The subtle nature of presidential debate influence.

• Author: Pfau, Michael

Televised debates have become an intrinsic communication form in contemporary presidential campaigns. Every presidential election campaign since 1976 has featured one or more debates. In addition, they are a watershed event in most campaigns: a moment in which other campaign activity slows and media, pundits, and the attentive public focus their attention on debates (Jamieson & Birdsell, 1988).

Most of us who study political communication believe that debates are superior to other communication forms in that they offer an opportunity for candidates to advocate the relative superiority of their positions via a communication venue that facilitates clash, depth, and unfiltered access. Indeed, some studies confirm that debates do, in fact, provide more depth in candidate discourse than other communication forms (Ellsworth, 1965; Mortensen, 1968). Contention among political communication scholars mainly concerns how this communication form might best serve American democracy (i.e., what format is optimal).

Political communication scholars operate from the premise that debates matter--that they make a difference: in people's knowledge and perceptions and, therefore, in campaign outcomes; and by strengthening the democratic process itself However, the search for debate effects has been elusive.

Many studies have addressed the question of debate effects since general election televised debates commenced in 1960. The results of these studies tend to conflict, although, on balance, they suggest modest effects.

There are learning effects, for certain. The preponderance of empirical studies indicate that presidential debates conducted during the general election enhance viewer learning about the candidates and their issue positions (Abramowitz, 1978; Apker & Voss, 1994; Becker, Sobowale, Cobbey, & Eyal, 1980; Benoit, Webber, & Berman, 1998; Chaffee, 1978; Chaffee & Dennis, 1979; Dennis, Chaffee, & Choe, 1979; Drew & Weaver, 1991; Jamieson & Adasiewicz, 2000; Katz & Feldman, 1962; Kelly, 1962; Lanoue, 1992; Lemert, 1993; Lichtenstein, 1982; Lupfer & Wald, 1979; Mayer & Carlin, 1994; Miller & MacKuen, 1979; O'Keefe & Mendelsohn, 1979; Pfau & Eveland, 1994; Sears & Chaffee, 1979; Twentieth Century Task Force, 1979; Zhu, Milavsky, & Biswas, 1994). However, potential for voter learning varies across presidential contests, thereby explaining the results of the handful of dissenting studies that found no significant learning effects (e.g., Bishop, Oldendick, & Tuchfarber, 1980; Hagner & Rieselbach, 1980). Learning is more l ikely in presidential elections without a popular incumbent and in debates which occur earlier in an election campaign (Hellweg, Pfau, & Brydon, 1992). However, the bulk of studies, including most of those reporting statistically significant learning effects, indicate that the magnitude of such effects are modest and, inevitably, they fall far short of pre-debate expectations (Graber & Kim, 1978; O'Keefe & Mendelsohn, 1979; Hellweg et al., 1992).

If learning effects are clear, albeit modest, persuasive effects have been far more equivocal. This fact has frustrated political communication scholars ever since the first televised general election presidential debates in 1960.

Many studies suggest that presidential debates impact voter attitudes toward candidates and, under certain conditions, voting disposition (Barnett, 1981; Becker, Pepper Wenner, & Kim, 1979; Benoit & Wells 1996; Ben-Zeev & White, 1962; Brydon, 1985; Casey & Fitzgerald, 1977; Chaffee Choe, 1980; D. Davis, 1979; M. Davis, 1982; Geer, 1988; Kelley, 1983; Krivonos, 1976; Ladd & Ferree, 1981; Leuthold & Valentine 1981; Middleton, 1962; Pfau & Kang, 1991; Pfau & Eveland, 1994; Pfau, Cho, & Chong 2001; Robinson, 1979; Roper, 1960, 1977; Swerdlow, 1984; Tannenbaum, Greenberg, & Silverman, 1962; Walker & Peterson, 1981). However, the persuasive influence of presidential debates is not as clear as the learning effects of debates. Many studies failed to detect significant persuasive effects. Instead, these studies suggest that debates reinforce attitudes, but they seldom change them (Abramowitz, 1978; Apker & Voss, 1994; Benoit, McKinney, & Holbert, 2001; Bishop et al., 1980; Davis, 1979; Eadie Krivonos, & Goodman, 1977; Feigert & Bowling, 1980; Hagner & Rieselbach, 1980; Katz & Feldman, 1962; Kennamer, 1987; Lemert, Elliott, Bernstein, Rosenberg, & Nestvold, 1991; Lubell, 1962; McLeod, Durall, Ziemke, & Bybee, 1979; Miller & Mackeun, 1979; Mulder, 1978; Payne, Golden, Marlier, & Ratzman, 1989; Rose, 1979; Rouner & Perloff, 1988; Sebald, 1962; Sigelman & Sigelman, 1984; Simons & Liebowitz, 1979; Smith, 1977; Vancil & Pendel, 1984; Wald & Lupfer, 1978).

Again, context is critical. Persuasive effects are limited, mainly, to elections that lack a popular incumbent, where there are more prospective voters who are uninformed or conflicted (Geer, 1988; Hellweg, et al. 1992). As Becker and Kraus observed: "[Studies] demonstrate...that campaign communication...can affect voting decisions rather directly--at least when 'one of the candidates is not well known, many voters are undecided, the contest appears to be a close one, and party allegiances are weak"' (1978, p. 267). Persuasive effects also are more likely in presidential debates which occur earlier in an election campaign, before prospective voters' attitudes solidify (Chaffee & Choe, 1980).

Overall, the search for persuasive effects has proven to be disappointing, especially given the prominent role that debates have assumed in contemporary presidential campaigns. Some studies indicate no significant influence. Others find a difference, but effect sizes (the variance accounted for in dependent variables, such as attitudes or voting disposition) have proven consistently disappointing. The evidence to date suggests small, if sometimes important, persuasive effects.

Nonetheless, I maintain that televised presidential debates typically produce significant effects (except when the number of undecided or conflicted prospective voters is unusually small, as is often the case in elections that feature a popular incumbent), but that these effects are subtle and, therefore, they have been largely undetected by most studies to date. These subtle effects stem, in part, from candidates' relational messages, which are an intrinsic feature in all human communication (Watzlawick, Beavin & Jackson, 1967), but are especially pronounced in circumstances in which the communication form facilitates real or perceived contact with a communicator (Burgoon & Hale, 1987), as it does in most televised communication (Pfau, 1990). Relational effects defy detection via traditional methods, in part because they are both verbal and nonverbal but most assessment tools are biased toward verbal effects, and in part because they take shape over time but most assessment is short term. These subtle effect s also emanate from what Owen (1991) terms "the interwoven nature" of contemporary campaign communication. In today's "interwoven" communication environment, the effects of any one communication form are inevitably commingled with the effects of other forms. Commingled effects defy detection because scholars tend to view the effects of particular communication forms in isolation from other forms.

This essay argues that both sources of effects, those that stem from candidates' relational communication and those that are commingled with the influence of other communication forms, are of statistical and practical importance. The essay also calls for research to assess both of these subtle sources of presidential debate influence.

RELATIONAL INFLUENCE

In part, because many presidential debate scholars come out of rhetorical and/or argument backgrounds and, in part, because they believe that democracy would be better served if more debate viewers focused on candidates' verbal discourse, political debate research has tended to overemphasize the content of debates and to ignore communication form (Hellweg et al., 1992; Kraus, 1988). Most research operates from the maxim that "to the extent that presidential debates influence receivers at all, they do so via their content" (Heliweg et al., 1992, p. 72).

Yet, Jamieson (1988), Graber (1987), and others have argued previously that contemporary television is a unique communication medium. It emphasizes visual over verbal content. Graber (1987) explains that, "When candidates for political office are shown on the television screen, audiences ... use the pictures to judge the candidates' personality ..." (p. 77). What candidates say matters, not so much due to an intrinsic substantive value, but because of what it communicates about the candidate as a person. In short, what candidates say and how they say it cultivates impressions of them. "Debates push personal characteristics to the foreground" (Meadow 1987, p. 209). Lang affirms the ascendancy of persona over substance in televised debates. She explains that: "What matters most in them [debates] is not the substance of what the candidates say ... but how well they say it and whether the candidate projects the image he strives to project" (1987, p. 213).

What Lang is talking about is termed presentational style. Viewers make holistic judgments about candidates based on their presentational style. The judgments are about the candidates as people: their ethos, image, or persona. Hinck (1993) explains that, an audience comprehends and reacts to a debate in its entirety, as a dramatic conflict of character, [and] not as a discursive argument over policies" (pp. 214-215). Thus, it comes as no surprise that many of the studies that reported persuasive effects for presidential debates attributed them more to holistic judgments about persona than to substantive content (Becker et al., 1979; Benoit et al., 1998; Berquist & Golden, 1981; Bowes & Strentz, 1979; Brydon, 1985; Casey & Fitzgerald, 1977; D. Davis, 1979; M. Davis, 1982; Lang & Lang...