The Manchester Union Leader's influence in the 1996 New Hampshire Republican primary.

Farnsworth, Stephen J. ; Lichter, S. Robert

In a media environment dominated by large corporations answerable to stockholders, newspapers and television news programs are often criticized for being bland and inoffensive as they try to protect the financial bottom line (Picard 1998; Sanford 1999). No one, however, would ever make such a claim against the Manchester Union Leader, New Hampshire's astonishingly aggressive, independent newspaper. For decades, its front-page, fire-breathing editorials have attacked politicians and policies in the most strident of terms, hurling such insults as "Dopey" Dwight Eisenhower, Jerry "the Jerk" Ford, and Edmund "Moscow" Muskie (cf. Buell 2000; Veblen 1975; White 1970, 1973). When President Clinton visited the state after the impeachment controversy, the paper ran a front-page banner headline "Mr. President, you're a disgrace!" above the masthead (February 19, 1999).

Were it not for the New Hampshire primary, the Union Leader would be little more than a curiosity on the national stage, a museum piece offering the opinionated invective that American journalism largely outgrew a century or more ago (Moore 1987; Schudson 1978; White 1973). Both the Union Leader and its Sunday edition, the New Hampshire Sunday News, have circulations too small to convey much influence beyond the state (66,250 daily and 93,768 on Sunday for the year ended March 31, 1997, according to the Union Leader report to the Audit Bureau of Circulations). New Hampshire without its first primary likewise would not be expected to figure prominently in national politics, as the state is home to fewer than one out of every two hundred Americans.

But that primary changes everything for both the Granite State and the Union Leader. Every four years, and for much of the time in between, Republican presidential candidates woo the state's electorate and its dominant paper, long thought to help determine which Republicans leave New Hampshire as likely nominees and which ones leave as ex-presidential candidates (Buell 2000; Mayer 1987; Veblen 1975). Ronald Reagan received a major boost from the paper on his way to winning the 1980 contest, and commentator Pat Buchanan became a viable alternative to President Bush in 1992 thanks to the paper's aggressive support and a weakening economy. Of course, the paper's influence has its limitations. Not even the Union Leader could make a winner out of Pete DuPont, the paper's favored candidate in 1988, or Steve Forbes, backed by the paper in 2000 (Adams 1987; Brady and Johnston 1987; Ceaser and Busch 1993, 1997; Palmer 1997; Robinson 1978).

This study examines the relationship between Union Leader news and editorial coverage and candidate preferences by likely New Hampshire voters during the weeks before the state's 1996 Republican primary. This study uses a content analysis of the paper's news stories and published editorials and opinion columns conducted by the Center for Media and Public Affairs, a nonpartisan media research institute in Washington, DC. The content analysis, which looks at the dimensions of candidate horse-race standings as well as more substantive matters, is compared to daily tracking polls conducted by American Research Group Inc.

The New Hampshire Primary and the Union Leader

Many scholars and politicians complain about New Hampshire's privileged position atop the presidential primary calendar, but no one has been able to do much about it. New Hampshire keeps moving its primary earlier to stay ahead of other competing state contests, and candidates who propose that New Hampshire not go first are quickly dismissed by the state's electorate, as happened to presidential aspirant Sen. Phil Gramm (R-TX) in 1996 (Buell 2000; Ceaser and Busch 1997; Mayer 1997; Trent 1998).

While New Hampshire may be unrepresentative of the rest of America-the state's population is 98 percent white--both New Hampshire and equally unrepresentative Iowa have been able to retain their advantageous positions largely because would-be presidential candidates fear to offend such powerful electorates and because the other forty-eight states have been unable to agree on an alternative arrangement (Palmer 1997). New Hampshire does not stay first in line because of some general sense of respect for its primary. Peirce and Hagstrom (1983, 205) are among those who have condemned the state's position in the nomination process in harsh terms, describing New Hampshire as "the state that permitted America's most bizarre and self-centered political environment to flourish."

Other researchers offer more positive assessments regarding the Granite State and its electorate, more of whom read the Union Leader than any other paper. Buell (2000) noted that the primary electorates in New Hampshire are actually quite close to the composition of both parties' primary electorates nationally. Buell (2000) and Brereton (1979, 1987) have observed that the state has a better record at picking the eventual nominees, particularly Republican ones, than many other states with later primaries.

New Hampshire's relatively small size and its political culture's intense focus on the primary offer some advantages for some candidates, as many voters can meet individual candidates face to face at coffees and town meetings or at least read about them extensively in the state's locally oriented publications (Brereton 1979, 1987; Duncan 1991; Sprague 1984; Rueter 1988; Palmer 1997). The existence of widespread retail politics in the Granite State gives dark-horse candidates greater opportunity to become competitive, if not to win the nomination. Poorly financed underdogs like Jimmy Carter--who won both Iowa and New Hampshire in 1976--could never have campaigned effectively if the first contest had been in a much larger and more expensive media environment like California or New York (Brereton 1979; Buell 1987; Robinson 1985).

But this perceived distinctiveness may be less true than it once was, and the media may explain why. Some researchers have suggested that the networks have begun to play a dominant role in the primary, making New Hampshire more like the rest of the country in terms of television-dominated campaigning than many of its supporters would like to admit (Brady and Johnston 1987; Farnsworth and Lichter 1999, 2002; Orren and Polsby 1987; Rosenstiel 1994).

The state's dominant and stridently conservative paper, the Union Leader, figures prominently into attacks on New Hampshire's leadoff position, particularly on the part of reporters and of candidates subjected to the paper's attacks (e.g., Peirce and Hagstrom 1983; White 1970, 1973). Academic researchers who have considered the paper's influence have reached conflicting conclusions. Moore (1987) concluded that the Union Leader had a powerful effect on the 1980 Republican primary, which was won by Ronald Reagan, but Palmer (1997) argued that George Bush's poor performance at a debate in Nashua had at least as much to do with Reagan's ultimate victory in the state as did the support Reagan received from the Union Leader. In the 1992 GOP contest, Buell (2000) argued generally in support of the Union Leader's influence, though he found in that contest that ideology was roughly as important to a voter's choice as whether that voter was a reader of the Union Leader. The paper's readers split 50-50 on Bush and Buchanan in the 1992 primary, while voters who did not read the paper favored Bush by a 63-36 margin.

The Union Leader's basic approach in influencing the state's primary voters is twofold according to past research: praise relentlessly the favored candidate and attack without mercy the candidate who poses the biggest threat to that favored candidate (Brereton 1987; Buell 2000; Moore 1987). Indeed, Moore (1987) suggested that the paper's greatest influence may be in deterring votes for the candidate the paper campaigns against rather than helping the paper's endorsed candidate.

Buell (2000) speculated that the Union Leader's influence may have declined somewhat following the death of William Loeb in 1981, the expanding influence of Manchester's WMUR-TV, and the fact that the Boston Globe and the Boston television stations are available throughout much of the state (cf. Palmer 1997). Even so, Buell found that the Union Leader's twice-endorsed candidate, Pat Buchanan, did better in towns where the paper's penetration was the greatest and did the worst in those communities where the paper's penetration was the least in both the 1992 and 1996 primary contests.

Some broad issues regarding general media influence are considered in this study of the Union Leader. Central among them is the "bandwagon effect," which suggests that positive media commentary boosts the candidates who receive such treatment (Bartels 1985; Patterson 1980, 1994). Along these same lines, Farnsworth and Lichter (1999) found that an "antibandwagon effect" was particularly powerful for a candidate (in this case Bill Clinton) who received overwhelmingly negative news coverage before the 1992 Democratic New Hampshire primary. The Union Leader approach of launching aggressive attacks on some candidates and offering aggressive support for others (Buell 2000; Moore 1987) creates the potential for both a powerful bandwagon effect and a strong antibandwagon effect, perhaps greater than would be found for less strident news outlets.

On another key issue, though, one might expect the paper's influence to be quite different from that of the news media generally. Horse-race coverage, in effect treating politics as a form of sports, often plays a greater role in citizen candidate preferences than does coverage of more substantial matters (cf. Patterson 1980, 1994). The Union Leader, after all, offers a very different stimulus than most newspapers or television stations. The paper's priming and agenda setting encourages the use of substantive evaluations, which suggests that substantive evaluations may also be quite important in a study of this paper's influence (Iyengar and Kinder 1987).

Methods and Measures

This study examines the interrelationship between the Union Leader's news and editorial coverage and daily tracking poll results starting on January 25, 1996, and ending on February 19, the day before the primary. Daily poll standings for the top four candidates--Pat Buchanan, Bob Dole, Lamar Alexander, and Steve Forbes--were derived from surveys of at least 450 likely Republican primary voters conducted by American Research Group (with an error of at most [+ or -] 4.6 percentage points). The content analysis used here provide twenty-four days for analysis, all the days covered by the tracking polls. These data are more extensive than past surveys of Republican contests, which generally involved relatively few respondents or survey points or both (Bartels and Broh 1989; John 1989; Moore 1984).

The tracking poll standings used in this study as the dependent variables combine the results of questions asked over three-day periods. The paper's morning news coverage on January 25 (the first day's coverage included in this study) was compared against tracking poll results collected January 25 to 27. While a portion of the poll results may be affected by media coverage on subsequent days (say in this case by a story in the paper on Jan. 26 or 27), this approach includes the lagged effects of news stories as well as immediate impacts. Separate one-day standings in a tracking poll of this size would be too small for effective analysis. A different lag would reduce the effect of the stories being examined and increase the impact of subsequent news reports and editorials as well as reduce the number of cases available for analysis.

All but two of the twenty-four available tracking polls contain three days of surveys. The final two tracking polls (survey sets ending Feb. 18 and 19) involved only two days of polling, though the sample sizes for those two tracking polls were expanded to take account for the smaller number of days of polling used. The paper's news and editorial coverage for February 16 is excluded from this study because the last three-day tracking poll (Feb. 15 to 17) is analyzed against the news printed February 15, and the first of the two-day tracking polls (Feb. 17-18)was analyzed against news from February 17. The other two-day tracking poll, conducted February 18 to 19, was compared against news from February 18.

A total of 385 Union Leader news stories and 65 editorials and opinion columns published on those twenty-four days are used in this analysis. The number of news articles in the paper on the Republican campaign during this period ranged from a low of 7 on January 27 to a high of 64 on February 9, with a daily mean of 16.04. The Union Leader, which first endorsed Buchanan in September 1995 for the February 20, 1996, primary, ran its largest number of editorials during the survey period (9) on February 8 and 11, with a mean number of editorials and columns relating to the campaign of 2.7 per day. The paper published no editorials and no opinion columns related to the presidential nomination contest on four of the twenty-four days contained in the study: January 27 and February 3, 10, and 13.

The content coding process used here involves analyzing and tabulating individual statements contained in news articles and editorial columns that contain explicitly positive or negative judgments of the candidates with respect to their political prospects (the horse race) or to more substantial matters, which include character, policy stances, and the candidate's professional competence. Trained coders carefully examined the content of each news story, opinion column, and editorial, with an intercoder reliability of at least .80 for all content analysis variables used here.

The substantive independent variables relate to a candidate's capacity to be president and include discussions of character, policy issues, and previous job performance. Under the coding system used here, some campaign-related measures are considered substantive evaluations if those issues are linked to character issues rather than to horse-race matters. Examples of campaign issues that may relate to character and thus are classified as substantive evaluations here include whether a candidate is said to be making misleading or deceptive campaign advertisements, is too close to special interests, or is conducting a divisive campaign. All of these would be coded as negative substantial evaluations. (Further information on the coding process is available in the appendix.)

A candidate said to be gaining ground has received a positive horse-race evaluation, while one thought to be dropping back has received a negative horse-race evaluation. Mixed evaluations are excluded from this analysis, as are neutral references, which have been shown in past research to have no effect on tracking poll standings in New Hampshire (Farnsworth and Lichter 1999).

These four measures are calculated for the paper's news articles; they are also calculated for editorials and opinion columns, which offer less even-handed opinionated perspectives on the day's events than do the news pages. This study, then, contains a total of eight media coverage measures per candidate: four news content measures and four editorial content measures.

Buchanan, the winner of the 1996 New Hampshire primary with 27.2 percent of the vote, led Dole on only two of the twenty-four tracking polls in this study: the periods February 14 to 16 and 15 to 17. Buchanan and Dole, who finished second with 26.2 percent of the vote, were tied in a third poll, conducted February 13 to 15. Dole led in the twenty-one other tracking poll periods, though that lead was sometimes smaller than the margin of error. Dole's support ranged from 23 to 33 percent during the survey period, while Buchanan's support ranged from 13 to 28 percent in the tracking polls used in this study.

Lamar Alexander, who finished third with 22.6 percent of the vote, was never higher than third in any of the tracking polls. Steve Forbes, who finished fourth with 12.2 percent, was in second place behind Dole in the first six tracking polls, but he fell to a second-place tie with Buchanan in the January 31 to February 2 poll. After remaining within a percentage point or two of each other for several days, Buchanan pulled ahead of Forbes for good in the tracking poll of February 5 to 7, leaving Forbes and Alexander to battle for third place during the final two weeks before the New Hampshire contest. Forbes was heavily criticized for using his immense personal fortune to compete with Dole and for his unusually negative television advertisements. A number of experts trace his decline to the reaction to that high-spending and combative approach (Denton 1998; Dover 1998; Just 1997).

Hypotheses

Hypothesis 1: The more positively a candidate is evaluated in the paper, the more positively the candidate would be viewed by the New Hampshire Republican primary electorate. The more negative the coverage, the more negatively the candidate will be viewed by likely voters.

This would be consistent with the agenda-setting ideas of Iyengar and Kinder (1987), who demonstrated that public opinion can be shaped to a considerable degree by media content.

Hypothesis 2: The impact of the paper's coverage would be the most powerful for Buchanan, the candidate receiving the most positive one-sided coverage.

This would be consistent with the schema model presented by Graber (1988), which suggests that media coverage frames public views of the political world.

Hypothesis 3: Accounts covering matters of substance should be more significant to tracking poll standings than are the horse-race stories.

Although this hypothesis departs from the research of Patterson (1980, 1994), the Union Leader's focus on substantive materials like issues and character would be expected under the schema model presented by Graber (1988) to prime Union Leader readers to focus on such issues rather than the horse race as they assess the candidates. These substantive issues are more subject to interpretation by the Union Leader than the more objective realities of horse race journalism, which can hardly stray far from the results found in public opinion surveys.

Results

Table 1 shows the number of positive and negative horse-race and substantive evaluations in Union Leader editorials and opinion columns for the four major Republican candidates. Clearly, the paper's editors made it clear how they felt in the opinion section. Pat Buchanan, the paper's favored candidate, received by far the largest number of directorial evaluations in those sixty-five articles: two hundred total directional references. The paper's support for Buchanan in its editorials and columns was overwhelming: he held a twelve-to-one advantage in positive horse race references to negative ones and a nearly three-to-one advantage in positive substantive evaluations versus negative ones. The other candidates were treated far more negatively: Dole, Alexander, and Forbes all received more than 60 percent of their directional evaluations (in editorials and columns) in the substantive negative evaluation category, as compared to only 16 percent negative substantive evaluations for Buchanan. Positive substantive evaluations of Buchanan were expressed in the editorial pages of the Union Leader eight times more frequently than they were for Dole, nearly fifteen times more frequently than they were for Forbes and twenty-two times more frequently than they were for Alexander.

Table 2 shows the number of positive and negative horse-race and substantive evaluations in Union Leader news articles for the same four candidates. While the coverage was less one-sided than on the editorial page, the paper's news writers leaned in the same direction as the paper's opinion writers. Buchanan, the paper's favored candidate, received 342 directional evaluations in the news columns and claimed both the largest percentage of positive horse race evaluations (60 percent) and the lowest percentage of substantive negative evaluations (11 percent).

Similar patterns across the two tables were found for the other candidates as well. Although the results were less one-sided in the news columns, the results were still negative for the candidates not endorsed by the paper. Forbes, who received the most directional evaluations in the news section, 619, also received the greatest percentage of negative substantive evaluations: 41 percent of all his directional evaluations in the paper's news columns. This is a decline from the editorial page, where 63 percent of the Forbes references were negative and substantive. But the overall message conveyed about this candidate remained quite negative even in the ostensibly more objective news section. Dole, the Republican frontrunner and eventual nominee, received the second largest amount of news coverage, with 618 directional evaluations, and the second largest percentage of substantive negative evaluations at 24 percent. The news results were considerably better than the 61 percent he recorded in this category on the editorial page.

The paper provided much more substantive coverage than do many news outlets, particularly the big-three network television evening news programs (cf. Farnsworth and Lichter 1999, 2002, 2003). While the coverage in the news section is somewhat similar to TV in terms of the attention paid to the horse-race aspects of the campaign, the editorial page offered a great deal more substantive coverage than that found on the three networks. The three networks devoted a far greater percentage of their attention to the horse-race aspect of the 1996 nomination campaign than did the Union Leader overall, with 47 percent of the stories focusing on the horse race (Owen 1997). A study of the 1992 Democratic nomination contest found a similar pattern: 58 percent of the network news coverage of Bill Clinton and 65 percent of the coverage of Paul Tsongas focused on the horse-race (Farnsworth and Lichter 1999). The greater focus on substantive coverage in Union Leader editorials suggests that substantive evaluations may be more important in this study than are the horse-race measures, which generally are seen as most significant to the study of network television's influence on voters.

While the Union Leader, like many other papers in America, argues that its opinions are kept separate from its news columns, the same candidates were favored and opposed in both sections. The main differences between the departments were of degree, not direction, as the news pages offered a less one-sided perspective than the opinion pages.

We now turn to the question of the extent to which the news and editorial coverage of a given candidate in the Union Leader affects the candidate's standing in subsequent tracking polls. We use net horse-race measures and net substantive evaluation measures to reduce the number of variables in the ordinary least squares (OLS) regression model, a particularly important step in this analysis because of the small sample size (N = 24 for all candidates). Positive horse-race measures are subtracted from negative ones to create a net horse-race measure for each candidate each day. The substantive evaluation measures were calculated the same way. There are two sets of net horse-race and net substantive evaluation measures in this analysis: one for news coverage, the other for editorials and opinion columns. The paper's coverage on a given morning is then compared to the results in the tracking polls that began later that same day.

OLS regression is the preferred statistical technique for these data because other techniques, including discriminant analysis and probit, would require collapsing the dependent variable, thereby reducing the effectiveness of the analysis. Furthermore, some of the potential alternative techniques (notably discriminant analysis) are very sensitive to the small sample sizes found here (Tabachnick and Fidell 1989).

Table 3 provides OLS regression models for the top four finishers in the 1996 GOP contest. The table shows unstandardized coefficients (b) with asterisks to denote statistical significance. Standard errors are in parentheses.

Reports of the Union Leader's declining influence may be premature, judging from the analysis found in Table 3. As predicted, the paper's greatest influence was on Buchanan's standings in the tracking polls. This model shows an adjusted r-squared value of .60, the highest of any of the four equations presented here. Only one variable, the editorial section horse race measure, fails to achieve statistical significance in the Buchanan analysis.

When the Union Leader spoke out in favor of Buchanan on its pages, and the paper often did so, voters listened. When the Union Leader gave the more conventional horse-race coverage in its news columns, New Hampshire voters listened. This pattern is consistent with the focus of the coverage in the two sections of the paper, or in the priming or agenda setting of voters, to use the terms of Iyengar and Kinder (1987). Substantive evaluations represented 60 percent of the directional coverage of Buchanan in editorials and columns; horse-race coverage represented 69 percent of the news coverage of the primary's eventual winner. The paper wanted to convince voters that Buchanan should be the choice for substantive reasons, and voters became increasingly convinced across the survey period. The paper sought to convince voters that Buchanan was a viable candidate, and the voters increasingly became convinced as the primary drew closer. Even a paper that offers as extensive a substantive analysis as does the Union Leader still can help exert a powerful bandwagon effect like that seen in media research of media outlets that employ more conventional approaches regarding the provision of news and opinion.

The third statistically significant measure, substantive evaluations in the news columns, indicates that the more that was said negatively about Buchanan, the higher his poll ratings became. While a positive relationship seems more consistent with the variables previously discussed, the news columns of a newspaper are not the paper's voice, not even in a paper as opinionated as the Union Leader. Any criticism of Buchanan by any quoted source, including a competing candidate, registers here as a negative substantive evaluation of Buchanan, regardless of whether the editors might disagree with the allegation. The results here might be a sort of "rally round the conservative" effect where fence-sitters reading the Union Leader might turn more solidly to Buchanan as he is attacked by his rivals and their supporters. This seems to be a particularly safe speculation given the paper's aggressive and consistent defense of Buchanan on its editorial pages.

The Alexander model was the only other statistically significant one in Table 3. There were two inverse statistically significant relationships: the more the editorial page attacked Alexander substantively and dismissed his chances in its horse race coverage, the better he did in the polls. One should note that Alexander had by far the smallest number of directional observations on the editorial pages, only twenty-three across twenty-four days--a figure that is a fraction of the two hundred directional references Buchanan received. Given the small number of references, the Alexander results should be treated with some caution.

The Alexander results may be an effect that relates largely to those readers of the Union Leader who read the paper because they have to--it is the most influential newspaper in the state--not because they agree with its editorial opinions. That is not a small group, according to some researchers (Palmer 1997). Buchanan won only one of the state's three largest cities (Manchester, which is the largest). Dole won both second-largest Nashua and third-largest Concord, even though both cities are located within the primary circulation area of the Union Leader (Buell 2000). Both cities have their own daily newspapers that compete with the Union Leader, and so the Manchester paper's influence may be less powerful in these communities.

Contrary to the findings of past research on the Union Leader, its harshly negative treatment of Dole and Forbes had virtually no impact on their standings with likely New Hampshire primary voters. The results suggest that a significant number of primary voters who are undecided turn to the state's dominant newspaper for direction concerning which candidate to choose. People who have already made up their minds seem less influenced by the paper's advice, whether it comes from the news pages or the editorial columns. Moore (1987) thought the paper's greatest influence may be when it attacks, but that does not appear to be the case in the 1996 GOP contest. Along these same lines, there is not much evidence of an antibandwagon effect on the candidates not favored by the paper (Farnsworth and Lichter 1999).

Conclusion

The 1992 and 1996 primaries marked the first times that sufficient data existed to test effectively for in-state relationships between day-to-day poll standings and news content (cf. Farnsworth and Lichter 1999, 2002). While our previous research studies suggested a powerful news influence for network television, this project demonstrates that the patterns of influence can vary considerably depending on the political party and media outlet being studied. Research on subsequent New Hampshire primaries could help buttress the general conclusions here of sometimes powerful news influence in New Hampshire, particularly as it relates to the Union Leader (Buell 2000; Moore 1987; Palmer 1997; Veblen 1975). Future researchers might consider whether the rise of the Internet has created new sources of media influence in the Granite State, or whether the 2000 primary voters were attending largely to the in-state media outlets in cyberspace.

Further research into the New Hampshire primary might consider the extent to which other campaign factors, including paid candidate advertising and the number of campaign appearances, may interact with news media coverage to affect public evaluations of candidates in the primary. Such questions were beyond the reach of this study and of nearly all other past research on the primary but may represent fruitful areas for future study of this pivotal nomination contest.

The findings here also suggest additional areas for further study that relate to the relative standings of particular candidates. Will the patterns of other candidates who receive very little directional coverage show unexpected effects similar to that found in the Alexander analysis? And will future viable conservative challengers endorsed by the paper trigger a "rally round the conservative" effect in future primary contests?

Although this research is not primarily about New Hampshire's fitness to be the first primary state, the results here can contribute to the ongoing debate over whether there is anything special about New Hampshire. Does this study help strengthen the state's claim that it should go first because of its smallness, which its supporters say allows for face-to-face retail politics? The 1992 Democratic results suggested that New Hampshire was a state heavily influenced by network news, a contest that differed little from media-dominated campaigns found throughout America (Farnsworth and Lichter 1999, 2002). This study of the 1996 contest suggests that the state's quirky and biased leading paper is particularly influential in Republican contests. Although one should not make a conclusive argument against New Hampshire's primary status based on two primaries, the results here hardly strengthens the case for keeping the first primary in the Granite State.

Indeed, the 1996 results here suggest, as did the 1992 findings, that the small-town, church-supper vision of the New Hampshire primary may be more of an ideal than a reality. Media influences are important predictors of the behavior of the New Hampshire electorate, be they Democrats or Republicans, in these contested primaries. This is exactly the complaint many researchers make about national elections (Patterson 1990, 1994). Simply put, New Hampshire in 1996, as it was in 1992, was not as dominated by face-to-face retail politics--and not as immune to media influences--as the first primary's supporters would like to think.

Appendix News Story Content Coding

Substantive topics:

Candidate personal background/character

Candidate professional background/qualifications

Substantive policy issues

Candidate campaign conduct

News media coverage

Campaign debate performance

Horse-race topics:

Campaign horse race

Candidate strategies and topics

"Campaign horse race stories report on the progress and prospects of the campaign horse race. These could include poll results, predictions of electoral performance, or assessments of a candidate's chances. It may also include characterizations of the candidate's strength, popularity, standing with the public, or other specific advantages or disadvantages of the candidate. This could include overall assessments of a candidate's chances or speculation over the impact of the day's events on the candidate's relative standing. Handicapping of the candidates, predictions of what might happen or other such statements would also be included here" (Farnsworth and Lichter 2002, 83).

"Candidate strategies and tactics stories focus on describing the plans, mechanisms, tactics, strategies and locations of a candidate's (or several candidates') presidential campaign. Describing which states, ethnic groups, interest groups, etc., that a campaign is targeting its appeals, or an analysis of why a particular candidate is stressing a certain issue, or using certain language in their campaigns, would be included in this category. Describing the effort that goes into planing a rally, speech or other event would be coded here' (Farnsworth and Lichter 2002, 83).

Source: Election 1996 Codebook (Washington, DC: Center for Media and Public Affairs, 1996). TABLE 1 Editorial Coverage of Candidates Buchanan Dole n % n % Horse-race evaluation positive 75 38 21 13 Horse-race evaluation negative 6 3 30 19 Substantive evaluation positive 88 44 11 7 Substantive evaluation negative 31 16 95 61 Total directional evaluations 200 101 157 100 Alexander Forbes n % n % Horse-race evaluation positive 2 9 18 20 Horse-race evaluation negative 2 9 9 10 Substantive evaluation positive 4 17 6 7 Substantive evaluation negative 15 65 55 63 Total directional evaluations 23 100 88 100 Source: Center for Media and Public Affairs. Data available from the authors upon request. Note: The totals are the summed directional content analysis for editorials and opinion columns during the survey period. Evaluations coded "mixed" or "neutral" are not included. Percentages may not add up to 100 because of rounding. TABLE 2 News Coverage of Candidates Buchanan Dole n % n % Horse-race evaluation positive 205 60 226 37 Horse-race evaluation negative 30 9 109 18 Substantive evaluation positive 68 20 137 22 Substantive evaluation negative 39 11 146 24 Total directional evaluation 342 100 618 101 Alexander Forbes n % n % Horse-race evaluation positive 78 34 189 31 Horse-race evaluation negative 36 16 89 14 Substantive evaluation positive 65 28 89 14 Substantive evaluation negative 52 23 252 41 Total directional evaluation 231 101 619 100 Source: Center for Media and Public Affairs. Data available from the authors upon request. Note: The totals are the summed directional content analysis for the Union Leader's news articles during the survey period. Evaluations coded "mixed" or "neutral" are not included. Percentages may not add up to 100 because of rounding. TABLE 3 Ordinary Least Squares Regression Analysis: Union Leader Coverage and Tracking Poll Standings Buchanan Dole b SE b SE News, horse race 0.32 ** 0.09 0.11 0.09 Editorial, horse race -0.14 0.16 -0.23 0.24 News, substantive evaluations -0.74 *** 0.20 -0.20 0.13 Editorial, substantive evaluations 0.66 ** 0.21 0.07 0.12 Adjusted [R.sup.2] .60 *** .07 Alexander Forbes b SE b SE News, horse race 0.18 0.42 -0.003 0.05 Editorial, horse race -10.23 * 4.19 0.12 0.13 News, substantive evaluations -0.56 0.28 -0.04 0.04 Editorial, substantive evaluations -1.43 * 0.58 0.10 0.09 Adjusted [R.sup.2] .30 * .002 Source: Center for Media and Public Affairs; American Research Group Inc. Data available from the authors upon request. * p < .05. ** p < .01. *** p < .001.

References

Adams, William C. 1987. As New Hampshire goes ... In Media and momentum: The New Hampshire primary and nomination politics, edited by Gary R. Orren and Nelson W. Polsby. Chatham, NJ: Chatham House.

Bartels, Larry M. 1985. Expectations and preferences in presidential nominating campaigns. American Political Science Review 79:804-15.

Bartels, Larry M., and C. Anthony Broh. 1989. The polls--A review: The 1988 presidential primaries. Public Opinion Quarterly 53:563-89.

Brady, Henry E., and Richard Johnston. 1987. What's the primary message: Horse race or issue journalism? In Media and momentum: The New Hampshire primary and nomination politics, edited by Gary R. Orren and Nelson W. Polsby. Chatham, NJ: Chatham House.

Brereton, Charles 1979. First step to the White House: The New Hampshire primary 1952-1980. Hampton, NH: Wheelabrator Foundation.

--. 1987. First in the nation: New Hampshire and the premier presidential primary. Portsmouth, NH: Peter E. Randall.

Buell, Emmett H., Jr. 1987. "Locals" and "cosmopolitans": State newspaper coverage of the New Hampshire primary. In Media and momentum: The New Hampshire primary and nomination politics, edited by Gary R. Orren and Nelson W. Polsby. Chatham, NJ: Chatham House.

--. 2000. The changing face of the New Hampshire primary. In In pursuit of the White House 2000: How we choose our presidential nominees, edited by William G. Mayer. Chatham, NJ: Chatham House.

Ceaser, James and Andrew Busch. 1993. Upside down and inside out: The 1992 elections and American politics. Lanham, MD: Rowman & Littlefield.

--. 1997. Losing to win: The 1996 elections and American politics. Lanham, MD: Rowman & Littlefield. Denton, Robert E., Jr. 1998. Communication variables and the dynamics of the 1996 presidential campaign. In The 1996 presidential campaign: A communication perspective, edited by Robert E. Denton Jr. Westport, CT: Praeger.

Dover, E. D. 1998. The presidential election of 1996: Clinton's incumbency and television. Westport, CT: Praeger.

Duncan, Dayton. 1991. Grass roots: One year in the life of the New Hampshire presidential primary. New York: Viking.

Farnsworth, Stephen J., and S. Robert Lichter. 1999. No small town poll: Network coverage of the 1992 New Hampshire primary. Harvard International Journal of Press/Politics 4 (summer): 51-61.

--. 2002. The 1996 New Hampshire Republican primary and network news. Politics & Policy 30 (1): 70-88.

--. 2003. The nightly news nightmare: Network television's coverage of U. S. presidential elections, 1988-2000. Lanham, MD: Rowman & Littlefield.

Graber, Doris. 1988. Processing the news. New York: Longman.

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

John, Kenneth E. 1989. The polls-A report: 1980-1988 New Hampshire presidential primary polls. Public Opinion Quarterly 53 (winter): 590-605.

Just, Marion R. 1997. Candidate strategies and the media campaign. In The elections of 1996: Reports and interpretations. Chatham, NJ: Chatham House.

Mayer, William G. 1987. The New Hampshire primary: A historical overview. In Media and momentum: The New Hampshire primary and nomination politics, edited by Gary R. Orren and Nelson W. Polsby. Chatham, NJ: Chatham House.

--. 1997. The presidential nominations. In The elections of 1996: Reports and interpretations. Chatham, NJ: Chatham House.

Moore, David. 1984. The death of politics in New Hampshire. Public Opinion 7 (February-March): 56-57.

--. 1987. The Manchester Union Leader in the New Hampshire primary. In Media and momentum: The New Hampshire primary and nomination politics, edited by Gary R. Orren and Nelson W. Polsby. Chatham, NJ: Chatham House.

Orren, Gary R., and Nelson W. Polsby, eds. 1987. Media and momentum: The New Hampshire primary and nomination politics. Chatham, NJ: Chatham House.

Owen, Diana. 1997. The press' performance. In Toward the millennium: The elections of 1996, edited by Larry J. Sabato. Needham Heights, MA.: Allyn & Bacon.

Palmer, Niall A. 1997. The New Hampshire primary and the American electoral process. Westport, CT: Praeger.

Patterson, Thomas E. 1980. The mass media election: How Americans choose their president. New York: Praeger.

--. 1994. Out of order. New York: Vintage.

Peirce, Neil R., and Jerry Hagstrom. 1983. The book of America: Inside the 50 states today. New York: Norton.

Picard, Robert G. 1998. Media concentration, economics and regulation. In The politics of news the news of politics, edited by Doris Graber, Denis McQuail, and Pippa Norris. Washington, DC: Congressional Quarterly.

Robinson, Michael J. 1978. TV's newest program: The "presidential nominations game." Public Opinion 1 (May-June): 41-46.

--. 1985. Where's the beef? Media and media elites in 1984. In The American elections of 1984, edited by Austin Ranney. Durham, NC: Duke University Press.

Rosenstiel, Tom. 1994. Strange bedfellows: How television and the presidential candidates changed American politics, 1992. New York: Hyperion.

Rueter, Theodore. 1988. Reflections on the New Hampshire primary. PS: Political Science and Politics 21 (spring): 273-77.

Sanford, Bruce W. 1999. Don't shoot the messenger: How our growing hatred of the media threatens free speech for us all. New York: Free Press.

Schudson, Michael. 1978. Discovering the news. New York: Basic Books.

Sprague, Stuart. 1984. The New Hampshire primary. Presidential Studies Quarterly 14 (Winter): 127-31.

Tabachnick, Barbara G., and Linda S. Fidell. 1989. Using multivariate statistics. 2d ed. New York: HarperCollins.

Trent, Judith S. 1998. The beginning and the early end. In The 1996 presidential campaign: A communication perspective, edited by Robert E. Denton Jr. Westport, CT: Praeger.

Veblen, Eric P. 1975. The Manchester Union Leader in New Hampshire elections. Hanover, NH: University Press of New England.

White, Theodore H. 1970. The making of a president 1968. New York: Pocket Books.

--. 1973. The making of a president 1972. New York: Bantam Books.

Stephen J. Farnsworth is associate professor of political science and international affairs at Mary Washington College in Fredericksburg, Virginia. He is the coauthor, with S. Robert Lichter, of The Nightly News Nightmare: Network Television's Coverage of U.S. Presidential Elections, 1988-2000 (Rowman & Littlefield, 2003).

S. Robert Lichter is president of the Center for Media and Public Affairs, a nonpartisan media research organization in Washington, DC. He is the coauthor, with Stephen J. Farnsworth, of The Nightly News Nightmare: Network Television's Coverage of U.S. Presidential Elections, 1988-2000 (Rowman & Littlefield, 2003).

AUTHORS' NOTE: The authors thank Mary Carroll Willi and Richard Noyes of the Center for Media and Public Affairs, Dick Bennett of American Research Group Inc., Dr. Carole Kennedy of San Diego State University, and the referees and editors for their assistance and comments. Thanks to Mary Washington College for its financial support and to Edwin Egee V for research assistance. An earlier version of this article was presented at the 2000 annum meeting of the Western Political Science Association. Any errors are the responsibility of the authors.