Polling and the Media |
Of Polls, Mountains
U.S. Journalists and Their Use of Election Surveys
THOMAS E. PATTERSON is the Bradlee Professor of Government and the Press at Harvard University’s John F. Kennedy School of Government.
Address correspondence to the author; e-mail: thomas_patterson{at}harvard.edu.
 |
Abstract |
|---|
 TOP Abstract Polls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
Polls are a prominent feature of U.S. election news coverage. Although polls are used to explain voter opinion, they are employed mostly to fuel horse-race coverage and to craft images consistent with the candidates’ positions in the race. Moreover, U.S. journalists sometimes misinterpret polls by slighting the possibility that changes in candidate preference are the result of survey error rather than real change. On balance, U.S. journalists’ dependence on polls adversely affects the quality of American election coverage.
Everyone knows that Harry Truman charged from behind in 1948 to edge out Thomas Dewey on the strength of his "give-’em-hell" style and a barnstorming whistle-stop campaign that traversed America. Truman’s blunt talk and boundless energy captured Americans’ imagination, gaining their confidence and winning their votes.
Truman’s closing rush makes a nice story, but it was not the story that journalists told during the 1948 campaign. Instead, the press portrayed Truman as a weak candidate whose stridency was a sign of desperation. In its November 1, 1948, issue, Newsweek described Truman as "a woefully weak little man, a nice enough fellow but wholly inept." Although some reporters believed Truman was making inroads, they were far outnumbered by the skeptics. In a front-page story the day before the election, the New York Times declared, "The rosy prospect of victory for the Truman ticket on Election Day finds no credence outside Mr. Truman’s kitchen cabinet." Time polled forty-seven journalists; all of them predicted a Truman defeat.
America’s journalists were suffering from a shortage of polls. America’s leading polling firm, Gallup, released its final survey three weeks before the election, which, like earlier ones, showed Truman far off the lead. Expecting him to lose, reporters supplied Truman with an image fit for a loser.
The Harry Truman we know today—he of "give-’em-hell" and whistle-stopping fame—became the press’s story line after the votes were counted on Election Day. Faced with a surprise winner, journalists deftly gave Truman a flattering new look.
Today’s journalists do not have to worry that they will badly misjudge the outcome of a presidential race. So many polls are taken during elections that reporters could not possibly miss the mark as widely as those who covered the 1948 campaign. However, the other poll-related problems that affected election coverage in 1948 continue to plague campaign news. This article will show that journalists continue to craft superficial images tied to the candidates’ support in the polls and that they continue to construct election narratives rooted in the candidates’ positions in the race. The article will also highlight a problem unknown in 1948—the tendency of journalists to report small changes from one poll to the next as a manifestation of actual change in voters’ preferences rather than as a reflection of survey error.
|
Polls by the Hundreds, in America Anyway |
|---|
|
TOPAbstractPolls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
As late as the 1956 presidential election, polls were a relatively small part of U.S. election coverage. However, Gallup, Harris, and other commercial pollsters increased their polling efforts in the 1960s, and by the 1970s most of the major news organizations were conducting their own polls. Between 1976 and 1988, the number of media-sponsored election polls nearly tripled (Ladd and Benson 1992
). In each of the presidential elections since the 1980s, several hundred polls were reported, with roughly half of them coming after Labor Day (Erikson and Wlezien 1999; Rosenstiel 2005).
American journalists are not alone in their fascination with polls. Election surveys, including media-sponsored polls, are increasingly a part of European election coverage (Donsbach 2001). However, European journalists have not embraced polls as fully as have their American counterparts (Brettschneider 1997; Hardmeier 1999). European journalists also use polls more selectively than American journalists—for example, by cutting back sharply in their poll reporting when covering a one-sided race (Farnsworth and Lichter 2003; Norris et al. 1999).
These differences owe in part to variation in American and European journalism. Although "objectivity" is a norm of both traditions, the concept has a different meaning in Europe than it does in the United States (Kleinnijenhuis and DeRidder 1998; Patterson 1998). For European journalists, objectivity extends to critical scrutiny of the forces behind political action, including partisan ideology. Although European reporters aim more to inform than to persuade, partisan neutrality is not for them a strict command.
For American journalists, whose reporting model dates to the Progressive Era, objectivity is nearly synonymous with partisan neutrality. U.S. reporters are expected to withhold suppositions that would favor one party over the other. Of course, American journalists do not hesitate to expose political scandals that are damaging to a particular party, and they do not ignore blatantly false or deceptive partisan claims (Patterson and Donsbach 1996). Nevertheless, they typically refrain from compositions that suggest one party has the better candidates or policies. It is a model of reporting that European newspaper journalists find puzzling and unnecessarily confining. An "intellectual apartheid" is how Martin Walker of the London-based Guardian describes America’s straitjacketed political reporting (quoted in Katz 1993, p. 24).
Partisan neutrality leads U.S. journalists to shy away from full-bore assessments of election issues. Instead, as Paul Weaver (1972, p. 69) notes, American journalists tend to see campaigns as "a game played by individual politicians for advancement, gain or power." Although policy problems and debates are part of the game, "these are [newsworthy largely] insofar as they affect, or are used by, players in pursuit of the game’s rewards." Or, as Larry Bartels (1988, p. 32) expresses it, "In covering a presidential campaign, the media tell us more about who is winning and who is losing than they do about who is fit to be president."
Polls fit nicely into this game-centered perspective. They supply the data that enable U.S. reporters to tell their audience who is ahead and who is behind. And because of their scientific basis, polls fall safely within American journalists’ conception of objectivity (Gitlin 1991). U.S. reporters thus have more reason than their European counterparts to believe that frequent polling is a vital tool of election reporting.
|
Feeding the Horses |
|---|
|
TOPAbstractPolls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
In the United States and Europe (Media Tenor 2002
, 2005), the use of polls spurs horse-race reporting—the tendency to treat elections as if they were sporting events where the paramount goal is to get across the finish line in first place. Only in America, however, have polls elevated the horse race to the point where it overshadows all else (Farnsworth and Lichter 2003). Election coverage is necessarily selective. Campaigns include political activity and strategy, but they also encompass an entire voting public, an array of special interests, and a range of governing problems. The story possibilities are endless, which means that journalists must find some way to narrow their selections. U.S. journalists’ game-centered angle on campaigns inclines them to direct their gaze at the horse race.
Polls serve to keep it there. Trial-heat polls are a constant reminder of the candidates’ standing in the race, as well as grist for strategy-centered stories. Even in election reports where polls are not the main subject, they often underpin the storyline (Welch 2002). A study of network nightly news coverage of the 1988–2004 presidential general elections (Farnsworth and Lichter 2005) found that the horse race is easily the dominant theme of campaign news. It accounted for roughly half or more of election stories in each of the five general elections studied, reaching a high of 71 percent in 2000. Even in the 2004 general election, which was characterized by unusually powerful issues, including Iraq and the economy, the horse race consumed as much news time as did all election issues combined. The policy issue that Americans said they cared the most about in 2004—the economy—received less than 5 percent of the total coverage (Media Tenor 2005).
|
Precision Journalism, Sometimes |
|---|
|
TOPAbstractPolls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
Polls have been trumpeted as an instrument for "precision journalism" (Meyer 1972
). And indeed, polls can enable journalists to report more accurately on what the public is thinking. Polls are not foolproof, however. Late in the 1992 campaign, for example, USA Today noted that Bill Clinton had dropped 2 percentage points since the previous day’s poll and that George Bush had gained 2 points. "The race is getting tight" is how USA Today described the change. When a subsequent tracking poll showed Clinton and Bush at their earlier levels, USA Today carried the headline, "Bush’s Surge Trickles Away." None of the observed movement exceeded sampling error, however, raising the question of whether any actual change in the candidates’ positions had occurred (Rhee 1996). Shaw and Roberts (2000), on the basis of their assessment of election polls, concluded that shifts in voter support within a relatively short period without an identifiable triggering event are typically a consequence of polling error. In a time-series analysis of 1996 presidential general election polls, Erikson and Wlezien (1999) found that movement in candidate support from one poll to the next was more often attributable to sampling error than to actual change in voter preferences. Differences in the way that the various polling organizations selected their samples, screened their respondents, and worded their questions also accounted for much of the observed variation. Real variation—that associated with actual shifts in preferences—accounted for the least amount of movement and, when it occurred, was typically short-lived.
Ironically, saturation polling can also protect journalists from exaggerating the significance of small changes in voter preference. In October 2004 journalists found themselves reporting on a tight race while trying to make sense of the conflicting outcomes of a slew of polls, including those of Gallup, Harris, Zogby, Rasmussen, and a half-dozen news organizations. Some of those polls showed George W. Bush and John Kerry in a dead heat, but others had Kerry ahead, and still others had Bush on top. Adding to the confusion were complaints from the opposing sides that some pollsters were manipulating their surveys in an attempt to boost their favorite candidate. The controversy led journalists, pundits, and bloggers into a public discussion of survey methods and their limits (Blumenthal 2005). Among the news stories on the subject was a front-page New York Times article on October 18 entitled "Bush Leads. Make That Kerry: Why Can’t the Pollsters Agree?" From a cursory assessment through Lexis-Nexis of poll coverage, it appears that the controversy prompted journalists to exercise more care in the reporting of national polls during the closing weeks of the 2004 campaign.
Whether the effect will be a lasting one is an open question. The CBS commentator Eric Sevareid once compared election reporters to alcoholics, saying that their vow to improve on the reporting of the last election is all but forgotten when the wine of a new campaign touches their lips. In fact, the lesson of the conflicting national polls was a bounded one even in 2004. In the closing days of the 2004 campaign, as it became clear that the election would be settled in the battleground states, major news organizations shifted their focus to these states. The New York Times, for example, ran a "State by State Series" that honed in on individual battleground states. Although the horse race was the persistent theme of these articles, few of them cautioned the reader about shortcomings in the statewide polls that informed the stories. An October 30 New York Times story on Wisconsin, for example, referred to two recent Wisconsin polls that "showed Mr. Kerry holding a one- or two-point lead." Nothing more was said about the polls, not even the identity of their sponsors. The polls appear to have been a Zogby poll released on October 26 that had Kerry up by 2 points and an American Research Group poll released on October 28 that had Kerry ahead by a point. Both of these Wisconsin surveys were based on 600 likely voters, which meant that the observed difference in Kerry’s and Bush’s support was within sampling error. Moreover, the Zogby organization had been singled out in news articles earlier in October as a firm whose survey methods were particularly controversial.
|
Transcendent Images |
|---|
|
TOPAbstractPolls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
Election polls provide snapshots of what voters are thinking at various moments in a campaign. Journalists use polls to gauge the public mood but, aside from the reporting of trial heats, do so only sparingly. In a study of election coverage in 2000, Larson (2001)
found that a mere 1 percent of poll mentions referred to survey questions about the candidates’ leadership qualities.
Paradoxically, surveys heighten journalists’ attention to the candidates, rather than to the voters themselves. Rhee (1996) found that as poll references increase in news stories, so, too, do references to candidates’ strategies. Journalists use polls transcendently, as a basis for claims about the candidates.
This tendency derives from the age-old definition of news as events (Robinson and Sheehan 1983). Candidates’ activities are events. Voters’ attitudes are not. Although voters’ partisan loyalties and policy preferences are the major influences on the vote, these influences are complex and not easily analyzed or reported. Moreover, because these influences are relatively stable, they are poorly suited to journalists’ need to say something new each day. Campaign activity has less impact on the vote, but it is more dynamic and is more easily grasped and reported.
Of course, reporters are aware of the larger forces driving the vote. Yet, with their sights aimed squarely at the candidates, they tend naturally to find their explanations for the vote in what candidates are doing. If a candidate is lagging in the polls, the problem must lie with the candidate’s personality or game plan. As John Kerry’s chances of getting the Democratic nomination seemed to fade in late 2003, and when he later slipped in the polls at the time of the summer Republican convention, journalists portrayed him as aloof and lifeless. But when in each case Kerry rebounded, he was presented as vigorous and savvy (Media Tenor 2005).
Farnsworth and Lichter’s study (2005) of network news coverage of the 2004 general election provides a more substantial basis for assessing the relationship between Kerry’s position in the 2004 race and his news image. After the GOP national convention and through the month of September, Kerry trailed Bush in the polls. And during September, Kerry also lagged behind Bush in terms of "good press," which Farnsworth and Lichter calculated by counting every favorable or unfavorable evaluation by a nonpartisan source (for example, a journalist, voter, or expert) of a candidate’s record, policy positions, and personal traits (but not including favorable or unfavorable horse-race evaluations). When Kerry then closed the gap during the October debates, the tone of the candidates’ coverage changed. Suddenly, Kerry was getting the "good press." The nonpartisan voices heard on the evening newscasts were substantially more likely to say favorable things about Kerry’s record, issues, and character than they were to say favorable things about these aspects of Bush’s candidacy.
The fact is, reporters have wide leeway in making claims about why candidates succeed or fail. Whether a candidate is doing well or poorly is clear enough from the polls, but the reasons why the candidate is doing well or poorly are not, leaving reporters relatively free to develop explanations of their own making (Rosenstiel 2004). When a candidate gains ground, reporters can attribute it to favorable aspects of his personality or campaign style. When a candidate starts to slide, negative qualities can be singled out.
What is said about the candidates must at some point be reconciled with their positions in the race. A reporter cannot routinely say that a candidate who is lagging in the polls has more personal appeal; to do so is to invite a complex explanation that might call into question the reporter’s objectivity. Accordingly, the temptation to say unfavorable things about a faltering candidate is nearly irresistible. Similarly, reporters find it hard not to puff up a surging candidate. When George H. W. Bush languished in the polls during the 1988 campaign, reporters said it was because he looked weak. Newsweek ran a Bush cover story entitled "Fighting the Wimp Factor." However, when Bush took the lead in polls after the GOP national convention, Newsweek declared in its September 12 issue that Bush had "banished . . . the wimp factor."
With their sights set squarely on the candidates, journalists find their explanations for poll results chiefly in what they see in the candidates rather than in what the polls themselves might reveal. George H. W. Bush’s stage presence undoubtedly played some part in his surge in the polls during the 1988 GOP convention, but a more substantial explanation was available to journalists if they had studied the polls they were trumpeting. The voters who were moving into his column were chiefly Republican-leaning voters who had not theretofore been paying much attention to the campaign (Shelley and Hwang 1991). In other words, Bush was reaping what first-time presidential nominees always reap during a party convention—their party’s less-involved loyalists.
The press’s portrayal of Bush was no different in kind than that of other recent nominees. Patterson (1993) found that media portrayals of the 18 major-party presidential nominees between 1960 and 1992 were governed primarily by poll position. A nominee riding a "bandwagon" received coverage predictable in its tone and imagery and different from the coverage given a nominee who found himself "losing ground" or a nominee who was a "likely loser" or one who was a "front-runner." Like Truman in 1948, each nominee was clothed in imagery consistent with his position in the race.
|
Thoughts about the Game They Portray |
|---|
|
TOPAbstractPolls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
A modern campaign would be unthinkable without media polls, and poorer for it if they were to disappear. Polls can generate interest in a campaign and, in some instances, clarify the voters’ options (Abramson et al. 1992
).
However, the use of polls can extend beyond reason. U.S. elections have reached that point. As reported by the American press, a campaign is a spectacular struggle: rapid followers, do-or-die encounters, strategy, tactics, winners, and losers. A campaign, to be sure, is all of these things, but it is also more than these things. It is an opportunity to choose the nation’s leadership. As it happens, voters do not necessarily need a lot of hard information about the candidates to make a reasoned choice (Popkin 1991). Yet, the more information they have, the more likely they will make such a choice. America’s poll-driven election coverage squeezes out content that would inform voters’ judgment (Farnsworth and Lichter 2003).
Poll-driven stories also distort the public’s perceptions of the candidates. They are portrayed as poll-minded strategists and saddled with flimsy, poll-derived images, which color citizens’ opinions in a largely negative way (Cappella and Jamieson 1997; Valentino, Buhr, and Beckmann 2001). Candidates are strategists, of course. But the fact that they dramatize their appeals and tailor their messages is nothing new. Such maneuvers are as old as politics itself. What is new is the penetrating intensity with which candidates’ activities are exposed, dissected, and criticized. It should occasion no surprise that as candidates have increasingly been portrayed as masters of strategy and manipulation, Americans would think less highly of them. The Gallup organization first asked voters their satisfaction with the presidential nominees in 1936. Through the 1960s the only nominee who, on balance, was perceived unfavorably at the end of the campaign was Barry Goldwater in 1964. Since then—that is, during the period of hyper-polling—more than a third of the nominees have wound up with an unfavorable image. Several factors account for the trend, but the tendency of the press to cast the candidates as strategic actors, whose every move is driven by a determination to win, is surely among them.
Nor should it occasion surprise that Americans have soured on campaigns. Elections have become negative affairs filled with horse-race commentary and analysis. A Vanishing Voter Project survey (Patterson 2002) at the end of the 2000 campaign found that twice as many respondents felt the campaign was "depressing, that it hasn’t been nearly as good as a campaign should be" as felt the campaign was "uplifting, that it made [you] feel better about elections." The same survey found that nearly two-thirds of Americans agree with the statement "political campaigns today seem like theater or entertainment rather than like something to be taken seriously." Cast as respondents in a show waged in their name but not of their doing, Americans have learned to see campaigns as stagecraft.
Polling cannot be blamed for all or even most of the ills of American campaigns. But this powerful technique has taken its place alongside big money, negative advertising, attack journalism, and other developments that in combination have diminished American elections.
|
References |
|---|
|
TOPAbstractPolls by the Hundreds,...Feeding the HorsesPrecision Journalism, SometimesTranscendent ImagesThoughts about the Game...References |
|---|
Abramson, Paul R., John H. Aldrich, Phil Paolino, and David W. Rohde. 1992. "‘Sophisticated’ Voting in the 1988 Presidential Primaries." American Political Science Review 86:55–69.[CrossRef][Web of Science]
Bartels, Larry M. 1988. Presidential Primaries and the Dynamics of Public Choice. Princeton, NJ: Princeton University Press.
Blumenthal, Mark M. 2005. "Toward an Open-Source Methodology: What We Can Learn from the Blogosphere." Public Opinion Quarterly 69:655–669.
Brettschneider, Frank. 1997. "The Press and Polls in Germany, 1980–1994." International Journal of Public Opinion Research 9:248–65.
Cappella, Joseph, and Kathleen Hall Jamieson. 1997. Spiral of Cynicism. New York: Oxford University Press.
Donsbach, Wolfgang. 2001. "Who’s Afraid of Election Polls." Paper presented at the annual ESOMAR conference, Amsterdam.
Erikson, Robert S., and Christopher Wlezien. 1999. "Presidential Polls as a Time Series." Public Opinion Quarterly 63:163–77.
Farnsworth, Stephen J., and S. Robert Lichter. 2003. The Nightly News Nightmare: Network Television’s Coverage of the U.S. Presidential Elections, 1988–2000. Lanham, MD: Rowman and Littlefield.
Farnsworth, Stephen J., and S. Robert Lichter. 2005. "The Nightly News Nightmare Revisited: Network Television’s Coverage of the 2004 Presidential Election." Paper presented at the annual meeting of the American Political Science Association, Washington, DC.
Gitlin, Todd. 1991. "Bites and Blips." In Communication and Citizenship: Journalism and the Public Sphere, ed. Peter Dalghren and Colin Sparks, pp.120–37. New York: Routledge.
Hardmeier, Sibylle. 1999. "Political Poll Reporting in Swiss Print Media." International Journal of Public Opinion Research 11:257–74.
Katz, Jeffrey L. 1993. "Tilt?" Washington Journalism Review 15 (January–February): 23–27.
Kleinnijenhuis, Jan, and Jan A. DeRidder. 1998. "Issue News and Electoral Volatility." European Journal of Political Research 33:413–37.[CrossRef]
Ladd, Everett Carll, and John Benson. 1992. "The Growth of News Polls in American Politics." In Media Polls in American Politics, ed. Thomas E. Mann and Gary R. Orren. Washington, DC: Brookings Institution Press.
Larson, Stephanie G. 2001. "Poll Coverage of the 2000 Presidential Campaign on the Network News." Paper presented at the annual meeting of the American Political Science Association, San Francisco.
Media Tenor. 2002. "Democracy Needs Content: Coverage on Schroeder and the SPD." Internet data release. Bonn: Media Tenor. October.
Media Tenor. 2005. "Contest Lacks Content." Media Tenor 1:12–15.
Meyer, Philip. 1972. Precision Journalism. Bloomington: Indiana University Press.
Norris, Pippa, John Curtice, David Sanders, Margaret Scammell, and Holli A. Semetko. 1999. On Message: Communicating the Campaign. London: Sage.
Patterson, Thomas E. 1993. Out of Order. New York: Knopf.
Patterson, Thomas E. 1998. "Political Roles of the Journalist." In The Politics of News: The News of Politics, ed. Doris Graber, Denis McQuail, and Pippa Norris, pp. 17–32. Washington, DC: CQ Press.
Patterson, Thomas E. 2002. The Vanishing Voter. New York: Knopf.
Patterson, Thomas E., and Wolfgang Donsbach. 1996. "News Decisions: Journalists as Partisan Actors." Political Communication 13:455–68.
Popkin, Samuel. 1991. The Reasoning Voter: Communication and Persuasion in Presidential Campaigns. Chicago: University of Chicago Press.
Rhee, June Woong. 1996. "How Polls Drive Campaign Coverage: The Gallup/CNN/USA Today Tracking Poll and USA Today’s Coverage of the 1992 Presidential Campaign." Political Communication 13:213–29.
Robinson, Michael, and Margaret Sheehan. 1983. Over the Wire and on TV. New York: Russell Sage Foundation.
Rosenstiel, Tom. 2004. "Character and the Campaign: What Are the Master Narratives about the Candidates and How Is the Public Reacting to Them?" Internet release. Washington, DC: Project for Excellence in Journalism. July. Available online at www.journalism.org/resources/research/campaign2004/character/default.asp (accessed November 22, 2005).
Rosenstiel, Tom. 2005. "Political Polling and the New Media Culture: A Case of More Being Less." Public Opinion Quarterly 69:698–715.
Shaw, Daron R., and Brian E. Roberts. 2000. "Campaign Events, the Media and the Prospects of Victory: The 1992 and 1996 US Presidential Elections." British Journal of Political Science 30:259–89.
Shelley, Mack C., II, and Hwarng-Du Hwang. 1991. "The Mass Media and Public Opinion Polls in the 1988 Presidential Election." American Politics Quarterly 19:59–79.
Valentino, Nicholas A., Thomas A. Buhr, and Matthew N. Beckmann. 2001. "When the Frame Is the Game." Journalism and Mass Communication Quarterly 78:93–112.
Weaver, Paul. 1972. "Is Television News Biased?" Public Interest 27:65–76.
Welch, Reed L. 2002. "Polls, Polls, and More Polls." Harvard International Journal of Press/Politics 7:102–14.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||