Search for a command to run...
Relative contributions of television news and campaign to U.S. voters' knowledge about candidate issue differences are compared. Empirical comparisons are based on interview data from six campaign surveys of voters, in various election settings from 1984 to 1992. In hierarchical regression analyses, after controls for demographic and political interest variables, measures of to television news consistently account for a significant increment of slightly more than 2 percent of variance in issue knowledge. Parallel measures representing to candidates' televised advertisements produce a much more variable pattern in terms of variance explained in knowledge. Usually effects of advertisements are less than those of news, and sometimes they are nonsignificant; but in one hotly contested ideological race informative effect attributable to advertisements exceeds that of TV news. These patterns hold up after further controls for other media use variables, including newspaper reading. A commonly repeated generalization in political communication literature is Patterson and McClure's (1976) conclusion that voters learn issue information from television advertisements but not from television news. The two assertions are often paired in syntheses of literature (e.g., Diamond 1978; Diamond and Bates 1984; Graber 1989; Jamieson 1993; Kaid 1981; Kraus and Davis 1981; Nimmo 1978; O'Keefe and Atwood 1981). The study has been cited at least 150 times in academic journals (Social Sciences Citation Index 1976-93), including recent publications by political scientists (Bartels 1993; FinXINSHU ZHAO is assistant professor of journalism and mass communication at University of North Carolina at Chapel Hill, and STEVEN H. CHAFFEE iS Janet M. Peck Professor of international communication at Stanford University. Public Opinion Quarterly Volume 59:41-65 ? 1995 by American Association for Public Opinion Research All rights reserved. 0033-362X/95/5901-0006$02.50 This content downloaded from 157.55.39.104 on Sat, 18 Jun 2016 07:11:55 UTC All use subject to http://about.jstor.org/terms 42 Xinshu Zhao and Steven H. Chaffee kel 1993); some mass communication researchers consider it a classic of political campaign literature (Weaver and Drew 1993). Often overlooked when citing Patterson and McClure's conclusions is limited scope of their study, conducted during a single campaign (Nixon-McGovern, 1972 election) in a single county (Onondaga County, New York). Subsequent voter surveys have sometimes found little correlation between knowledge and to political commercials (see, e.g., Drew and Weaver 1991), and others attribute clear learning effects to television news (see, e.g., Bartels 1993; Drew and Reeves 1980; Lasorsa 1986; McLeod and McDonald 1985; Neuman, Just, and Crigler 1992; Sears and Chaffee 1979). Patterson and McClure's result remains prominent in literature more because it was first to make an explicit contrast between learning from TV news and ads than because it is consistent with most subsequent studies. The rule that ads are important to issue learning while news is not affects both research and practice. Just, Crigler, and Wallach (1990), as one instance, decided in their campaign research not to study television news at all. They concentrated instead on commercials (and on televised debates), citing Patterson and McClure as justification for their design. Diamond and Bates (1984) reported that political campaign managers' beliefs regarding television news and ads are also affected by Patterson-McClure conclusion, guiding daily decisions in field campaigns. The proposition that ads are more informative than news is not grounded in any general theory (Kraus and Davis 1981, p. 278). Indeed, it runs counter to many people's intuition-a feature that has probably helped attract to it. The more common view of political commercials is probably that of prominent practitioner who called them the most deceptive, misleading, unfair and untruthful of all advertising (Ogilvy 1985, pp. 210-13). Broadcast journalists, while criticized on many sides, are generally conscientious reporters who strive to be informative (Halberstam 1979). While local news programs may emphasize trivial events, political campaigns do get considerable TV coverage-partly, perhaps, in response to criticisms of television news following PattersonMcClure report. To infer that voters do not benefit from following news on television implies that this extensive professional effort goes for naught. Uncontrolled correlational studies suggest that TV news is less informative about politics than are newspapers, but this result does not hold up with controls for measurement error (Bartels 1993), prior knowledge (Chaffee and Schleuder 1986; Chaffee, Ward, and Tipton 1970), or questioning about attention to TV news rather than mere frequency of exposure (Chaffee and Schleuder 1986; McLeod and McDonald 1985). Still, empirical comparisons of ads versus news This content downloaded from 157.55.39.104 on Sat, 18 Jun 2016 07:11:55 UTC All use subject to http://about.jstor.org/terms Campaign Ads vs. TV News as Information Sources 43 effects have been outnumbered in literature by repetition of Patterson and McClure's conclusions, such as their assertion (1976, p. 54) that network TV news is simply not informative. Few surveys have asked parallel questions about TV news and ads, so that any doubt a given study might cast on one part of contrast does not extend to other.' Almost all relevant comparisons have been one-campaign (and usually one-locale) studies, potentially as limited in generalizability as was original. In this article we report a series of surveys intended to test generalization with data from a variety of election settings, using measures designed to provide a clearer comparison of news versus ads as agencies of voter learning about political issues. Our general method is correlational, and we include here an extensive set of control variables. Individual differences in knowledge might be associated with differential to either news or ads for reasons quite apart from effects of latter on former. Knowledgeable people do, it is well established, seek further information in areas where they are already expert (Sears and Freedman 1967). So do well-educated people, and years of schooling is one of several controls that we routinely enter before our tests of specific effects of TV news and ads. We also enter stringent behavioral controls where available, including other kinds of political knowledge (apart from issue positions of candidates) and use of newspapers.