Content-Type: text/html

                Campaign Ad Coverage


Submitted to the MacDougall Student Award




   Campaign Advertising Coverage in the 1990s Elections
                                                    : A Content Analysis




       Running Head: Campaign Ad Coverage





                                            Submitted
                                                   by

                                          Young Min

           Doctoral Student
                                                            School of Journalism
      College of Communication
                 University of Texas at Austin


                                             Address: 3373 Lake Austin Blvd. Apt. # D.
            Austin, TX 78703


                                                      <a href="/cgi-bin/wa?LOGON=A3%3Dind0109B%26L%3DAEJMC%26E%3D8bit%26P%3D8428689%26B%3D--%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D_171707671%253D%253D_%26T%3Dtext%252Fplain%3B%2520charset%3Diso-8859-1" >[log in to unmask]</a>
                                                   Telephone:  (O) 512-471-1926
                                                                       (H) 512-708-9201


     Submitted to the Newspaper Division and the MacDougall Student Award
                                                     of the 2001 conference
               of the Association for Education in Journalism and Mass Communication






    Campaign Advertising Coverage in the 1990s Elections
                                            : A Content Analysis


              ABSTRACT

This paper explores the discursive patterns and styles of campaign advertising coverage. Specifically, it examines how the news sets contextual frames for political ads, attending to how two prestigious newspapers--the New York Times and the Washington Post--covered the 1992, 1996, and 2000 general-election advertising campaigns. An analysis of 118 ad stories indicates that ad coverage in the 1990s has paid more attention to challengers than incumbents, to presidential than state or local races, and to negative than positive ads. While employing an investigative and research driven style of reporting, the press has applied a "double-standard" to the assessment of political ads; it has tended to deflate the authenticity of campaign ads, but more often than not it has reinforced the causes of the campaigners concerning the political effectiveness of those ads. Most importantly, the press has exhibited a Republican bias in coverage and a Democratic bias in tone in reporting on advert
ising campaigns. Overall, the campaign ad coverage in the 1990s has shifted its attention from the effectiveness of the charges and countercharges to their accuracy, focusing more on the substance of candidates' issue positions. This shift may encourage candidates to engage each other with the matters that are more essential and relevant to governance.




Campaign Advertising Coverage in the 1990s Elections
                                                  : A Content Analysis


Many scholars agree that negativity has been a commanding attribute of both political journalism and campaign advertising (Ansolabehere and Iyengar, 1995a; Jamieson, 1992; Patterson and McClure, 1976). On the former claim, some go as far as to say that American journalism is "out of order." For example, Patterson (1994) contends that campaign coverage, governed by a deeply cynical view of politics and politicians, threatens the media's ability to make a constructive contribution to the electoral dialogue. On the latter claim, political advertisers have also been regarded as a negative industry. According to Finkel and Geer (1998), negative political ads, especially negative presidential ads, have continually increased since 1976. As politics gets more and more personalized in media-centered campaigns, discrediting an opponent has become a more popular technique for attracting voters.
Campaign ads, meanwhile, have recently attracted great attention from the news media, and this increase in campaign ad coverage has blurred the traditional division between the paid and the unpaid media (Ansolabehere and Iyengar, 1995b; Jamieson, 1992; West, 1993). Noting this growing interaction between the two media, this paper aims at exploring the patterns and styles in news discourse concerning advertised candidate messages.
According to Kendall (2000), campaign advertising has spawned negative coverage. "Ad watch" features, a new genre of campaign journalism popularized through the 1990s elections, have particularly contributed to the negative aura of political journalism by dissecting political ads and exposing lies and distortions in them. The regular reporting of political ads, indeed, focuses more on negative spots than on positive ones, thus convincing candidates of the advantage of negative appeals in gaining free coverage. Campaign ad coverage, which embodies this "spiral of negativity" between paid and unpaid campaign communication, however, has not attracted much scholarly attention. Campaign advertising is one of the most important forms of direct candidate communication (along with political speeches and debates), and campaign journalism has increasingly found a newsworthiness in candidates' paid communications, yet few studies have illuminated the nature of the news coverage of political
advertising.
The news, at the same time, sets the contexts for candidates' paid communication, while ensuring that more citizens will be exposed to the ads. As Jamieson (1992) points out, news reporting provides a "frame" through which viewers can interpret an ad. If the news accounts are inconsistent with the ad, the power of the ad will be decreased. If the news commentaries reinforce the ad, in contrast, the persuasiveness of the ad message will increase (Ansolabehere and Iyengar, 1994; Jamieson, 1992). It is important, therefore, to understand the contexts within which a campaign ad is being presented in the news in assessing the electoral success of the paid communication. Considering that more and more people have become about as likely to see ads through the news as to see them directly (West, 1993), it can be deemed even more valuable to investigate what kinds of ads get into the news arena and how journalists portray these ads.
In analyzing campaign ad coverage, this study also attempts to provide subsequent effects studies with at least a rough guide for investigating how the news media's recirculation and recontextualization of campaign ads influence the impact of the ads per se on the electorate. Specifically, the present study will focus on campaign advertising mediated by prestigious national newspapers which best embody the "new long journalism," by which Barnhurst and Mutz (1997) refer to a highly analytic, complex, and interpretive style of news reporting.[1]
                   Recent Studies of Campaign Advertising Coverage
News discourse relative to direct candidate communication has displayed its own unique styles and patterns. In her analysis of network news coverage of candidate speeches in the 1992 primaries, Kendall (1993) observes that the news media "regularly reduced the candidates to gesturing, voiceless figures, the power of their language and ideas lost in reporters' summaries" (p. 131). Jamieson (1992) reports that "post-debate coverage focuses almost exclusively on the candidates' debating styles and on how their performance would affect the horse race" (p. 194). Yet the nature of the news coverage of candidates' communicative performances in general, and their paid communications in particular, lacks due attention, thus constituting a lacuna in the campaign journalism literature.
Media attention to political advertising greatly increased in the 1992 presidential election, where for the first time the ads were regularly scrutinized by the news media for accuracy and effectiveness (Kendall, 2000). Until 1992, the news media had focused mainly on the potential impact, that is, the effectiveness of campaign commercials, rather than monitoring the accuracy or fairness of the ads (Jamieson, 1992). Buchanan (1996) explains this intensification of media's attention to advertising as a reformist reaction to the corrupt, dirty ad campaigns in the 1988 presidential election, emphasizing that it discourages candidates' outright deception to subject their messages to the electorate to close scrutiny.
Swanson's (1997) account of American campaign journalism helps to better understand the news media's growing interest in campaign ads. According to Swanson, while focusing on campaign horse races and dramatizing electoral contests, journalists have demonstrated their resistance against politicians' manipulation through an interpretive style of campaign reporting, where reporters actively mediate the candidates' words and actions for the audience. That is, political advertising has been a contemporary focus of this interpretive style of reporting which has governed American journalism since the 1960s.
Semetko and her colleagues (1991) assert that this kind of intrusive reporting has more often than not employed negative approaches. Their analysis of the 1984 U.S. presidential election coverage shows that journalists contextualize candidates' conduct more frequently in deflating tones than in reinforcing ones. As Patterson (1994) observes, the majority of negative coverage consists of journalistic evaluative remarks about candidates' ideas, job performance, issue positions, and personal characteristics, rather than straight reports on candidates' negative campaigning. These findings enable us to expect that negative valence would prevail in campaign ad coverage.
The news media, indeed, tend to frame direct candidate communication through debates, speeches, and ads, based on the "strategy schema" rather than on "the problem-promise-performance approach," according to Jamieson (1992). She argues that the so-called "who-won-the-game" frame places its attention on the presumed intent or effectiveness of a candidate's communicative performance, rather than on its legitimacy or relevance to governance. Ad-watch features in the news media, however, have attempted to supervise campaign ads in terms of accuracy, fairness, or legitimacy. It would be interesting, therefore, to explore whether ad-watch kinds of news stories in the 1990s provide a different framework other than the "strategy/game" schema. The contexts that news provides in reporting campaign ads are important, since those contexts can determine whether news simply lends unpaid access to campaign ads, legitimizing possibly deceptive ads with no treatment of their accuracy, or it punish
es the campaign that has delivered misleading messages to the voters, providing them with correct information.
In short, how the news sets contextual frames for political commercials is an important inquiry in analyzing campaign ad coverage. More specifically, these contextual frames include, for example, who is commenting about the ad, whether the commentaries deflate or reinforce the ad message, and which criteria are employed in assessing the ad--informative accuracy, political effectiveness, or something else.
What kinds of candidate ads attract the news media's free attention would be another important question. Jamieson's (1992) study on advertising and news coverage in the 1988 and 1990 election campaigns shows that news tends to feature oppositional, rather than self-promoting candidate ads, thus skewing the public perception of the negativity of campaigns. More concretely, ads appearing in network news were nearly twelve times more oppositional than self-promotional. The print media were six times more likely to mention an oppositional than a self-promotional ad. In addition to the tone or the type of strategies of appeals, what other attributes of campaign advertising can determine its attractiveness to the news media?
Few studies have pursued this inquiry. Jamieson (1992) indicates that network news can be drawn to such factors as "visuality," "editability," and "humorousness." This may not be the case for the print media, which are the focus of this study. In analyzing all the stories about campaign ads appearing on the "CBS Evening News" from 1972 through 1992, West (1993) examines the content of the ad references (i.e., the general topics those ads address) such as personal qualities, domestic performance, specific policy statements, international affairs, campaigns, party, etc. Those ads reported on in the network news refer more often to specific policies than to any other topics.
Building on West's findings, this paper will investigate campaign ad coverage in terms of the type of the ads appearing in the news (i.e., issue-focused, image-focused, dovetailing, etc).[2] It will also subsequently examine what kinds of policy issues or candidate image attributes the ads deal with. Additional attention can be concentrated on the characteristics of the candidates sponsoring the ads in the news, such as their partisan identifications, the type of office they are seeking (presidential, Congressional, gubernatorial, etc.), the status of their candidacy (incumbent, challenger, open-seat, etc.), and so on.  More analytically, what kinds of ads (or whose ads) are more likely to lead to deflating tones (or reinforcing tones) of news commentaries will also be investigated.
Since 1968, Jamieson contends, there has been a change in the relationship between news and political advertising; that is, "Where news once contextualized ads, visually evocative and easily edited oppositional ads backed by candidate speeches and pseudo-events now have the capacity to shape the language and pictures of news" (p. 135). Several inter-media agenda setting studies have supported her point by demonstrating that the advertising media shape the news agenda of campaign issues or candidate image attributes rather than vice versa (Lopez-Escobar et al., 1998; Robert and McCombs, 1994).
The grammar of evaluation developed by the 1990s' ad-watch journalism, however, also has had a significant impact on candidates' strategies in their paid communications (West, 1993), and this indicates that there could have been another reversal in the relationship between the free and the paid media. This study may not provide decisive evidence supporting this "re-reversal," if, indeed, any exists. A thick description of more recent trends in campaign ad coverage, however, can provide at least an approximate picture of how the two channels of campaign communication now interact with each other.
 Methods
As Riffe et al. (1998) point out, the external validity of content analysis can be increased if the content being explored is important. The increasing pervasiveness and importance of ad coverage as a new genre of campaign journalism have been noted in the above discussions. In this regard, this study content-analyzed campaign advertising coverage in the 1992, 1996, and 2000 general election campaigns. This time frame was selected because a more interpretive and critical style of advertising reporting has been developed since the 1992 campaign, when ad-watch features first appeared in the news media. Although this study focused mainly on the general election coverage, it would be interesting for a future study to extend the time frame to the primaries period since, as Kendall (2000) indicates, the primary coverage has its own unique patterns and styles different from those of the general election coverage.
News stories published from the first of September to the eve of Election Day in the New York Times and the Washington Post, two prestigious elite newspapers which best embody a highly analytic, complex, and interpretive style of news reporting, were consulted. According to Buchanan (1996), more than half of the 200 ad watch stories in the sample for his analysis of the 1992 election coverage were from these two newspapers.
All news articles which contain the word "ad" in their headlines were collected via the Lexis-Nexis database. A total of 207 were collected from the New York Times, while a total of 96 came from the Washington Post. Using a stratified random sampling method, about 50 percent of the initially retrieved articles were selected, and, among these, a total of 32 news stories were found to be irrelevant to advertising campaigns and were excluded. A total of 118 articles were finally included in the sample for this particular content analysis.[3]
The entire news story was consulted as both the unit of context and the unit of analysis. Coding categories were devised to follow from these research focuses: 1) the characteristics of the sponsoring candidate especially in terms of partisan affiliation, the office sought, and the status of the candidacy; 2) the tone (negative, positive, etc.)[4] and type of the ad, and the type of issues or candidate image attributes addressed in the ad; 3) the identities of sources commenting about the ad (journalists, academics, sponsoring campaigns, opposing campaigns, other experts, etc.); 4) the criteria used for assessing the ad (accuracy, effectiveness, etc.); 5) the tone of the news assessment (deflating, reinforcing, neutral, etc.)[5] (For more details, see Appendix I). Holsti's reliability coefficients for the various coding categories, meanwhile, ranged from 0. 78 to 1.00.[6]
When appropriate (for instance, if the ad in a news story dealt with more than one issue or image attribute which were judged to be almost equally emphasized), multiple-coding was allowed. Content analysis data were computerized so that various statistical patterns could be examined. In addition to this quantitative analysis of campaign ad coverage, its qualitative aspects were also explored, primarily in terms of what kinds of lexicons of the press reviews of political commercials feature negative or positive tones.
                                                     Results
Whose paid communication is more likely to be reported on in the news?
The data indicate that the ad-watch coverage focused on the mainstream candidates with little access for minor parties, as Robinson and Sheehan (1983) suggest. Among a total of 120 cases analyzed, only 8 percent of campaign ad coverage reported on the ads of Independent candidates or special interest groups. Ross Perot's ads in 1992 and 1996 mainly constituted the small amount of Independent candidate coverage. Special interest groups included labor unions, the Coalition for Voters' Choice to change New York City's term limit law (1996), supporters of a proposed $ 3.8 billion transportation bond act in New York (2000), and so on.
The remainder of the coverage concentrated on the two major parties' candidates: Republican ads (53 percent), Democratic ads (33 percent), and both Republican and Democratic ads (6 percent). Apparently, Republican candidates were more likely to get free media coverage on their campaign ads than were their Democratic counterparts, and this Republican bias was also consistent within the presidential ad coverage, as illustrated in Figure 1. If, as Ansolabehere and Iyengar (1995b) claim, mere exposure to ad-watch stories per se strengthens the viewers' preferences toward the sponsoring candidates, regardless of the tone of the ad-watch analyses, the "partisan bias" in the sheer amount of ad coverage may affect the electoral outcome. This hypothesis, however, needs more
rigorous empirical tests, which are beyond the scope of this study.[7] As discussed later, the
Figure 1. Partisanship of the Sponsoring Candidates of the Presidential Ads in the News

"partisan imbalance" in the quantity of ad coverage seems to have been counterbalanced
 by another "political imbalance" in the tone of the press commentary.
The status of the candidacy of the sponsoring candidate also influenced the amount of media attention. As displayed in Table 1, challengers' ads were more likely to attract unpaid coverage than were those of incumbents or open-seat candidates. As Robinson and Sheehan (1983) contend, the news media may attend more to challengers in their efforts to compensate for the presumed incumbent advantages.
Taking these insights together, it can be hypothesized that Republican candidates might benefit from their status as challengers, not from their partisan affiliation. This hypothesis, however, was not substantiated; although Republicans were featured as challengers more frequently than were Democrats in the ad coverage, the difference was not significant ((ý (1, N = 74) = 1.216, p = .270). Since there have been two Republican
     Table 1. Status of the Candidacy of the Sponsoring Candidates by Publication
    Publica-
             tion
Candidacy
New York
   Times
Washington
     Post

    Total

 Incumbent

 Challenger

 Open-Seat

    20.0 %

    38.8

    34.1

    37.1 %

    51.4

      5.7

   25.0 %

    42.5

    25.8

 Total (N =)

       85

       35

    120











challengers in the three presidential elections in the 1990s, however, the Republican bias in the presidential ad coverage might be related to the Republicans' status as challengers.
Consistent with the conventional wisdom that political journalism is execucentric, presidential ads (49 percent) outnumbered other levels of campaign ads which included senatorial (41 percent), Congressional (5 percent), gubernatorial ads (2 percent), and so on. The New York Times' heavy emphasis on the State of New York Senatorial elections, such as the close Clinton-Lazio election of 2000, however, seems to have offset the presidentially-skewed ad coverage to some degree. Overall, however, as Dunn (1995) points out, the decline in press coverage from prominent to down-ballot offices was substantial, resulting in almost no press coverage of state-level elections in these elite newspapers.
What kinds of ads are more likely to attract media coverage?
        As noted by Jamieson (1992) and Kendall (2000), the news media were more likely to be drawn to "negativity" than to "positivity." Sixty percent of the ad coverage mediated negative commercials, while 33 percent of it focused on positive spots. This pattern was consistent across both news organizations, as shown in Figure 2. Recalling that the print media were six times more likely to mention an oppositional than a self-promotional ad in the 1988 and 1990 elections (Jamieson, 1992), the negativity-skew seems to have considerably decreased in the ad coverage of the 1990s, when negative ads have been only twice as likely to appear in the news as have positive ads.
        Validating West's (1993) observation that ads in the news refer more often to specific policies than to other considerations, this analysis also shows that issue-focused advertising was more frequently "recirculated" in the free channels than its image-focused counterpart. Sixty-four percent of the ads covered by the two prestigious

  Figure 2. The Valence of the Campaign Ads in the News (by Publication)

newspapers featured specific issues, while only 19 percent of them attacked or self- promoted personal qualities or traits, such as leadership, credibility, commitment to the people, integrity, etc. Nine percent of the ads in the news were found to dovetail issues and images in one format. In summary, negative issue advertising was the most frequently appearing in the campaign ad coverage of the 1990s, as Figure 3 outlines.
        If Republicans relied on "negativity" or "substantiality" in their ads more frequently than did Democrats, the Republican bias in the amount of ad coverage may have originated not from the partisan attribute of the sponsoring candidates but from the negative tone or issue-focused nature of the Republican ads. This hypothesis, however, was not confirmed; although Republicans were more likely to employ negative appeals

             Figure 3. The Type and Valence of the Campaign Ads in the News

than were their opponents, the difference was not substantial ((ý (1, N = 98) =2.388, p = .122). Republicans, indeed, were found to be slightly less issue-oriented than were Democrats although the difference was not significant ((ý (1, N = 88) = 2.030, p = .154).
The top three topics which were most frequently addressed in the issue-focused ads, meanwhile, were "the budget and taxes" (21 percent), "health care" (14 percent), and "the economy" (12 percent)--traditionally the most controversial campaign issues. Credibility, meanwhile, was more frequently mentioned in the image-focused ads than any other personal attributes.[8]
Who are the referees evaluating the campaign ads?
Among a total of 120 cases, nearly 93 percent (n = 112) of the news stories assessed a campaign ad in terms of accuracy, effectiveness, or both accuracy and

          Figure 4. Identities of Sources Commenting About the Ads in the News

effectiveness, rather than solely reflecting the advertise candidate messages. Only 7 percent of the ad coverage reported political ads simply as usual campaign activities without any evaluative comments. These findings imply that campaign ad journalism employs a largely interpretive and analytic style of reporting. More precisely, the ad-watch coverage embodies an "investigative" and "research-driven" style of reporting, in that it often conducts in-depth analyses (or comparisons) of candidates' issue positions or other claims, in scrutinizing the truthfulness of their ad messages, as discussed later.
        As suggested in Figure 4, the analysis of the identities of the sources commenting about the ads in the news reveals that the majority of the ad coverage contained journalistic contextualizing remarks. Among a total of 152 identified commentators,
journalistic sources (N = 83) accounted for 55 percent. Even further, the news articles where journalists directly contextualized campaign ads in some ways accounted for more than 70 percent of the "analytic and interpretive" ad coverage.[9] Again, this 70 percent coverage where reporters interpreted the ads for the audience can be deemed more investigative and more research driven than the remaining coverage where partisan or expert sources, other than journalists, function as the main commentators on the ads.
Excepting journalistic sources, the opposing campaigns were the most frequently cited commentators in the ad coverage. Here, news tends to place campaign ads in negative and strategic contexts by quoting the opposing campaigns' commentaries more frequently than the sponsoring campaigns' or other experts'.
What criteria do the news media use to assess campaign ads?
According to Jamieson (1992), the news media had reviewed political ads mainly in terms of effectiveness until they began to monitor campaign ads for accuracy in 1992.
          Figure 5. Criteria for Ad Assessment in the Campaign Ad Coverage

The analysis of the criteria for ad assessment employed in the news reveals that the press has monitored campaign ads more frequently for both accuracy and effectiveness than for accuracy or for effectiveness only. As Figure 5 depicts, 28 percent of the ad-watch stories focused primarily on the accuracy, fairness, or legitimacy of the ad messages under review, while 14 percent of the reports employed the electoral successfulness of the ads as their evaluative yardsticks. The press sometimes utilized other criteria for assessment, such as nastiness, subliminality, or campaign spending. For instance, the September 9, 2000 edition of the New York Times focused on the increasing vitriol in both the Clinton and Lazio campaign ads. The October 6, 2000 edition of the Times pointed out that Lazio's new ad, sponsored by the Republican National Committee, violated his soft-money spending agreement with the Clinton campaign. A Times ad watch on September 13, 2000, dealt with various comments
 about the subliminal message of the "RATS" commercial, the RNC ad criticizing Al Gore's health care plans.
This pattern was also evident when the analysis was limited to the presidential ad coverage, although slightly more attention was paid to the "accuracy" aspects of the presidential candidates' ads: accuracy (31 percent), effectiveness (15 percent), and both accuracy and effectiveness (36 percent).
How do the journalists contextualize campaign ads?

A further analysis looked at the tone of the press commentaries in terms of each criterion (i.e., accuracy or effectiveness). In monitoring the accuracy of campaign ads, the overall tone of the press reviews was negative. As Table 2 informs, the majority of the news stories which dealt with the accuracy of campaign ads deflated the causes of the
sponsoring candidates (i.e., suspected inaccuracies and possible deceptiveness in the paid candidate communications). The negative tone of the ad coverage in monitoring the accuracy of the campaign ads was featured by the following terms: deceptive, erroneous, illegitimate, improper, inappropriate, misleading, ridiculous, uncertain, unfair,

Table 2. The Tone of the Press Commentary by Main Evaluative Criteria
                   Accuracy
                 Effectiveness
1992
1996
2000
Total
1992
 1996
2000
Total

Deflating

Reinforcing

Neutral

50.1 %

26.9

23.1

72.7 %

0

27.3

50.0 %

20.6

29.4

56.1 %

17.1

26.8

 6.3 %

62.5

31.3

38.5 %

46.2

15.4

35.3 %

41.2

23.5

28.6 %

47.6

23.8
Total (N =)
  26
   22
   34
   82
  16
   13
   34
   63

etc., as adjectives; distort, dwarf, exaggerate, ignore, misrepresent, oversimplify, overstate, etc., as verbs; glancing blows upon the truth, many questions, minimal exactitude, a perversion of truth, unprecedented sleaziness, etc., as noun phrases. As such, highly evaluative lexicons were featured in the ad-watch coverage, and such evaluative components might function as information shortcuts for the people who delegate their judgmental tasks to the media in elections, as Popkin (1991) posits.
Only 17 percent of the news accounts, meanwhile, reinforced the accuracy of the ad messages. The remaining coverage presented the political ads as "only partially accurate" (i.e., accurate in some aspects but inaccurate in others), or offered competing comments from both the opposing campaigns and the sponsoring campaigns. In order to balance two different points of view, the news frequently relied on "semantic-moves" lexicons, e.g., if, although, but, however, etc. Overall, the accuracy assessment coverage conveyed highly substantial information by concentrating primarily on the comparisons of the competing candidates' issue positions.
In contrast, the tone of the news stories highlighting the political potential of campaign ads was much more positive (N = 63). Nearly 50 percent of the coverage reinforced the effectiveness of the political ads being scrutinized (i.e., acknowledged that those ads might successfully appeal to the voters). Thirty percent of the news articles presented the ads under review as politically incompetent or risky. The rest of the coverage presented both the political potential and risks the ads might hold or analyzed the hidden strategies pursued by the sponsoring candidates through their ad campaigns.
In commenting on the political potential or risks of the ads under review, the press placed itself as a campaign consultant. The majority of the analyses of the effectiveness of the ads were "strategically" framed, in that they focused on the hidden motives or expectations of the sponsoring campaigns or on how the ads might possibly advantage or disadvantage the horse races, as manifested in the following examples:

"The advertisement packs a tremendous amount of damaging information into 30 seconds, portraying Dole as an aged, obstructionist Washington insider_Viewers might regard the ad as informative rather than negative." (The New York Times, September 25, 1996: A18)

"U.S. Sen. Charles S. Robb (D) has_erased the lead held by Republican rival George Allen in Virginia's Senate race, gaining ground in a TV advertising onslaught that has reshaped the contest in a virtual dead heat." (The Washington Post, October 29, 2000: A01).

"The advertisement, which injects a touch of humor in what otherwise has been a largely humorless campaign, could help Mrs. Clinton win votes upstate." (The New York Times, October 29, 2000: A40)

"This ad does not once mention Mr. Zimmer. Instead, it touts Mr. Torricelli's endorsements_The switch to a more positive spot may reach some voters who have been turned off by the negative ads." (The New York Times, November 1, 1996: B6)

"While Mr. Bush may have the Republican's traditional advantage when it comes to tax-cutting, right now tax-cuts are not one of the top concerns of voters." (The New York Times, September 16, 2000: A10)

"Sarcasm is a dangerous weapon in commercials because if no humor is evident, sarcasm can come across as petty and malicious and turn off swing voters. The tone may also distract from the message_.The timing of this commercial suggests a certain desperation by Republicans." (The New York Times, September 1, 2000: A22)


The news sometimes indirectly reinforced or deflated the effectiveness of the ads, by portraying the sponsoring candidates' other campaign activities favorably or unfavorably, or by presenting the poll data on the candidates in the ad-watch stories.
Meanwhile, only a small portion of the ad coverage mediated campaign ads in neutral tones, and this is inconsistent with previous findings, which have suggested that the majority of campaign coverage is neutral in its valence (Just et al., 1999; Patterson, 1994; Robinson and Sheehan, 1983).
Which attributes of campaign ads affect the tone of the news commentaries on the ads?
The analysis of the kinds of campaigns ads that led to more negative or positive coverage suggests that the partisanship of the sponsoring candidate did significantly matter in determining the tone of the free coverage.
As Tables 3 and 4 report, there appeared to be a Democratic slant in the tone of ad assessments in the news, contrary to the Republican bias in the quantity of the ad coverage. Chi-Square analyses demonstrate that Democratic candidates were significantly more likely to receive favorable news analyses about their paid communications than were Republican candidates, both in terms of accuracy ((ý (2, N = 74) = 7.159, p < .05) and effectiveness  ((ý(2, N = 59) = 12.813, p < .01). This finding does not confirm previous observations that there is no clear evidence of the news media's political bias toward the Democratic Party or against the Republican Party, although most journalists are liberal in their personal beliefs (Hofstetter, 1976; Robinson

Table 3. Tone of Accuracy Assessment by Partisanship of the Sponsoring Candidate
        Partisanship
Accuracy

  Democrat

  Republican

     Total

Deflating

Reinforcing

Neutral


   41.9 %

   29.0

   29.0

     65.1 %

       7.0

     27.9

     55.4 %

     16.2

     28.4

Total

       31

       43

       74
 Table 4. Tone of Effectiveness Assessment by Partisanship of the Sponsoring Candidate
        Partisanship
Effectiveness

  Democrat

  Republican

     Total

Deflating

Reinforcing

Neutral


     9.1 %

    77.3

    13.6

     40.5 %

     29.7

     29.7

    28.8 %

    47.5

    23.7

Total

      22

       37

       59

and Sheehan, 1983; Weaver and Wilhoit, 1991). Patterson (1994), however, contends that whether there exists a systemic attempt by a liberal press against Republican candidates is still a question, pointing out that the vogue for interpretive reporting may encourage the unconscious forms of partisan bias.
The valence of the advertising (i.e., negative or positive), meanwhile, was found to affect the tone of the news commentary on the ad to a marginal degree. Another Chi-Square test shows that the news media presented positively appealing ads slightly more favorably than negatively appealing ads, both in terms of accuracy ((ý(2, N = 79) =
5.172, p = .075) and effectiveness ((ý(2, N = 57) = 5.829, p = . 054), although the differences were not statistically significant. The type of office sought (presidential, senatorial, Congressional, etc.) or the status of the candidacy (incumbent, challenger,
open-seat, etc.) was found to have little impact on the tone of subsequent news coverage.
This "partisan" imbalance in the tone of the ad coverage was also apparent when the analysis focused only on the presidential ad coverage. The ad-watch coverage was more critical toward the accuracy or legitimacy of the ads sponsored by Republican
   Figure 6. The Tone of Accuracy and Effectiveness Assessments on the Ads in the News
                                   by Partisanship of the Sponsoring Candidate



presidential candidates than that of the ads advocating Democratic presidential candidates ((ý(2, N = 37)  = 6.130, p  < .05). The presidential ad coverage also presented Republican ads as significantly less effective than Democratic ads ((ý(2, N = 27) = 10.795, p  < .01). Other attributes of presidential ads, such as ad valence and type, or the status of the candidacy of the sponsoring candidate, did not influence the tone of the press commentary.
This political bias in the campaign ad coverage should carefully be contextualized, since, if Republican candidates more frequently employed negatively appealing ads than did their Democratic rivals, they might suffer not from their partisanship but from their strategies of negative appeals. This hypothesis was, however, not supported in the present data; although Republicans preferred negative ads over positive spots slightly more than did Democrats, the difference was not significant ((ý(1, N = 98) = 2.388, p = .122).
Summary and Discussions
In analyzing the coverage of ad watches in the 1990s, this study has confirmed several previous observations on campaign journalism. First, the ad coverage was found to be primarily presidential, although the two prestigious newspapers allocated a considerable amount of space on key senatorial elections. State-levels of elections, overall, received almost no coverage. Second, negative advertising has been more likely to attract free coverage than has positive advertising. Negative ads in the news often dealt with substantial information on candidates' issue positions, rather than attacking the opposing candidates' personal qualities or traits. In this regard, the ritualized condemnation of negativity in political discourse among scholars and reporters should be cautioned, as Jamieson (1992) points out. Third, issue-focused ads have been more frequently "recirculated" in the news channels than have image-focused ads, confirming West's (1993) observation.
        The campaign ad coverage in the general elections of the 1990s was also found to employ a largely interpretive and analytic style of reporting. The substantial majority of the coverage made evaluative remarks on candidate ads in terms of accuracy, effectiveness, or both accuracy and effectiveness, and the majority of the remarks were comprised of journalists' analyses.
Consistent with West's (1993) finding, the news has tended to place the authenticity of advertised candidate messages in negative contexts. More often than not, however, it has reinforced the causes of the advertisers in assessing the electoral effectiveness of the ads. Unlike the conventional wisdom, the ad coverage has been featured in either negative or positive tones, rather than in neutral tones. In more than 50 percent of the news stories in which campaign ads were monitored for both accuracy and effectiveness (N = 44), meanwhile, the press commentary on the accuracy and that on the effectiveness of the ad under review were inconsistent with each other; in most cases, the press pointed to the inaccuracies in the ad, but acknowledged its political potential. This "double-standard" may weaken the power of the coverage of ad watches to depress deceptive or illegitimate paid campaigns.
Most importantly, the press in the 1990s elections was found to have both "access imbalance" and "tone imbalance" in reporting on advertising campaigns, with the partisan direction of one bias being the inverse of the other. Which bias might have more significantly affected the electoral effects of the campaign ads may depend on whether mere exposure to ad-watch reports can solidify the audiences' favorability toward the sponsoring candidates, or whether the ways in which journalists cover the ads can magnify or mitigate the persuasive power of the ad messages.
The imbalance in the amount of coverage can be seen in part as a structural bias, in that it might stem to some extent from the press's tendency to focus more on challengers than incumbents. This study, however, does not provide any empirical evidence as to "why challengers attract more free coverage."
The "tone imbalance," meanwhile, was more apparent in the press's reviews on the effectiveness of campaign ads than its accuracy checks; as Patterson (1994) points out, reporters' vision of politics as a strategic game may introduce random, unconscious partisanship into their campaign reporting. If, as Jamieson (1992) contends, news reporting provides a "frame" through which the audience interprets ads, "the Democratic skew" in the tone of the ad coverage may have a significant impact on the consumers of the political ads, regardless of whether it is a structural or an ideological bias. Since the "tone bias" was not correlated to the news media's inclination to focus more on challengers than on incumbents or to feature negative spots more often than positive ads, it might have little to do with the media's structural characteristics in newsgathering routines or journalistic professional norms (Hofstetter, 1976). Whether the "tone imbalance" observed in this study embodies the pres
s's ideological orientation also cannot be determined, however, until a much wider range of news media organizations are examined for their advertising coverage.
Overall, unlike the 1980s, the advertising journalism in the 1990s has shifted its attention from the effectiveness of the charges and countercharges to their accuracy, focusing more on the substance of candidates' issue positions. This move may encourage political candidates to engage each other with the matters that are more essential and relevant to governance.
[1] Notes.

 This study places its emphasis on the print coverage of campaign ads (perhaps, mostly
  television ads) rather than on television ad coverage. This is primarily because major
  studies of political ad journalism have focused on network ad watches rather than on
  print ad watches (Ansolabehere and Iyengar, 1995b; Jamieson, 1992; West, 1993). All
  these studies alike contend that the visual representation of campaign ads in television
  news overwhelms the critical analyses of the advertised messages per se, thus simply
  reinforcing the causes of the advertisers. The researchers point to this as one of the
  "unintended" ill effects of network ad watches. Newspaper ad watches which
  concentrate on the analysis of the contents of candidate ads, however, may play a
  different role, contributing more to the ultimate purpose of ad watches (i.e., monitoring
  candidates' paid communications for accuracy, fairness, or legitimacy). This is another
  reason why this study attends mainly to the campaign ad coverage in the print media.

[2]  Based upon Thorson et al.'s (1991) definitions, issue advertising is operationalized as
  "that referring only to general topics of political concern," and image advertising, as
  "the sum of the personal characteristics and professional abilities of the candidate" (p.
   467).

[3]  The initially retrieved news stories (n =  303) were stratified in terms of day of the
  week, since "cyclic variation of content for different days of the week" required that
  all different days of the week be proportionally represented (Riffe et al., 1993, p.
  134). Among the news articles grouped per day of the week, nearly 50 % of the
  stories were sampled using random numbers. Through this process, a total of 150 news
  stories were sampled. Among them, a total of 32 irrelevant articles, most of which
  reported on business ads, were identified and excluded; thus, a total of 118 stories were
  included in the final sample. This sample included two stories, each of which reviewed
  two ads with exactly the same portioning; each story was counted as two rather than
  one, and thus, the total number of cases under examination was 120. The following table
  displays the number of cases in the final sample per publication and per year. The
  number of news stories in the population appears in the parentheses.

                                     1992            1996             2000             Total
   New York Times       26 (63)         17 (42)        42 (102)         85 (207)
   Washington Post        9 (24)         13 (38)        13  (34)          35  (96)
              Total              35 (87)         30 (80)        55 (136)       120 (303)

[4]  As Buchanan (1996) suggests, in this study, negative advertising was understood as that
  focusing mainly on "comments on an opponent's personal traits, character, record, and
  policy positions" (p. 95). Similarly, positive advertising was operationalized as that
  focusing mainly on self-promoting comments on personal traits, character, record, and
  policy positions by the candidate him/herself.

[5]  These coding categories for the tone of press commentaries were borrowed from
  Semetko et al.'s (1991) content analysis of the tone of journalists' contextualizing
  remarks in campaign coverage.

[6]  About 8 percent of the news stories in the sample (N = 9) were randomly selected for
  the coder reliability test. According to Kaid and Wadsworth (1989), only 5 to 7 percent
  of the entire sample is enough for a reliability test (cited in Riffe, Lacy, and Fico, 1998).
  The second coder was given a brief coding instruction before analyzing the selected 9
  articles. Specifically, Holsti's reliability coefficients for the coding categories were
  came up with as follows: 1.00 for the partisanship, office sought, and status of the
  candidacy of the sponsoring candidate and for advertising tone;  0.89 for advertising
  type and for the criteria for ad assessment; 0.80 for the identities of sources commenting
  about the ad and for the tone of effectiveness assessment; and 0.78 for the tone of
  accuracy assessment.

[7]  In an unreported experimental study that the author conducted, the tone of the press
  commentary on the ad under review was found to influence individuals' evaluations of
  and their willingness to vote for the sponsoring candidate. More specifically, people
  exposed to the news analysis deflating the ad message were less likely to perceive the
  sponsoring candidate as credible (F (1, N = 88) = 9.299, p = .003) and less likely to vote
  for him/her (F (1, N = 88) = 6.314, p = .014) than were those exposed to the news
  analysis reinforcing the ad message. That is, the impact of ad-watch reports on the
  electorate may differ, depending on how journalists contextualize the ads. "How the ad
  is presented in the news," therefore, can be more important than "whose ad is covered."

[8]  A total of 117 issues were identified in the issue-focused ads while a total of 35 image
  traits were distinguished in the image-focused ads.

[9]  The "analytic and interpretive" ad coverage refers to the 112 news stories which
  evaluate campaign advertising in some ways. Among these, the articles which
  contain journalistic remarks (N = 83) account for 74.1 percent.
 References


Ansolabehere, S. D., and Iyengar, S. (1994). Riding the wave and claiming ownership
over issues: The joint effects of advertising and news coverage in campaigns. Public Opinion Quarterly, 58: 335-357.

Ansolabehere, S. D., and Iyengar, S. (1995a). Going Negative: How Political
Advertisements Shrink and Polarize the Electorate. New York: Free Press.

Ansolabehere, S. D., and Iyengar, S. (1995b).Winning through advertising: It's all in the
context. In J. A. Thurber and C. J. Nelson (Eds.), Campaigns and Elections American Style (pp. 101-111). Boulder, Colorado: Westview.

Barnhurst, K. G., and Mutz, D. C. (1997). American journalism and decline in event-
centered reporting. Journal of Communication, 47: 27-53.

Buchanan, B. (1996). Reviewing Presidential Politics: Campaigns, Media and the Public
Interest. Lanham, Maryland: Rowman & Littlefield.

Cappella, J. N., and Jamieson, K. H. (1996). Spiral of Cynicism: The Press and the
Public Good. New York, NY: Oxford University Press.

Dunn, A. (1995). The best campaign wins: Local press coverage of nonpresidential races.
In J. A. Thurber and C. J. Nelson (Eds.), Campaigns and Elections American
Style (pp. 112-115). Boulder, Colorado: Westview.

Finkel, S. E., and Geer, J. G. (1998). A spot check: Casting doubt on the demobilizing
effect of attack advertising. American Journal of Political Science, 42: 573-
595.

Hofstetter, R. (1976). Bias in the News: Network Television Coverage of the 1972
Election Campaign. Columbus: Ohio State University Press.

Jacques, W., Meilinger, F., Balmoris, M., Gerns, C., and Denby, S. (1993). Some aspects
of major newspaper coverage of the 1992 presidential debates. American Behavioral Scientist, 37: 252-256.

Jamieson, K. H. (1992). Dirty Politics: Deception, Distraction, and Democracy. New
York: Oxford University Press.

Jamieson, K. H., and Birdsell, D. S. (1988). Presidential Debates. New York: Oxford
University Press.

Kendall, K. E. (2000). Communication in the Presidential Primaries: Candidates in the
Media, 1912-2000. Westport, CT: Praeger.

Kendall, K. E. (1997). Presidential debates through media eyes. American Behavioral
Scientist, 40: 1193-1207.

Kendall, K. E. (1993). Public speaking in the presidential primaries through media eyes.
American Behavioral Scientist, 37: 240-251.

Lichter, S. R., and Noyes, R. E. (1996). Good Intentions Make Bad Bews: Why
Americans Hate Campaign Journalism. Lanham, Maryland: Rowman & Littlefield.

Lopez-Escobar, E., Llamas, J. P., McCombs, M., and Lennon, F. (1998). Two levels of
agenda setting among advertising and news in the 1995 Spanish elections. Political Communication 15, 225-238.

Patterson, T. E., and McClure, R. D. (1976). The Unseeing Eye. New York: Putnam.

Patterson, T. E. (1994). Out of Order. New York: Vintage Books.

Perloff, R. M. (1998). Political Communication: Politics, Press, and Public in America.
Mahwah, NJ: Lawrence Erlbaum.

Popkin, S. (1991). The Reasoning Voter: Communication and Persuasion in Presidential
Campaigns. Chicago: University of Chicago Press.

Riffe, D., Aust, C. F., and Lacy, S. R. (1993). The effectiveness of random, consecutive
day and constructed week sampling in newspaper content analysis. Journalism Quarterly, 70: 133-139.

Riffe, D., Lacy, S., and Fico, F. G. (1998). Analyzing Media Messages: Using
Quantitative Content Analysis in Research. Mahwah, NJ: Lawrence Erlbaum.

Roberts, M., and McCombs, M. (1994). Agenda setting and political advertising: Origins
of the News Agenda. Political Communication, 11: 249-262.

Robinson, M. J., and Sheehan, M. (1983). Over the Wire and on TV: CBS and UPI in
Campaign '80. NY: Russell Sage Foundation.

Semetko, H. A., Blumler, J. G., Gurevitch, M., and Weaver, D. H. (1991). The Formation
of Campaign Agendas: A Comparative Analysis of Party and Media Roles in
Recent American and British Elections. Hillside, NJ: Lawrence Erlbaum.

Swanson, D. L. (1997). The political media complex at 50: Putting the
1996 presidential campaign in context. American Behavioral Scientist, 40: 1264-1282.

Thorson, E., Christ, W.G., and Caywood, C. (1991). Effects of issue-image strategies,
attack and support appeals, music, and visual content in political commercials. Journal of Broadcasting & Electronic Media, 35: 465-486.

Weaver, D. H., and Wilhoit, G. C. (1991). The American Journalist: A Portrait of U.S.
News People and Their Work, 2nd ed. Bloomington, IN: Indiana University
Press.

West, D. M. (1993). Air Wars: Television Advertising in Election Campaigns, 1952 -
1992. Washington, D.C.: Congressional Quarterly.












Appendix I. Coding Sheet


1.      Text # ____________

2.      Date/Year __________

3.      Publication  __________________

4.      The size of the story  _________ words

5.      The geographical nature of the audience of the ad
1.      National audience
2.      Key battle states
3.      Regional audience (specify, e.g., Michigan)  ______________
4.      Not identified

6.      Partisanship of the candidate sponsoring the ad
1.      Democrat
2.      Republican
3.      Others (specify) _____________
4.      Cannot identify

7.      The type of the office the sponsoring candidate is seeking
1.      Presidential (president, vice-president)
2.      The U.S. Senate
3.      The U.S. House of Representatives
4.      Governor
5.      State Senate
6.      State House of Representatives
7.      others  (specify)  ____________________

8       The status of the candidacy of the sponsoring candidate
1.      Incumbent
2.      Challenger
3.      Open-seat candidate
4.      Cannot identify

9       The valence of the ad
1.      Negative
2.      Positive
3.      Cannot identify

10      The type of the ad
1.      Issue-focused
2.      Image-focused
3.      Dovetailing (combination of issue and image)
4.      cannot identify

11      The type of issues (if the ad concerns specific policy issues)
1.      Abortion & life issues
2.      Budget & tax
3.      Campaign finance reform
4.      Economy (business, employment, agriculture_)
5.      education
6.      health care
7.      social security and retirement security
8.      social welfare (welfare reform, child welfare, working families, disabilities, immigration, etc.)
9.      defense
10.     environment
11.     energy
12.     law and enforcement
13.     Internet and technology
14.     others (specify) ________________

12      the attributes of candidate images (if the ad concerns candidates' personal qualities or traits)
1.      leadership
2.      credibility (trustworthiness, honesty, etc.)
3.      caring
4.      morality  (integrity, family-man, etc.)
5.      knowledgeable (intelligent)
6.      others (specify)  ___________________

13      The criteria for ad assessment used in the news
1.      Accuracy (fairness, legitimacy, etc.)
2.      effectiveness
3.      others (specify) _____________________
4.      mixed  (specify, e.g. 1+2 or 2+3 or 1+2+3) _____________
5.      no assessment
6.      cannot identify

14      The identities of sources commenting about the ad
1.      journalists
2.      academics
3.      sponsoring campaigns
4.      opposing campaigns
5.      other experts (specify) ______________
6.      cannot identify

15      The tone of the news commentary on the accuracy of the ad (i.e., if the ad is being assessed in terms of accuracy)

1.      Deflating  (presented as inaccurate)    lexicons ___________________
2.      Reinforcing (presented as accurate)    lexicons ___________________
3.      Neutral                                  lexicons ___________________
4.      Cannot identify

16      The tone of the news commentary on the effectiveness of the ad (i.e., if the ad is being assessed in terms of effectiveness)

1.      Deflating  (presented as ineffective)   lexicons ___________________
2.      Reinforcing (presented as effective)   lexicons ___________________
3.      Neutral                                      lexicons ___________________
4.      Cannot identify