ISSA Proceedings 2010 – Variations of Standpoint Explicitness In Advertising: An Experimental Study On Probability Markers

ISSA2010Logo1. Introduction
Empirical research has demonstrated that variation in standpoint explicitness matters. In several research reports, explicit articulations of a standpoint or conclusion have been compared to more implicit articulations. Meta-analyses of such reports (Cruz, 1998; O’Keefe, 1997, 2002) have shown that messages with explicitly stated standpoints are more persuasive than messages without such standpoints. Such effects were not found for advertising messages, for which the conclusion – buy this product – seems relatively straightforward, regardless of the articulation of the conclusion (Cruz, 1998).

There are different ways in which explicit conclusions may be articulated, one of which is the use of probability markers. Advertising research has compared hedges (which mark a standpoint as moderately probable) and pledges (which mark a standpoint as very probable). In this study, it was investigated whether the reputation of the brand affects the persuasiveness of hedges and pledges. Based on a study conducted by Goldberg and Hartwick (1990), it was expected that hedges would be more persuasive for low-reputation brands, whereas pledges would be more persuasive for high-reputation brands. This expectation was put to a test in an experiment.

2. Standpoint explicitness
The pragma-dialectical approach to argumentation views argumentation as reasonable discourse aimed at resolving a difference of opinion (Van Eemeren & Grootendorst, 1992, 2004). In order to discuss reasonably, a set of rules for critical discussion is proposed. One of the rules holds that parties should express themselves clearly, unambigiously and explicitly, because this allows for critical scrutiny. This means, for instance, that the proponent has to explicitly put his or her standpoint on the table. Although normatively reasonable, standpoint explicitness may seem to threaten the persuasive effectiveness of the proponent. As O’Keefe (1997, p. 2) summarizes, greater explicitness “invites closer scrutiny, counterargument, objection, rejection”. A number of studies have empirically investigated whether greater standpoint explicitness is associated with less persuasive effectiveness. These studies have been summarized in statistical meta-analyses (Cruz, 1998; O’Keefe, 1997, 2002). O’Keefe divided these studies into two categories: studies on conclusion omission (messages with or without a conclusion) and studies on conclusion specificity (the conclusion is explicit, but may be general or specific). In O’Keefe (2002), which contains more studies than O’Keefe (1997), the meta-analysis involving 35 comparisons demonstrated that more explicit articulation of standpoints was found to be more persuasive than less explicit articulation. This result was found for both the conclusion omission studies and for the conclusion specificity studies. In another meta-analysis with a different set of studies, Cruz (1998) reached the same conclusion. In sum, this means that the normative consideration of standpoint explicitness is in line with empirical results (cf. O’Keefe, 2007).

A meta-analysis summarizes findings from primary research. Some primary research reports may have findings that deviate from the general conclusion. This is the case for two advertising studies mentioned in Cruz (1988): Kardes (1988) and Sawyer and Howard (1991). As O’Keefe (1997, 2002) notes, Kardes (1988) is not a study on standpoint explicitness, but on specificity of supporting arguments. Contrary to the general findings on standpoint explicitness, the advertising study reported in Sawyer and Howard (1991) showed that the implicit standpoint was more persuasive than the explicit standpoint. Cruz (1998) gives two explanations for this result. In the first place, the advertising text was shorter than the texts in the average other study. In longer texts, explicit standpoints are needed to comprehend the proponent’s standpoint, whereas this is less likely the case for shorter texts. In the second place, the genre of advertising may play a role: “one conclusion is readily understood in all advertisements: Buy the product” (Cruz, 1998, p. 222). As a result, for advertising texts it seems that standpoint explicitness does not matter.

3. Probability markers
The purpose of an advertisement is to positively affect people’s attitude towards the product, attitude towards the brand, purchase intention and – ultimately – actual purchase. The message that an ad conveys is generally related to the benefits of the product or service: product X has benefit Y, leads to Y, gives you Y (cf. Darley & Smith, 1993). This is a descriptive standpoint or claim that can be true to some degree. An example is given in (1).

(1) Our nasal spray helps you breathe freely.

This uniformity in advertising message structure does not mean, however, that advertisers do not vary in the way they put forward claims. In one particular field of study, the interest has been on the effectiveness of probability markers that can be used in claims. A probability marker signals the degree to which a claim is true (Berney-Reddish & Areni, 2005, 2006). A pledge, such as ‘absolutely’ or ‘undoubtedly’, signals complete certainty of the claim, such an in example (2). A hedge, such as ‘likely’ and ‘possibly’, signals that the claim is not necessarily true, such as in example (3).

(2) Our nasal spray always helps you breathe freely.

(3) Our nasal spray in most cases helps you breathe freely.

A few studies have compared the relative persuasiveness of hedges and pledges in advertising claims. Berney-Reddish and Areni (2005, 2006) compared the two probability markers in four texts for different products, and showed that hedges and pledges were equally persuasive. A similar finding was reported in Hornikx, Pieper and Schellens (2008), who had participants rate eight different cosmetics claims with these markers.

Two characteristics of these experiments offer suggestions for future research. A first characteristic is that the experiments used multiple message designs with simple, abstract claims. The use of a multiple message design improves internal validity, but disadvantages ecological validity. It is an empirical question as to how persuasive hedges and pledges are in a more realistic advertising setting. This leads to the following research question:
Research question: Is there a persuasive difference between hedges and pledges in advertising claims in a realistic advertisement?

A second characteristic of the experiments – a consequence of the first characteristic – is that the proponent of the standpoint (the brand) was not identified. There are reasons to believe that the brand affects how people are persuaded by advertising claims. Goldberg and Hartwick (1990) reasoned that the effectiveness of the claim that brands put forward partially depends on their reputation. Brands with a high reputation are in a better position to express a strong claim than brands with a lower reputation. With an experiment for the brand Miro, Goldberg and Hartwick (1990) indeed demonstrated that extreme claims (e.g. ‘Miro came first against the world’s top 100 products in its category’) were more persuasive than less extreme claims (e.g., ‘Miro came twentieth against the world’s top 100 products in its category’) when the brand was introduced as a high-reputation brand, and that less extreme claims were more persuasive than extreme claims when the brand was introduced as a low-reputation brand. This relationship may also apply to hedges and pledges, as a claim with a pledge may be considered as a more extreme claim, and a claim with a hedge as a less extreme claim. Based on the study of Goldberg and Hartwick (1990), the following hypothesis was formulated:

Hypothesis: A hedge is more persuasive than a pledge in an ad for a low-reputation brand and a pledge is more persuasive than a hedge in an ad for a high-reputation brand

4. Method
An experiment was set up to answer the research questions and to test the hypothesis. Dutch participants were given a description of a company profile, an advertisements for nasal spray of that compay, and a questionnaire that contained the relevant dependent measures.

4.1 Material
Participants were told that an American company, Sinus Relief, was considering the introduction of their nasal spray on the Dutch market. Before participants were invited to read a potential advertisement, they were given background information about that company. Participants received a fictitious, but realistic company profile from the Wall Street Journal in which the company Sinus Relief was described. This procedure to manipulate brand reputation was borrowed from Goldberg and Hartwick (1990).

In one version of the article, Sinus Relief was presented as a high-reputation brand, and in another version, the company was presented as a low-reputation brand. The two versions each contained 190 words distributed over three paragraphs, but differed with respect to the company’s characteristics, such as number of years in business (more than 60 years vs. 10 years), sales volume (86 million vs. 3 million), market share (48% vs. 4%), and number of employees (2100 vs. 78).

This manipulation was checked in a pretest among 50 Dutch participants, of whom 60% was female, and of whom 68% had followed higher education. The participants were on average 30.84 (SD = 12.05) years old (range 20-62). Participants responded on 5-point semantic differentials (very bad – very good) to three statements: “The reputation of Sinus Relief among employees is”, “The reputation of Sinus Relief among customers is”, and “The reputation of Sinus Relief among investors is” (α = .87). In the high-reputation text, which was read by half of the participants, the reputation of Sinus Relief was perceived as higher (M = 4.25, SD = 0.38) than in the low-reputation text (M = 2.39, SD = 0.73); F (1, 48) = 128.95, p < .001, h2 = .73.

Next to the company profile, the material consisted of two versions of an advertisement for a nasal spray from Sinus Relief. One version contained hedges, the other pledges. A number of markers were pretested among other participants (16 Dutch students): ‘always’ (9.19) and ‘absolutely’ (8.69) scored highest on a 10-point probability scale and were used as pledges, whereas ‘in most cases’ (6.00) and ‘usually’ (5.25) scored much lower and were used as hedges. Note that scores below the midpoint of the scale mean that a marker indicate improbability rather than probability, which would have made such a marker inappropriate to function as a hedge. In order to emphasize the use of markers, not one but two markers were used in text (4); ‘always’ and ‘absolutely’ as pledges, and ‘in most cases’ and ‘usually’ as hedges:

(4) “Got a cold? We know how annoying that is. Our nasal spray brings relief. It will [always / in most cases] help you breathe freely. Sinus Relief: [absolutely / usually] the best choice for your nose”.

The two ads each contained a picture of a woman, a brand logo, the product, and a text.

4.2 Participants
A total of 137 Dutch people participated in the study, of whom 51.8% was male, and of whom 69.3% had followed higher education. None of these people had participated in either of the two pretests. The participants were 33.51 years old on average (SD = 13.47), with ages from 18 to 67. The four groups of participants (see ‘Design’) did not differ in mean age (F (3, 132) < 1), or levels of education (c2 (15) = 16.50, p = .35), but differed in gender distribution (c2 (3) = 9.55, p < .05). This difference in gender distribution does not seem to have affected the results, because there was no main effect of gender on the dependent measures (F (4, 132) < 1).

4.3 Design
The experiment had a 2 (high vs. low reputation) x 2 (pledges vs. hedges) between-subjects design.

4.4 Instrumentation
The persuasiveness of the ads was measured on the basis of attitude towards the product, attitude towards the brand, and purchase intention. Attitude towards the ad was measured separately from persuasiveness (cf. Hornikx & O’Keefe, 2009).

Attitude towards the product was measured using four 5-point semantic differentials: good – bad, low – high quality, inattractive – attractive, and effective – ineffective (α = .78). Attitude towards the brand was measured using four 5-point semantic differentials: positive – negative, unreliable – reliable, good – bad, and expert – inexpert (α = .83). Purchase intention was measured with 5-point Likert scales that followed three statements: “I would like to receive more information about this nasal spray”, “I consider buying this nasal spray”, and “I would definitely buy this nasal spray if I needed nasal spray” (α = .76). Attitude towards the ad was measured using four 5-point semantic differentials: beautiful – ugly, not interesting – interesting, pleasurable – not pleasurable, inattractive – attractive (α = .84). In addition, the reputation manipulation was checked with the same statements that were used in the pretest (α = .86). The questionnaire ended with questions about participants’ age, gender, nationality, and highest education.

4.5 Procedure
Dutch people were invited individually to fill in the questionnaire at different locations in a Dutch city (e.g., railway station, shopping centre, university). People were not rewarded for their participation, which took about 7 to 10 minutes. After the questionnaires had been collected, the real research purpose was revealed, and participants were thanked for their cooperation. There were no disturbances during the experiment.

4.6 Statistical tests
The research question and the hypothesis were evaluated through a 2 (reputation) x 2 (marker) analysis of variance, where reputation and marker were both between-subjects factors.

5. Results
Before addressing the research question and the hypothesis, it was first checked whether the reputation manipulation was successful. As in the pretest, the brand reputation was perceived as higher in the high-reputation conditions (M = 3.92, SD = 0.58) than in the low-reputation conditions (M = 2.37, SD = 0.69); F (1, 134) = 200.97, p < .001, h2 = .60. Furthermore, there was a main effect of reputation on persuasiveness (F (3, 130) = 12.28, p < .001, h2 = .22). For the high-reputation version, the attitude towards the product (F (1, 132) = 21.73, p < .001, h2 = .14), the attitude towards the brand (F (1, 132) = 34.40, p < .001, h2 = .21), and the purchase intention (F (1, 132) = 13.61, p < .001, h2 = .09) were higher than for the low-reputation version. Such a main effect did not occur for the attitude towards the ad: participants’ liking of the ad was not affected by the reputation of the brand (F (1, 132) < 1).

The research question about the persuasive difference between hedges and pledges was answered on the basis of the main effect of marker. There was no main effect of marker on persuasion (F (3, 130) < 1) or on attitude towards the ad (F (1, 132) < 1). It was expected that a hedge would be more persuasive than a pledge in an ad for a low-reputation brand and that a pledge would be more persuasive than a hedge in an ad for a high-reputation brand, but the relevant interaction effect between reputation and marker did not occur, neither for persuasion (F (3, 130) < 1), nor for attitude towards the ad (F (1, 132) < 1). Table 1 gives the descriptive statistics for the dependent measures in the four conditions.

Table 1. Persuasiveness and attitude towards the ad in function of brand reputation and marker

6. Conclusion and discussion
The present study investigated the persuasiveness of hedges and pledges in a realistic product advertisement for a fictituous brand that was presented as having a high or low reputation. The level of reputation was expected to interact with the type of marker. That is, high-reputation brands may benefit more from pledges than from hedges, whereas low-reputation brands may benefit more from hedges than from pledges. The results did not support the hypothesis: there was no interaction effect between reputation (high or low) and marker (hedge or pledge). This occurrence of a non-significant interaction effect cannot be attributed to the manipulation of reputation. In the first place, the manipulation proved to be successful: the high-reputation brand was perceived to have a higher reputation than the low-reputation brand. In the second place, the reputation manipulation affected participants’ response to the subsequently presented ad: ads were found to be more persuasive when they followed the high-reputation journal article than when they followed the low-reputation journal article.

Whereas earlier studies used abstract claims without any context, the present study used a more realistic setting with a fictitious ad, containing text and images, designed for a specific brand. In this context too, hedges and pledges were found to be equally persuasive, corroborating findings reported in Berney-Reddish and Areni (2005, 2006), and Hornikx et al. (2008). Suggestions for future research follow from characteristics of this study. Although having a higher ecological validity, the present study suffers from a low level of generalizability of the results as it involved only one ad. More experimental studies with ads for other products and brands should be conducted before conclusions about a possible relationship between markers and brand reputation can be drawn. Furthermore, it would be wise to also include conditions without markers, so that the persuasiveness of hedges and pledges can be assessed: are claims with markers more or less persuasive than claims without any marker?

REFERENCES
Berney-Reddish, I. A., & Areni, C. S. (2005). Effects of probability markers on advertising claim acceptance. Journal of Marketing Communications, 11(1), 41-54.
Berney-Reddish, I. A., & Areni, C. S. (2006). Sex differences in responses to probability markers in advertising claims. Journal of Advertising, 35(2), 7-16.
Cruz, M. G. (1998). Explicit and implicit conclusions in persuasive messages. In M. Allen, & R. W. Preiss (Eds.), Persuasion: Advances through meta-analysis (pp. 217-230). Cresskill, NJ: Hampton Press.
Darley, W. K., & Smith, R. E. (1993). Advertising claim objectivity: Antecedents and effects. Journal of Marketing, 57(4), 100-113.
Eemeren, F. H. van, & Grootendorst, R. (1992). Argumentation, Communication, and Fallacies: A Pragma-Dialectical Perspective. Hillsdale, NJ: Lawrence Erlbaum Associates.
Eemeren, F. H. van, & Grootendorst, R. (2004). A Systematic Theory of Argumentation: The Pragma-Dialectical Approach. Cambridge: Cambridge University Press.
Goldberg, M., E. & Hartwick, J. (1990). The effects of advertiser reputation and extremity of advertising claim on advertising effectiveness. Journal of Consumer Research, 17(2), 172-179.
Hornikx, J., & O’Keefe, D. J. (2009). Adapting consumer advertising appeals to cultural values: A meta-analytic review of effects on persuasiveness and ad liking. Communication Yearbook, 33, 38-71.
Hornikx, J., Pieper, M., & Schellens, P. J. (2008). Versterkende, afzwakkende en numerieke markeringen in claims over cosmetica-producten: Maken ze claims overtuigender? Tijdschrift voor Communicatiewetenschap, 36(1), 3-14.
Kardes, F. R. (1988). Spontaneous inference processes in advertising: The effects of conclusion omission and involvement on persuasion. Journal of Consumer Research, 15(2), 225-233.
O’Keefe, D. J. (1997). Standpoint explicitness and persuasive effect: A meta-analytic review of the effects of varying conclusion articulation in persuasive messages. Argumentation and Advocacy, 34(1), 1-12.
O’Keefe, D. J. (2002). The persuasive effects of variation in standpoint articulation. In F. H. van Eemeren (Ed.), Advances in pragma-dialectics (pp. 65-82). Amsterdam: Sic Sat.
O’Keefe, D. J. (2007). Potential conflicts between normatively-responsible advocacy and successful social influence: Evidence from persuasion effects research. Argumentation, 21(2), 151-163.
Sawyer, A. G., & Howard, D. J. (1991). Effects of omitting conclusions in advertisements to involved and uninvolved audiences. Journal of Marketing Research, 28(4), 467-474.