Asia Pacific Advances in Consumer Research Volume 1, 1994 Pages 261-264
USING DRAMA TO PERSUADE: THE EFFECTS OF INVOLVEMENT AND AD FORM ON PERSUASION
Siew Meng Leong, National University of Singapore
Swee Hoon Ang, National University of Singapore
Lynn Heng, Singapore Post
This study examines how consumers of varying levels of involvement process information from drama and argument ads. Such indicators of persuasion as expression of feeling, expression of belief, and verisimilitude were examined. Results indicated that drama ads were more persuasive than argument ads regardless of the level of consumer involvement. Implications of the findings are discussed.
Given the vast amount of financial resources devoted to advertising, practitioners are rightly concerned with its effectiveness. A persuasive ad is one that is able to communicate the right message to the right audience at the right time to effect the right actions. Such persuasiveness is dependent on many factors, involving either the ad itself (e.g., style of presentation), or the consumer (e.g., level of involvement), or both.
Persuasion studies in several disciplines have distinguished between reasoned argument and drama or story (Bruner (1986) in psychology; Booth (1961), Chatman (1978), and Scholes (1981) in literature; Goldberg (1982) in theology; Bennett and Feldman (1981) in law; Fisher (1984), Goldman (1992), Leiss, Kline, and Jhally (1990), and Williamson (1978) in communication; White (1981) in history; and McCloskey (1985) in economics). The distinction is important because both work in fundamentally different ways, and both have different strengths and weaknesses as they use different mechanisms to persuade. As television ads can be classified as either dramas or arguments or hybrids of these forms, this distinction between dramas and arguments matters to the study of advertising.
However, empirical research in this manner of classifying ads is lacking. To date, besides Wells (1988) who was the first to apply it to advertising, only Deighton, Romer, and McQueen (1989) have argued for the extent of dramatization as a different form of advertising. Thus, this study extends consumer research in this area to the realm of information processing. It attempts to study the effectiveness of these two forms of advertising on individual's information processing under varying levels of involvement.
What is known about lectures (or arguments) and dramas outside of television can make useful contributions to our understanding of how drama and argument ads work, and what makes these particular styles in advertising successful. The distinction between lecture and drama (or argument and drama ads) also has implications for how we conduct advertising research, for example, in copy testing for unfinished ads. An additional contribution to researchers is that this study offers an opportunity to explore a new territory. Much is known about "lectures", but our understanding of "drama" is lacking.
To practitioners, the distinction between lecture and drama offers the challenge to shake off the handicaps imposed by lecture-based research. Advertising conveys information, and lectures do that very well. However, advertising can also invoke inferences and preview feelings that the product can produce (Wells 1988). Advertisers who do not take advantage of these effects are ignoring part of what their investment in advertising can add to their brands. This also offers an opportunity to outflank competitors who bind themselves to information narrowly defined. "Logic does not always work . . . when logic fails, imaginative advertisers shift to overdrive" (Wells 1988).
Therefore, the objectives of this paper are: (a) to examine the different manner in which drama and argument ads affect consumer perceptions, and (b) to investigate how the level of involvement affects the nature of information processing and hence, the effectiveness of the two types of ads in persuading consumers.
The next section provides a literature review on drama and argument ads as well as on involvement, a possible moderating factor which will be investigated in this study. Based on this, two sets of hypotheses are developed for empirical testing. The research method used is then explicated, followed by the results produced therefrom. Theoretical and managerial implications of the findings are then discussed and directions for future research suggested.
Several conceptualizations of advertising form are available in the literature (cf. Aaker and Myers 1987). Wells (1988) asserts that one dimension is the extent to which an ad can be dramatized. Two ad forms-drama and argument-can be conceived as extreme points on a scale constructed with plot, character, and narration attributes marking the transitions along it (Deighton, Romer, and McQueen 1989).
Specifically, the argument ad form has a narrator but no plot or character. The narrator serves to underscore an event's meaning, explaining the relevance of its claim. As its indicative mood explicitly points to what the audience would believe and why, the audience tends to evaluate the claims objectively, conscious of the persuasive intent of the commercial (Deighton et al. 1989).
As a plot (comprising a stable state of affairs that is breached to induce a crisis and finally redressed; Bruner 1986), and human characters or protagonists are introduced, and the narrator removed, there is a transition towards the drama form. Characters serve to make human values salient (Scholes 1981), so that when placed in conjunction with the product, the ad can express claims of product value. Without the narration, advertising has to depend on verisimilitude to establish what the depicted events are worth to the claim. Verisimilitude is when the actions or events in the commercial seem plausible and authentic. They can therefore draw viewers into the commercial as they identify with the characters and events in it. Thus, in a drama ad, explicit claims are not used in exchange for the power of empathy (Deighton et al. 1989).
Further, Deighton et al. (1989) also contend that the essential difference between argument and drama ads is that between telling and showing. In telling, a narrator does some of the audience's thinking, explaining events and warranting their meaning. In showing, the verisimilitude of events alone, through the ad's ability to build empathy, determines how well the claim is backed up or supported.
Following the pioneering work of Wells (1988), Deighton et al. (1989) argued that drama ads appeal to personal experience where ad claims are processed subjectively. By evoking expression of feelings and verisimilitude in consumers, expressions of beliefs about the brand are created. In contrast, argument ads appeal to objectivity, inviting counter- and support-arguments.
Involvement has been found to be a moderating factor in ad processing (Burnkrant and Sawyer 1983; Petty, Cacioppo, and Schumann 1983). Specifically, drama or argument ads may have a greater impact on consumer information processing under varying levels of product involvement. Such involvement should also moderate the persuasiveness of ad forms.
As a drama ad appeals to subjective truth, we expect expression of feelings and conviction of its verisimilitude to occur with an audience with low involvement. Hence, persuasion of drama ads under low involvement is evidenced by a positive relationship between expression of feelings and conviction of verisimilitude with expression of belief in the ads. For argument ads, there is less appeal to subjective truth, resulting in minimal expression of feelings and verisimilitude. Thus, their persuasiveness under conditions of low involvement will be lower as evidenced by a smaller positive relationship between expression of feelings and verisimilitude with expression of belief.
In contrast, a highly involved audience would require cogent arguments to be persuaded of the ad's value. Thus, drama ads would not be persuasive as argument ads. Hence, no relationship between expression of belief is expected with expression of feelings or verisimilitude under high involvement-drama condition. Exposure to argument ads under high involvement would lead to the generation of more cognitive responses and greater expression of belief but few expressions of feelings and verisimilitude. Hence, under such circumstances, there will also be no relationship between expression of feelings and verisimilitude with expression of belief.
Based on these arguments, the following hypotheses are advanced for empirical testing in this research:
H1a: Under low involvement, verisimilitude is more positively correlated with expression of belief for drama ads than argument ads.
H1b: Under high involvement, verisimilitude is not correlated with expression of belief for both drama and argument ads.
H2a: Under low involvement, expression of feeling is more positively correlated with expression of belief for drama ads than argument ads.
H2b: Under high involvement, expression of feeling is not correlated with expression of belief for both drama and argument ads.
A 2x2 between-subjects design was used. The factors were involvement (low versus high) and ad form (drama versus argument). Subjects were 109 undergraduate students, randomly assigned to treatment cells of sizes ranging from 22 to 31.
Ad Form. Television commercials were preferred to print or radio ads because they could more realistically and vividly portray the distinction between drama and argument ads. The commercials used in this study were authentic and good advertising, by the criterion that firms in England and Australia were willing to air them on prime-time television. Based on pretesting, five pairs of commercials were chosen. Each drama ad had a correspondingly equal time duration to its argument counterpart. Pretest results showed that the two sets of ads did not vary significantly from each other on comprehensibility and entertainment value. These were measured by having subjects rate the extent to which they agreed to the ads being understandable and entertaining on seven-point scales respectively.
Manipulation checks involved two seven-point scales embedded among the items measuring the dependent variables. The first item measured the extent to which the subjects perceived the ad to be a drama ad ("1" for dramatic and "7" for not dramatic). The second, which was more indirect, measured the extent to which the subjects could see a "story" developing while watching the ad ("1" for telling a story and "7" for not telling a story).
Involvement. Involvement was manipulated following the procedure developed by Apsler and Sears (1968). In the low involvement condition, subjects were told that they will be watching ads of brands that are currently unavailable in a neighboring country. However, they will be made available there in three years' time. In the high involvement condition, they were told that the ads concerned brands currently not available in their own country, but which will be introduced in a year's time. Five seven-point semantic differential scales partially adopted from Zaichowsky (1985) were used as manipulation checks: personal relevance/personal irrelevance; of personal concern/not of personal concern; involving/uninvolving; significant/insignificant; matters to me/does not matter to me.
The three dependent variables measured were expression of belief, expression of feelings, and verisimilitude (EOB, EOF, and VERI respectively). There were two items measuring each dependent variable (see Table). Subjects indicated their extent of agreement to each item on seven-point scales.
Subjects were told that they were participating in a study concerning different ad forms. They were instructed on the characteristics of a drama and an argument ad. Subjects were then shown two local adsCone demonstrating a drama ad and another argument adCto educate them in visualizing differences between these ads forms. Next, the manipulation for involvement took place. Subjects were then shown the five test ads. After each ad was shown, an eight-item questionnaire was administered, containing the six dependent measures and two manipulation checks on ad form. When responses to the five test ads were completed, measures on the involvement manipulation as well as some demographic characteristics were obtained.
Ad Form. As the two items measuring ad form were reliable at a=.74, an overall measure for ad form was obtained by summing the items. The drama ads were perceived to tell a story and be more dramatic than the argument ads (x=5.68 versus 10.17; t=-12.08, p<.01). Hence, the manipulation of ad form seemed successful.
Involvement. High reliability was observed among the five measures of involvement (a=.90). Summing these scores, involvement was higher among subjects in the high versus low involvement condition (x=19.48 versus 13.52; t=4.67, p<.01). Therefore, the manipulation of involvement appeared to be successful as well.
ITEMS MEASURING EXPRESSION OF BELIEFS, EXPRESSION OF FEELINGS, AND VERISIMILTUDE
Given that only two items were employed to assess these constructs, satisfactory reliability was obtained for EOF and VERI (a=.95 and .64 respectively), but not for EOB (.43). Thus, the items for each dependent variable were summed for an overall score, with due qualification noted for the EOB measure.
H1. H1a stated that under low involvement, verisimilitude is more positively correlated with expression of belief for drama ads than argument ads, while H1b posited that under high involvement, verisimilitude is not correlated with expression of belief for both ad forms. Results showed that the correlation between VERI and EOB was .43 for both drama and argument ads under low involvement (p's<.01). It was .61 and .41 (p's<.01) for drama and argument ads respectively under high involvement. Hence, both H1a and H1b were not supported.
H2. H2a stated that under low involvement, expression of feeling is more positively correlated with expression of belief for drama ads than for argument ads, while H2b stated that under high involvement, expression of feeling is not correlated with expression of belief for both drama and argument ads. The results showed that there was a significantly positive relationship between EOF and EOB for the low involvement-drama ad condition (r=.48, p<.01), but not for the argument ad condition (r=.32, p<.10), thus supporting H2a. Under high involvement, an insignificant relationship between EOF and EOB was observed for argument ads (r=.15, p>.10), but not for drama ads (r=.43, p<.05). Thus, H2b was not supported.
The results indicated that contrary to expectations, the relative persuasiveness of drama ads through expression of feelings and verisimilitude over argument ads was evident across both levels of involvement. As predicted, the persuasiveness of drama ads over argument ads becomes evident under low involvement as they appeal to subjective truth rather than cogent argument. However, the expected reliance on objectivity in processing ad information favoring argument appeals over drama ads did not seem to materialize under high involvement.
One possible conceptual account that may be advanced for these findings is Petty and Cacioppo's (1986) Elaboration Likelihood Model (ELM). These scholars argue that the amount and nature of issue-relevant elaboration in which individuals are willing and able to engage in evaluating a message vary with personal and situational factors. A central route to attitude change is invoked when motivation and ability to process issue-relevant arguments increase, while peripheral cues become relatively more important determinants of persuasion as motivation and/or ability to process such arguments decrease.
Clearly, the findings in the low involvement condition may be interpreted from the ELM framework. Specifically, the drama ad form may serve as a peripheral cue which induces consumer persuasion under low involvement. However, problems arise with regard to interpreting the results concerning the high involvement subjects. In particular, it would seem consistent from this perspective that the argument ads would be favored as they appear to be more issue-relevant than dramatic appeals. Yet, such findings were not obtained.
One possibility to consider is that the central and peripheral routes may not be mutually exclusive. Although cognitive responses associated with central route were not measured in this research, some degree of peripheral processing may have occurred, thus resulting in drama ads being more persuasive over argument ads when subjective appeals are considered. The involvement manipulation may also have contributed to this. Pretest scores show that the high involvement subjects may actually be moderately involved as suggested by comparing their mean (19.48) to the scale midpoint of 20. Based on past research (Petty and Caccioppo 1986, p. 210), moderately involved subjects may use source or other peripheral factors to determine how much thinking they should engage in with respect to message-relevant arguments in an ad.
A second possibility surrounds the fact that specific cues cannot be assumed to be processed either centrally or peripherally. For example, Petty and Caccioppo (1981) found that their peripheral cue (an attractive source) also had an effect under high involvement. What occurred was that the source's hair served as a message-relevant argument for the shampoo brand used in the study's ad. Similarly, Yalch and Elmore-Yalch (1984) found that the number of arguments in a message served as a simple peripheral cue in their experiment. Applied in the present context, therefore, a specific ad form should not necessarily be conceived as triggering a particular path to persuasion. Thus, the narrators in the argument ads employed in the study may have served as negative peripheral cues to subjects.
Future research may extend our knowledge of the different processing mechanisms suggested by drama versus argument ads. Direct measures of cognitive response and brand beliefs could be collected to furnish further insights in this connection. Moreover, better measures may be developed for assessing the expression of belief dependent variable, given its lower internal consistency observed here relative to the measures for expression of feeling and verisimilitude.
Factors other than involvement may be incorporated into future experiments to investigate alternative moderating influences affecting the persuasiveness of these ad forms. These may include those which assess other dimensions of consumer motivation to process information such as their need for cognition (Caccioppo and Petty 1982). Use of other media (e.g., print) may also be considered. The use of TV commercials in this study may have biased subjects toward low involvement processing (Krugman 1965). Various manipulations of consumer ability to process information including ad repetition may also be considered. Finally, it may also be instructive to assess the persuasiveness of the various ad forms more explicitly from the perspective of the ELM. Such work would not only enhance the execution of the dramatic ad form by practitioners but also extend the boundaries of the ELM in consumer research.
Aaker, David A. and John G. Myers (1987), Advertising Management, 3rd edition, Englewood, NJ: Prentice-Hall.
Apsler, R. and D. O. Sears (1968), "Personality, Involvement, and Attitude Change," Journal of Personality and Social Psychology, 9, 162-166.
Bennett, W. Lance and Martha S. Feldman (1981), Reconstructing Reality in the Courtroom, New Brunswick, NJ: Rutgers University Press.
Booth, Wayne C. (1961), The Rhetoric of Fiction, Chicago, IL: University of Chicago Press.
Bruner, Jerome (1986), Actual Minds, Possible Worlds, MA: Harvard University Press.
Burnkrant, R.E. and Alan Sawyer (1983), "Effects of Involvement on Information Processing Intensity," in Information Processing Research in Advertising, ed. R.J. Harris, Hillsdale, NJ: Erlbaum.
Caccioppo, John T. and Richard E. Petty (1982), "The Need for Cognition," Journal of Personality and Social Psychology, 42, 116-131.
Chatman, Seymour (1978), Story and Discourse: Narrative Structure in Fiction and Film, Ithaca, NY: Cornell University Press.
Deighton, John, Daniel Romer, and Josh McQueen (1989), "Using Drama to Persuade," Journal for Consumer Research, 16 (December), 335-343.
Fisher, Walter R. (1984), "Narration as a Human Communication Paradigm: The Case of Public Moral Argument," Communications Monographs, 51 (1).
Goldberg, Michael (1982), Theology and Narrative, Nashville, TN: Abington.
Goldman, Robert (1992), Reading Ads Socially, New York: Routledge.
Krugman, Herbert E. (1965), "The Impact of Television Advertising: Learning Without Involvement," Public Opinion Quarterly, 29 (Fall), 349-356.
Leiss, William, Stephen Kline, and Sut Jhally (1990), Social Communication in Advertising: Persons, Products, and Images of Well-Being, New York: Routledge.
McCloskey, Donald N. (1985), The Rhetoric of Economics, Madison, WI: University of Wisconsin Press.
Petty, Richard E. and John T. Cacioppo (1981), "Issue-Involvement as a Moderator of the Effects on Attitude of Advertising Content and Context," in Advances in Consumer Research, Vol. 8, 20-24.
Petty, Richard E. and John T. Caccioppo (1986), Communication and Persuasion: Central and Peripheral Routes to Attitude Change, New York: Springer-Verlag.
Petty, Richard E., John T. Caccioppo, and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10 (September), 135-146.
Scholes, Robert (1981), "Language, Narrative, and Anti-Narrative," in On Narrative, W.J.T. Mitchell ed., Chicago, IL: University of Chicago Press.
Wells, William D. (1988), "Lectures and Drama," in Cognitive and Affective Responses to Advertising, Pat Cafferata and Alice Tybout eds., Lexington, MA: D.C. Heath.
White, Hayden (1981), "The Value of Narrativity in the Representation of Reality," in On Narrative, W.J.T. Mitchell ed., Chicago, IL: University of Chicago Press.
Williamson, Judith (1978), Decoding Advertisements, New York: Marion Boyers.
Yalch, Richard F. and Rebecca Elmore-Yalch (1984), "The Effect of Numbers on the Route to Persuasion," Journal of Consumer Research, 11 (June), 522-527.
Zaichowsky, Judith Lynne (1985), "Measuring the Involvement Construct," Journal of Consumer Research, 12 (December), 341-352.