated him as moderate, while 32 percent described him as very liberal.

For Bush, the differences were much slighter. Seventeen percent of Democrats rated Bush very conservative, compared to 8 percent of Republicans. This means that the ratio of Democrats to Republicans placing Bush at the extreme of the ideological scale was just more than 2:1, while the ratio of Republicans to Democrats situating Gore at the extreme was 5.5:1. Among the most intense partisans--those describing themselves as strong Democrats or strong Republicans--the differences were clear as well. Twenty-one percent of strong Democrats described Bush as very conservative, while twice as many strong Republicans (43 percent) described Gore as very liberal. In sum, Democrats viewed Bush in much less extreme terms than Republicans viewed Gore (see Figure 4).

[FIGURE 4 OMITTED]

This is not a phenomenon unique to the 2000 election. In most presidential campaigns dating back to 1972 (the first for which NES data on this question are available), voters across the political spectrum have viewed the Republican candidate in roughly the same ideological terms. The sole exception is Ronald Reagan in 1984, whom Republicans actually saw as substantially more conservative than other voters did. In every other case, it is when they evaluate the Democratic candidate that partisans diverge. Republicans always perceive the Democratic candidate as much more liberal than Democrats and independents perceive him to be. Bill Clinton is the clearest case: while Democrats and independents placed him at about the same ideological position as most other Democratic candidates, in 1996 strong Republicans thought Bill Clinton was more liberal than previous strong Republicans had found Michael Dukakis, Walter Mondale, and even George McGovern. In 2000, Bush voters were slightly more enthusiastic about their candidate than Gore voters were about theirs, but Bush voters' views of Gore were substantially less favorable than Gore voters' views of Bush. Brady and Sniderman (1985) observe that "liberals do not like conservatives; however, they do not dislike them nearly as much as conservatives dislike liberals" (p. 1066). Bill Clinton inspired unusual anger and dislike among conservatives; for many of them, "very liberal" may have been synonymous with "immoral." It is also true that conservatives perceive liberals generally--not just Democratic politicians--to be farther to the left on issues than they in fact are (Sniderman, Brody, and Tetlock 1991).

When the principal contenders converge on a virtually indistinguishable rhetoric that masks the differences in their policy agendas, there can be negative consequences for governing, as George W. Bush discovered in the first few months of his presidency. In the wake of criticism over some of his cabinet nominees, Bush offered this defense two days before his inauguration in an interview with Brit Hume on Fox News:

 The reason I sit here able to talk to you as a person soon to be
 sworn as president is because of the positions I took in the
 campaign. I was basically running against an incumbent. He had a
 pretty good economy going for him. And the world was kind of at
 peace. And I wish I could say it was my charming personality or the
 ability to string a couple of sentences together. The truth of the
 matter is that I am sitting here because I took firm positions on
 important issues and didn't back off. And I am not backing off the
 minute I arrive in Washington. Quite the contrary. I'm going to
 take those issues I campaigned on and campaign hard for their
 enactment because I believe it's the right thing for the country.
 (Fox News, January 18, 2001)

Despite Bush's protestations, citizens were largely unclear about his policy agenda. Nonetheless, there is little indication they translated that confusion into a more comprehensive judgment about Bush's ideology. If Bush's rhetoric had been entirely successful, we would expect that voters would have perceived him as more moderate than they had perceived past Republican presidential candidates. In fact, this was not the case. When we use NES data to compare ideological ratings of Bush to those of previous Republican candidates (the NES has asked this question since 1972), we see that Bush appears to have been perceived quite similarly to those his party had nominated in the past. The variance between the mean ratings is extremely small, ranging from a low of 4.8 (Nixon 1972) to a high of 5.3 (Reagan 1984) on a 7-point scale, with higher numbers being more conservative. Bush falls in the middle of his Republican counterparts, with a mean rating of 5.0, equivalent to a rating of slightly conservative.

Although the candidates' rhetoric did not translate into visible differences in the assessments voters made of their overall ideologies compared to that of previous candidates, and despite the similarity between Bush's positions and those of previous Republican nominees, the idea that Bush was a "different kind of Republican" became a major theme of campaign coverage. Early in the campaign, a study by the Project for Excellence in Journalism (2000) showed that fully 40 percent of comments about Bush in the press concerned the idea that he was "a different kind of Republican." This image of Bush resulted not from any difference in the issue positions he adopted but from his issue agenda decisions and stylistic differences in his campaign, including frequent appearances with members of minority groups. The Republican convention in Philadelphia featured so many African-American speakers that one conservative commentator compared the gathering to a Utah Jazz basketball game: an all-white audience in the stands watching black performers (Brooks 2000).

Conclusion

Does rhetorical convergence matter? If presidential campaigns were merely sound and fury, signifying nothing, then the answer would be no. But in fact, campaigns serve a number of vital functions. As Hart (2000) put it, "Every four years, the American nation reconstitutes itself, thereby giving its citizens an opportunity to reflect on who they are and what they want to become" (p. 8). Campaigns do more than choose leaders; each is a vehicle for the citizenry to assess the past and set an agenda for the future.

Although citizens may not have made a summary judgment about Bush's ideology, some reporters did. When Bush began to govern, he received criticism in some quarters for implementing a conservative policy agenda that seemed at odds with the rhetoric of the campaign. "The reviews are in," said Tucker Carlson on CNN's Crossfire one month after Bush's inauguration. "George W. Bush is more conservative than expected" (February 20, 2001). The problem was not that Bush was implementing an agenda that differed from the positions he took during the campaign but rather that his agenda was at odds with the rhetoric through which it was presented.

During a campaign, a candidate attempts to reduce a complicated policy position to a single message (perhaps one that has a few different elements). But when policies are implemented, citizens can see effects that may or may not correlate precisely with the rhetoric used to justify them. While in a campaign rhetoric usually overwhelms policy details, in the process of legislating and governing, there is time for details to receive more examination. When a CBS/New York Times poll conducted in mid-July 2001 asked respondents if they believed Bush was a "different kind of Republican" or a "typical Republican," 63 percent answered that he was a typical Republican (Berke and Elder 2001).

These data suggest some of the difficulty presidents can encounter when they move to implement their agendas. While vagueness often has strategic value in the context of a campaign, it presents a particular problem for governance. The further a candidate is from the center of public opinion on a given issue, the greater his incentive to be vague; if a candidate's position is popular, he will feel free to discuss it in detail. It is precisely on issues where the candidate's positions are unpopular--and therefore where he has been vague during the campaign--that he will encounter the greatest displeasure with his policy when he attempts to implement it. Rhetorical convergence is not problematic if both candidates mean what they say and genuinely intend to pursue similar policies. But this is rarely the case in a presidential election, not because the Democrat and Republican do not agree on anything but because both the candidates and the press tend to ignore the issues on which they do agree. Because the vast majority of voters do not delve into the details of policy but may remember the broad arguments a president made when he was a candidate, they are likely to punish a candidate if his policies appear to diverge from those they were led to expect.

(1.) The National Annenberg Election Survey employed a rolling cross-section methodology, in which a new set of respondents was interviewed every day for the duration of the presidential campaign, from late 1999 through early 2001. The data presented in this article concern awareness of candidate issue positions, on which there was only slow movement over the course of the campaign. To focus on what voters knew by the campaign's end, the data presented cover only the period from after the final Bush-Gore debate (October 18) to the day before election day (November 6). This sample contains 6,191 cases. Although the sample adheres well to actual demographic breakdowns in most cases, one variable in which it does not is education; as in most telephone surveys, the National Annenberg Election Survey overrepresents those with high education and underrepresents those with low education. For this reason, and because the article presents simple percentages of respondents, the data are weighted by number of adults and phone lines in household, age, race, Hispanicity, education, and region. Readers should note, however, that the results obtained using unweighted data vary from the presented results by no more than four percentage points in any case.

(2.) At the insistence of Democrats in Congress, the voucher provision was dropped from the education bill that eventually passed during Bush's first months in office.

(3.) These data are taken from the week following the final debate in each election: 1,028 interviews conducted October 17 to 23, 1996, and 997 interviews conducted October 18 to 24, 2000.

(4.) In Gore's case, all of the differences are statistically significant at p < .01. The same is true of Bush's scores in all cases except gays in the military; the difference there is significant at p < .10.

References

Alvarez, R. Michael. 1998. Information and elections. Ann Arbor: University of Michigan Press.

Berke, Richard, and Janet Elder. 2001. Bush loses favor despite tax cut and overseas trip. The New York Times, June 21, A1.

Brady, Henry, and Paul Sniderman. 1985. Attitude attribution: A group basis for political reasoning. American Political Science Review 79:1061-78.

Brooks, David. 2000. Report from Philly: Why the GOP is playing nice. Newsweek, August 3, 10.

Campbell, James. 1983. Ambiguity in the issue positions of presidential candidates: A causal analysis. American Journal of Political Science 27:284-93.

Cantril, Albert, and Susan Davis Cantril. 1999. Reading mixed signals: Ambivalence in American public opinion about government. Washington, DC: Woodrow Wilson Center Press.

Conover, Pamela Johnston, and Stanley Feldman. 1989. Candidate perception in an ambiguous world: Campaigns, cues, and inference processes. American Journal of Political Science 33:912-40.

Dalager, Jon. 1996. Voters, issues and elections: Are the candidates' messages getting through? Journal of Politics 58:486-515.

Downs, Anthony. 1957. An economic theory of democracy. New York: Harper.

Gilens, Martin. 2001. Political ignorance and collective policy preferences. American Political Science Review 95:379-96.

Green, Donald, and Ian Shapiro. 1994. Pathologies of rational choice theory. New Haven, CT: Yale University Press.

Hart, Roderick. 2000. Campaign talk: Why elections are good for us. Princeton, NJ: Princeton University Press.

Iyengar, Shanto, and Donald Kinder. 1987. News that matters. Chicago: University of Chicago Press.

Jacobs, Lawrence, and Robert Shapiro. 2000. Politicians don't pander: Political manipulation and the loss of democratic responsiveness. Chicago: University of Chicago Press.

Jamieson, Kathleen Hall, and Paul Waldman. 1997. Mapping campaign discourse: An introduction. American Behavioral Scientist 40:1133-38.

--, eds. 2001. Electing the president 2000: The insiders' view. Philadelphia: University of Pennsylvania Press.

Jarvis, Sharon. 1999. Party style: Partisanship in presidential campaign speeches, 1948-1996. Paper presented at the conference of the International Communication Association, San Francisco.

King, David. 1997. The polarization of American parties and mistrust of government. In Why people don't trust government, edited by Joseph Nye, Philip D. Zelikow, and David C. King. Cambridge, MA: Harvard University Press.

Krosnick, John. 1990. Americans' perceptions of presidential candidates: A test of the projection hypothesis. Journal of Social Issues 46:159-82.

Kuklinski, James, Paul J. Quirk, Jennifer Jerit, David Schwieder, and Robert F. Rich. 2000. Misinformation and the currency of democratic citizenship. Journal of Politics 62:790-816.

Lichter, S. Robert. 2001. A plague on both parties: Substance and fairness in TV election news. Harvard Journal of Press/Politics 6 (3):8-30.

Lodge, Milton, Marco Steenbergen, and Shawn Brau. 1995. The responsive voter: Campaign information and the dynamics of candidate evaluation. American Political Science Review 89:309-26.

Norpoth, Helmut, and Bruce Buchanan. 1992. Wanted: The education president. Public Opinion Quarterly 56:87-99.

Page, Benjamin. 1976. The theory of political ambiguity. American Political Science Review 70:742-52.

Pew Research Center for the People and the Press. 2000. Presidential debate clouds voters' choice. Available from www.people-press.org.

Project for Excellence in Journalism. 2000. A question of character: How the media have handled the issue and how the public has reacted. Available from www.journalism.org.

Shepsle, Kenneth. 1972. The strategy of ambiguity: Uncertainty and electoral competition. American Political Science Review 66:555-68.

Sniderman, Paul Richard Brody, and Philip Tetlock. 1991. Reasoning and choice: Explorations in political psychology. Cambridge: Cambridge University Press.

West, Paul. 2000. For Gore, a battle for green territory. The Baltimore Sun, October 24, A1.

Paul Waldman is associate director of the Annenberg Public Policy Center at the University of Pennsylvania.

Kathleen Hall Jamieson is Walter H. Annenberg dean and professor of communication at the Annenberg School for Communication of the University of Pennsylvania. Her latest book (with P. Waldman) is The Press Effect: Politicians, Journalists, and the Stories That Shape the Political World (Oxford University Press, 2002).