THE RELATIVE EFFECTIVENESS OF COMPARATIVE AND NONCOMPARATIVE
ADVERTISING: Evidence for Gender Differences in Information-Processing Strategies
Chang, Chingching. Journal of Advertising 36. Â1 (Spring 2007): 21-35.
_______________________________________________________________
Abstract
Drawing on gender differences in processing strategies and elaboration thresholds, this study showed that comparative advertising encouraged greater levels of brand-evaluation involvement among men but not among women. Instead, for women, attentiongaining comparative appeals encouraged inferences regarding the ad's manipulative intent.
Furthermore, these gender differences in processing led to evaluative consequences. For men, the greater brand-evaluation involvement brought about by comparative appeals led to more favorable ad and brand evaluations and greater purchase intentions. For women, the heightened perceptions of manipulative intent brought about by comparative appeals resulted in negative ad and brand evaluations and reduced purchase intentions. Finally, the attribute alignability of competing products moderated the gender-dependent effectiveness of comparative and noncomparative appeals. [PUBLICATION ABSTRACT]
_______________________________________________________________
Full Text
Headnote
ABSTRACT:
Drawing on gender differences in processing strategies and elaboration thresholds, this study showed that comparative advertisingencouraged greater levels of brand-evaluation involvement among men but not among women. Instead, forwomen, attentiongaining comparativeappeals encouraged inferences regarding thead's manipulative intent.
Furthermore, these gender differences in processing led to evaluative consequences. For men, thegreater brand-evaluation involvement brought about by comparativeappeals led to more favorable ad and brandevaluations and greater purchase intentions. Forwomen, the heightened perceptions ofmanipulative intent brought about by comparativeappeals resulted in negative ad and brandevaluations and reduced purchase intentions. Finally, theattribute alignability ofcompeting products moderated thegender-dependent effectiveness of comparativeand noncomparative appeals.
Comparativeappeals are commonly employed in advertisingand are believed to have many strategic advantages over noncomparative appeals. Forexample, comparative advertising can effectively encourage consumers to engage in relative judgments, generating either an associationor differentiationeffect (Droge and Darmon 1987; Pechmann and Ratneshwar 1991; Pettit-O'Malley and Johnson 1992). However, persuasion effects of comparative advertisingare not always superior to those ofnoncomparative advertising(Grewal et al. 1997). Therelative effectiveness of comparative versusnoncomparative advertisingmay vary as a function ofproduct characteristics (Putrevu and Lord 1994) or message content (Grewal et al. 1997). Thepresent study postulates that therelative effectiveness of comparative versus
noncomparative advertisingappeals also partly depends on thegender of thead perceiver.
Men and women differ in their information-processing strategies. Men are often seen as
"selective processors" who rely on heuristics or overall themes, whereas women are
categorized as "comprehensive processors" who focus on integrating detailed information (Meyers-Levy 1989; Meyers-Levy and Maheswaran 1991; Meyers-Levy and Sternthal 1991).
In addition, attention and elaboration thresholds forwomen are lower than those formen
(Meyers-Levy and Sternthal 1991). Based on this research, thepresent paper predicts that men will be less likely to engage in involved brandevaluation unless they are exposed to comparativeappeals, which have been shown to draw more attention and enhance message elaboration (Muehling, Stoltman, and Crossbare 1990). In contrast, due to their low elaboration thresholds, women will engage in involved brandevaluation regardless of the use of comparativeor noncomparative appeals.
Attention-drawing tactics may not be desirable forcomprehensive processors. Past research indicated that attentiongetting advertisingtactics sometimes lead consumers to infer that the advertiser is attempting to manipulate thead perceivers (Campbell 1995). However, inferring manipulative motives requires greater cognitive resources (Campbell and Kirmani 2000).
Therefore, it is proposed that comparativeappeals are more likely to encourage inferences of manipulative intent forwomen, who are comprehensive processors and have low elaboration thresholds, than formen, who are selective processors and have high elaboration thresholds.
It is predicted that these gender differences in processing will lead to divergent evaluative outcomes. Formen, increased brand-evaluation involvement brought about by comparative appeals will lead to more favorable ad and brandevaluations. Forwomen, increased inferences ofmanipulative intent brought about by comparativeappeals will result in lower ad and brandevaluations.
Recent research has suggested that product comparison is conducted through an alignment process involving three properties: commonality, alignable differences, and nonalignable
differences (Zhang and Fitzsimons 1999; Zhang, Kardes, and Cronley 2002; Zhang and
Markman 1998, 2001). Commonalities are matching attributes. Alignable differences are
"features that are comparable along thesame dimension" (Zhang and Markman 1998, p.
413); nonalignable differences are unique features ofone brandthat do not correspond with those of theother brand.
Thedifficulty ofcomparison increases as thealignability of theattributes decreases (Zhang and Fitzsimons 1999). Past research showed that themore complex an ad is, theless likely it is that ad perceivers will be able to elaborate on messages (Yalch and Elmore-Yalch 1984). For this reason, it is possible that thedegree ofalignability ofproduct attributes fortwo competing brandsmoderates therelationship between gender and theeffectiveness of comparativead appeals. As processing difficulty increases, comparativead appeals may cease to encourage men to engage in message elaboration or women to infer manipulative intent. Given this, the effectiveness of comparativeappeals formen and thenoncomparative appeals forwomen in
elevating ad and brandevaluations and purchase intentions may be attenuated when the two competing ads feature nonalignable, as opposed to alignable, product attributes. This paper will explore these hypotheses in a competitive adprocessing context, where competing ads are presented. This makes comparison ofattribute alignability possible regardless of the use of comparativeappeals. Indeed, thead-exposure context formagazine ads can be generally characterized as a context cluttered with ads forcompeting brands.
GENDER AND INFORMATION PROCESSING
Gender differences in performances ofcognitive tasks have been well documented in past research. Forexample, Klinteberg, Levander, and Schalling (1987) found that women prefer a reflective-sequential problem-solving strategy, whereas men prefer an impulsive-global strategy. In another study, men were more likely than women to adopt hypothesis-confirming strategies (Chung and Monroe 1998). Duff and Hampson (2001) showed that women make fewer errors than men when performing tasks that demand working memory. Gender differences have also been shown forinformation-processing strategies. Studies have found that women usually engage in a comprehensive, piecemeal-style processing, whereas men tend to engage in selective modes ofprocessing, relying on readily available heuristic cues (Meyers-Levy 1989; Meyers-Levy and Maheswaran 1991; Meyers-Levy and Sternthal 1991).
Meyers-Levy and Sternthal (1991) proposed a selectivity hypothesis that suggests that
genders differ in their elaboration thresholds. In light ofgender differences in information
processing, this hypothesis proposes that because women have lower thresholds than men, only a limited amount ofattention is required forthem to trigger an elaboration. These differences may explain why studies have found that women encode more detailed information than men (Seidlitz and Diener 1998), as well as greater recall ofnew information (Erngrund, Mantyle, and Nissan 1996). In addition, subtle changes in message order have been shown to encourage women to engage in more extensive message elaboration than men (Meyers-Levy and Sternthal 1991, Experiment Two), suggesting that women have lower elaboration thresholds than men.
Gender differences in elaboration thresholds may also lead men and women to respond
differently to incongruent messages. Forexample, use ofa moderately incongruent message, which is assumed to pass women's elaboration thresholds but not men's, resulted in women generating more accurate judgments than men (Meyers-Levy and Sternthal 1991, Experiment One). Moreover, women performed well on recognition tasks across both high and low-incongruity contexts, whereas men performed well only when incongruity was high (Meyers-Levy and Maheswaran 1991). These findings suggest that women may be constantly sensitive to relevant information, and therefore motivated to engage in message elaboration even when message incongruity is low. Men, on theother hand, may not be motivated to engage in detailed message elaboration unless encouraged to do so by situational factors, such as a high degree ofmessage incongruity.
Findings in advertisingand consumer behavior research are also consistent with theview that there are gender differences in information-processing strategies and thresholds for
elaborative processing. Forexample, Darley and Smith (1995) found that women's adprocessing strategies are more likely to be altered by subtle product changes than men's, which could be due to lower attention thresholds. In another study, women engaged in greater levels ofad elaboration than did men (Krugman 1966). Moreover, in shopping contexts, Laroche et al. (2000) demonstrated that women are involved in comprehensive and intensive information searches, whereas men are only engaged in selective information searches. Women also showed greater use ofavailable macro- and microinformation than men when shopping in a store (Cleveland et al. 2003). Due to these differences, women will be more involved in brandevaluations when reading advertisements than men, regardless of what advertisingappeals are used.
COMPARATIVEAPPEALS AND BRAND-EVALUATION INVOLVEMENT FORDIFFERENT GENDERS
Past research has paid little attention to gender differences in responses to comparative
advertisingappeals. A notable exception is theresearch conducted by Pfau (1994), who found that for men, moderate-intense comparative advertisingmore effectively generated favorable brandattitudes fora highinvolving product, but high-intense comparative advertisingmore effectively generated positive brandaffect fora low-involving product. Forwomen, however, moderate-intense comparative advertisingwas more effective than high-intense comparative advertisingin generating message agreement fora low-involving product. Drawing on the selectivity hypothesis, Pfau concluded that these differences could be explained by the fact that, as compared with women, more stimulation is needed to motivate men to engage in message elaboration.
Considering theprocessing differences described above, gender may play an important role in determining therelative effectiveness of comparativeand noncomparative advertising appeals, although this question has yet to be thoroughly explored. According to theselectivity hypothesis, low-involving noncomparative, attribute-based advertisingappeals may not exceed men's elaboration thresholds, and therefore may fail to motivate involvement in brand evaluations. In contrast, comparativeappeals, which are attention drawing (Muehling, Stoltman, and Grossbart 1990), should be able to exceed men's elaboration thresholds, motivating them to elaborate on thead messages. Therefore, comparativeappeals should be more likely than noncomparative appeals to engage men in brandevaluation. Forwomen, however, presenting competing products in thesame context may attract sufficient attention to render thead messages above their thresholds ofelaboration. In other words, both comparativeand noncomparative attribute-based appeals should exceed women's elaboration thresholds. Therefore, comparativeand noncomparative advertisingshould involve women in brandevaluations to thesame degree.
H1: Brand-evaluation involvement in a competitive-ad context is influenced by a significant interaction between gender and ad appeal type. Use of comparativeappeals encourages more involvement in brandevaluations than use ofnoncomparative attribute-based appeals formen but not forwomen.
COMPARATIVEAPPEALS AND INFERENCES OFMANIPULATIVE INTENT FOR
DIFFERENT GENDERS
Increases in attention sometimes evoke negative responses, such as inferences of
advertisers' manipulative intents, which are defined as "consumer inferences that the
advertiser is attempting to persuade by inappropriate, unfair, or manipulative means"
(Campbell 1995, p. 226). It is believed that consumers are aware of thepsychology of
persuasion and advertising, and this knowledge may affect how they respond to ad
messages (Friestad and Wright 1994,1995). This persuasion knowledge includes ideas
regarding what persuasion tactics are appropriate or fair (Friestad and Wright 1994). When persuasion tactics are not used appropriately, they are especially likely to encourage ad perceivers to ponder over why such tactics are employed and, in theprocess, will elicit perceptions ofadvertisers' manipulative intents (Campbell 1995).
Indeed, low ad credibility has been shown to be associated with greater levels ofperceived manipulative intent (Cotte, Coulter, and Moore 2005). Even though skepticism toward advertisingis a general phenomenon (Boush, Friestad, and Rose 1994), it is more magnified when certain tactics are employed. Forexample, comparativeappeals were rated lower in source credibility (Grewal et al. 1997) and were shown to elicit greater levels ofskepticism (Prasad 1976) and counterarguments (Wilson and Muderrisoglu 1980). Jain, Buchanan, and Maheswaran (2000) also found that compared with noncomparative appeals, comparative advertisingappeals are responsible formore negative attributive thoughts, which have been shown to be positively correlated with levels ofperceived manipulative intent (Cotte, Coulter,
and Moore 2005).
Moreover, Campbell (1995) argued that attention-getting advertisingtactics may be more likely to generate inferences ofmanipulative intent. In a similar vein, Ahluwalia and Burnkrant (2004) demonstrated that persuasion tactics that deviate from expectations motivate participants to pay more attention to thesources. Thus, it is reasoned that the
attentiondrawing characteristics of comparativeappeals may encourage inferences of
manipulative intent.
However, it is important to note that Campbell and Kirmani (2000) found that individuals'
cognitive capacities influence their inferences regarding persuasion agents' motives.
Participants were more likely to usepersuasion knowledge to infer that persuasion agents were being manipulative when they were not cognitively busy than when they were cognitively busy.
In parallel, it is argued in this paper that comparativeappeals will elicit inferences of
manipulative intent only if participants are comprehensive ad processors. Specifically,
because women are considered to be "comprehensive information processors" with low
elaboration thresholds, as noted above, the use ofdirect comparativeappeals to persuade women is probably not necessary, and may even evoke negative responses. In other words, if women are already motivated to elaborate on ad messages, attention-getting ad tactics, such as comparativeappeals, are likely to encourage inferences ofmanipulative intent. In contrast, formen, whose thresholds are higher, direct comparativeappeals should only encourage elaboration on product information and help them to see thesuggested benefits of theadvertised product without generating greater inferences regarding manipulative intent.
H2: Perceived manipulative intent ofads is influenced by a significant interaction between gender and ad appeal type. Forwomen, but not formen, comparativeappeals generate greater inferences regarding manipulative intent than do noncomparative appeals.
COMPARATIVEAPPEALS AND PERSUASION FORDIFFERENT GENDERS
Past research has indicated that participants' inferences regarding manipulative intent have evaluative consequences-thehigher thelevel ofperceived manipulative intent, theless persuasive themessage. In one study, forexample, greater perceived manipulative intent encouraged negative attributions of theadvertisers and led to more negative ad attitudes (Cotte, Coulter, and Moore 2005). Furthermore, negative ad attitudes induced by inferences about manipulative intent have been shown to result in more negative brandattitudes and purchase intentions (Campbell 1995). Similarly, Chang (2001) found that perceived manipulative intent influences ad evaluations, which, in turn, influence brandevaluations. Finally, research has demonstrated that when comparative advertisingis not used appropriately, it can evoke negative attributions toward advertisers and attenuate ad believability and brandattitudes (Jain and Posavac 2004). Forthis reason, in comparative advertisingsituations, where inferences ofmanipulative intent are encouraged, women are expected to have more negative evaluations of thead and brand, as well as lower purchase intentions.
On theother hand, drawing on theargument that message elaboration can lead to greater persuasion when message arguments are cogent (Cialdini, Petty, and Cacioppo 1981; Petty and Cacioppo 1983), it is reasoned in this paper that enhanced attention among men, who are selective information processors, can increase theopportunity formessage elaboration and allow them to understand thebenefits ofa product. Specifically, themore men's brandevaluation involvement is enhanced by comparativeappeals, themore likely they will respond favorably toward thead and brand.
H3: Ad attitudes (H3a), brandattitudes (H3b), and purchase intentions (H3c) are influenced by a significant interaction between gender and ad appeal type. Forwomen, noncomparative appeals are more effective than comparativeappeals, whereas formen, thereverse is true.
ATTRIBUTE ALIGNABILITY
Thestructural alignability of theattributes ofcompeting brandsmay alter thecomplexity of
processing product information. Prior research suggests that product comparison involves a process ofdetermining thecommonalities, alignable differences, and nonalignable differences between a pair ofproducts (Zhang and Fitzsimons 1999; Zhang, Kardes, and Cronley 2002; Zhang and Markman 1998, 2001).
According to Zhang and Markman (1998), "alignable differences are like commonalities, in that they are corresponding elements ofa pair, but they are also like differences, in that they are unlike elements that correspond" (p. 414). Because ofthis dual role, alignable differences are more likely to be a focus ofcomparison and to lead to judgmental consequences, as illustrated by thefinding that alignable differences are easier to recall than nonalignable differences (Zhang and Markman 1998). In addition, Zhang and Markman (2001) found that an alignable better brandis preferable to a nonalignable better brand.
Forcomparison, alignable differences may be used more than nonalignable differences
because processing theformer involves less cognitive effort. Forexample, alignable
differences were perceived as easier to compare than nonalignable differences (Zhang and Fitzsimons 1999). Moreover, ratings ofcomparison difficulty have been found to increase as attribute alignability decreases (Zhang, Kardes, and Cronley 2002).
In general, comparing nonalignable product attributes is a more complex task than
comparing alignable product attributes. Complexity ofinformation has been shown to exhaust individuals' cognitive resources and reduce their ability to pay attention to messages (Anderson and Jolson 1980) or elaborate on messages (Yalch and Elmore-Yalch 1984). Past research has documented that owing to men's lower likelihood to engage in comprehensive processing, they generate less accurate judgments than women when task complexity is high (Chung and Monroe 2001) and express less favorable attitudes than women toward ads that contain complex information (Putrevu, Tan, and Lord 2004). Leigh and Menon (1987) also showed that thesuperior learning effects ofincreasing ad-processing involvement are attenuated when message complexity is high. Therefore, when nonalignable attributes are featured, men may find it too complex to discern thedifferences between thetwo competing brandseven in thepresence ofattentiongetting comparativead appeals. Forthis reason, these
types of comparativeappeals will not effectively increase men's brandevaluation involvement.
However, when they are exposed to ads featuring alignable differences, which are less
complex to process, comparativeappeals may motivate them to be more involved in brand evaluation than would noncomparative appeals. Because women's elaboration thresholds are low, their brand-evaluation involvement may not be influenced by ad appeal type or attribute alignability.
H4: Brand-evaluation involvement is affected by a three-way interaction among attribute
alignability, gender, and ad appeal type. Forwomen, thetwo-way interaction between
alignability and ad appeal type is not significant. Formen, comparativeappeals induce greater brand-evaluation involvement than do noncomparative appeals only when thead features alignable attributes.
Past research has demonstrated that when processing simple information, increasing
participants' involvement results in participants producing more counterarguments (Wright 1974). However, when processing information that demands higher cognitive capacity, increasing participants' involvement had no impact on thenumber ofcounterarguments they generated. Theinclusion ofnonalignable attribute information may increase thecognitive resources required to process an ad message. Forwomen, if nonalignable attribute information is more difficult to process than alignable attribute information, comparativeads with nonalignable differences may not evoke more inferences ofmanipulative intent than noncomparative ads with nonalignable differences. Nevertheless, because men's elaboration thresholds are low, inferences regarding manipulative intent, which require more cognitive resources, may not be likely regardless ofad appeal type or attribute alignability.
H5: Inferences ofmanipulative intent are influenced by a threeway interaction among attribute alignability, gender, and ad appeal type. Forwomen, comparativeappeals induce greater perceived manipulative intent than do noncomparative appeals only when thead features alignable attributes. In contrast, formen, thetwo-way interaction between ad appeal type and attribute alignability is not significant.
In the same vein, attribute alignability likely moderates theinteraction between ad appeal type and gender forjudgmental consequences. Specifically, thegreater perceived manipulative intent evoked by processing alignable comparativeads should lead women to report less favorable ad and brandevaluations than they would after processing alignable noncomparative ads. When ads focus on nonalignable attributes, however, ad appeal type should have no effect on inferences ofmanipulative intent, and thus should also have no effect on ad and brandevaluations. Formen, on theother hand, greater brand-evaluation involvement encouraged by processing alignable comparativeads may lead to more positive ad and brandevaluations than when processing alignable noncomparative ads. Nonalignable comparativeads, however, should not effectively increase men's brand-evaluation involvement. As a result, theeffect ofad appeal type on ad and brandevaluations should also be attenuated formen when nonalignable attributes are featured.
H6: Ad attitudes (H6a), brandattitudes (H6b), and purchase intentions (Hoc) are influenced by a three-way interaction among attribute alignability, gender, and ad appeal type in a competitive-ad context. Forwomen, noncomparative appeals are more effective than comparativeappeals forads featuring alignable attributes, but not forads featuring nonalignable attributes. Formen, comparativeappeals are more effective than
noncomparative appeals forads with alignable attributes, but not forads with nonalignable attributes.
METHOD
Design
Thestudy featured three factors. Thetwo manipulated factors were ad appeal type (
comparative versusnoncomparative) and attribute alignability (alignable versus
nonalignable). The third factor was gender Participants, Stimuli, and Procedures
One hundred and eighteen college students were recruited from a metropolitan area in the Pacific Rim and paid fortheir participation. Fifty-three percent were female. Each participant was randomly assigned to one of thefour experimental conditions.
Theproduct selected was an electronic dictionary, a product commonly used by college
students in this region. Since thepurpose was to improve understanding of therelationship between gender and advertisingeffectiveness, it was important to select a product commonly purchased and used by both men and women. Thepercentage ofmale (80%) and female (85%) participants who owned theproduct did not differ significantly, ^sup 2^(1, n = 118) = .62, p = .29. In addition, ANOVAs (analyses ofvariance) showed that ownership did not have a significant impact on any of thedependent variables (all ps >.50), and theinteraction between gender and ownership was not significant forany of thedependent variables either (all ps >.42).
After reading brief instructions, participants viewed four ads in thefollowing order: a filler ad; an ad for theleading competitor, Besta; an ad for thetarget brand, AST; and finally, a second filler ad. They then evaluated theads and brandsin terms of thedependent variables described below and responded to a manipulation check.
Thestimuli ads were created by ad agency professionals. Visuals and layouts for the Besta and AST brandswere kept similar to reduce any possible confounding effects. In a pretest, participants (n = 35) were asked to focus on thegraphical aspects of theads and to rate their agreement with thefollowing statements: "Thevisual part of thead is likable," and "thevisual part of thead is attractive." Participants made ratings on a seven-point Likert scale with higher scores indicating greater agreement. Thetwo items were significantly correlated (for the Besta ad, Cronbach's = .87; for theAST ad, Cronbach's = .89), and were therefore averaged. TheBesta and AST ads did not differ significantly on visual appeal scores, t(34) = 1.49, p = .14 (M^sub Besta^ = 3.72, SD = .99; M^sub AST^= 3.49, SD = .93). Although the ad copy for theAST brandwas manipulated, thegraphics remained constant across conditions. Both thecopy and graphics for theBesta ad were held constant across conditions.
Independent Variables Ad Appeal Type The comparative-appeal ad listed theattributes of AST side by side with those ofBesta. The noncomparative-appeal ad simply listed theproduct attributes of AST without referring to Besta. As a manipulation check, participants rated on a seven-point Likert scale thedegree to which they agreed with two statements: "Thead compares thetarget brandwith a competitor's brand," and "thead shows thesuperiority of thetarget brandto a competitor's brand." Higher scores indicated greater agreement. Agreement was significantly higher for comparativeads than noncomparative ads forboth thefirst statement, F(1, 110) = 81.12, p = .01 (M^sub com^
= 5.14, SD = 1.66; M^sub noncom^ = 2.61, SD = 1.36), and thesecond statement, F(1, 110) = 35.45, p = .01 (M^sub com^ = 4.53, SD = 1.73; M^sub noncom^ = 2-75, SD = 1.52).
Therefore, themanipulation was deemed satisfactory.
Attribute Alignability
Zhang and Markman have stated that alignable differences can arise when "aspects of each item are placed in correspondence on thebasis ofsome matching relation between the objects" (2001, p. 14) and when "thecorresponding elements occupy different points on the same dimensions" (p. 14). Thefocus here was exclusively on thelatter situation. In a product comparison context, nonalignable differences can refer to attributes present in one product but not in theother (missing attributes) or to attributes that are uniquely associated with one brandbut not theother (unique attributes). Both types were investigated here.
Alignability was manipulated fortwo focal attributes that college students usein selecting an electronic dictionary: number ofword entries and number oflanguages. All ads also featured three common attributes that are less important: human-voice pronunciation, handwriting pad, and leather case. Attributes were chosen by a pretest (n = 35) in which participants used a seven-point scale to rate all attributes ofavailable brandson themarket according to theattribute's importance to them when making a purchase. Higher scores indicated greater importance. Thefocal attributes did not differ significantly from each other on ratings of importance, but both were rated significantly more important than each of thethree common attributes (see Table 1).
In addition to thethree common attributes present in all ads, each Besta ad also featured number ofvocabulary. In thealignable condition, theAST ad also featured number of vocabulary (with a higher number than Besta). In themissing nonalignable condition, the AST ad featured both focal attributes. Therefore, in comparison to this ad, theBesta ad was missing thenumber oflanguages attribute. In theunique nonalignable difference condition, the AST ad featured number oflanguages. Hence, in this case, number ofvocabulary was unique to Besta and number oflanguages was unique to AST. Theattributes present in each condition are listed in Table 2. Since thetwo nonalignable types (missing nonalignable and unique nonalignable) exerted no significantly different impacts on any of thedependent measures (all ps >.17), responses to them were collapsed in thefollowing analyses.
Dependent Variables
All scale items were rated on a seven-point Likert scale. Item scores foreach scale were
averaged, with higher scores indicating stronger agreement with themeasured construct.
Brand-Evaluation Involvement
Brand-evaluation involvement was assessed by thefollowing six items adopted from the
brand-evaluation strategy scale used by Laczniak and Muehling (1993): "I paid attention to what was stated in thead so I could evaluate theadvertised ad," "I paid attention to what was stated in thead to help me evaluate the brandfeatured in it," "I paid attention to what was stated in thead so that I could determine theattributes of the brandfeatured in it," "I paid attention to what was stated in thead so that I could determine thebenefits of the brand featured in it," "I paid attention to what was stated in thead so that I could rate thequality of the brandfeatured in it," and "I paid attention to what was stated in thead so that I could determine what the brandfeatured in it had to offer." Internal reliability was deemed satisfactory for thetarget ad (Cronbach's = .86).
Inferences of Manipulative Intent Six items were used to measure inferences of themanipulative intent of thead. Thesix items were taken from Campbell (1995): "Theway this ad tries to persuade people seems acceptable to me," "theadvertiser tried to manipulate theaudience in ways that I don't like," "I was annoyed by this ad because theadvertiser seemed to be trying to inappropriately manage or control theconsumer audience," "I didn't mind this ad: theadvertiser tried to be persuasive without being excessively manipulative," "this ad was fair in what was said and shown," and "I think that this advertisement is unfair." Cronbach's forthis scale was .89 for thetarget ad, indicating satisfactory reliability. Thefirst, fourth, and fifth items were reversescored.
Ad Attitudes Participants indicated thedegree to which they agreed that thead was "interesting," "good," "likable," "not irritating," and "pleasant." These items were adopted from Madden, Alien, and Twible (1988) and Mitchell and Olson (1981). Internal reliability of theitems was deemed satisfactory for thetarget ad (Cronbach's = .92).
BrandAttitudes
Participants indicated thedegree to which they agreed that the brandwas "good," "likable," "pleasant," "positive," and "high quality." These items were adopted from Chang (2002).
Scale reliability was deemed satisfactory for thetarget brand(Cronbach's = .94).
Purchase Intentions Three items ("probably," "likely," and "possibly") were adopted from Zhang (1996) to assess how likely participants were to purchase theproduct in thefuture. Cronbach's for this scale was .92 for thetarget brand, indicating satisfactory reliability.
ANALYSES AND RESULTS
One assumption ofthis study is that women are more involved in brandevaluations than men.
Supporting this assumption, ANOVA showed that female participants reported higher brand evaluation involvement ratings than male participants, F(1, 110) = 4.54, p = .04 (M^sub female^ = 3.03, SD = 1.30; M^sub male^ = 2.69, SD = 1.20; see Table 3).
The interaction between gender and ad appeal type for brand-evaluation involvement only approached significance, F(I, 110) = 3.49, p = .06. Since H1 was theory driven and
developed a priori, however, further simple effects tests were conducted (see Winer, Brown, and Michels 1991). As expected, results ofsimple effects tests (see Table 4) revealed that the effect ofad appeal type was not significant forwomen, F(1, 58) = .21,p= .65. For men, however, ad appeal type was significant, F(1, 52) = 10.56,p - .01, with comparative advertisingincreasing theextent of brand-evaluation involvement. Therefore, theresults of simple effects tests were consistent with thepredictions, providing partial support for
H1. Consistent with another assumption ofthis study, themain effect ofappeal type on inferences ofmanipulative intent was significant, F(1, 110) = 4.43, p = .04. Themeans were in the expected directions, with comparativeappeals generating higher ratings than noncomparative appeals (M^sub com^ = 3.73, SD = 1.14, M^sub noncom^ = 3.24, SD = 1.20). For inferences of manipulative intent, the interaction between gender and ad appeal type was significant, F(1, 110) = 5.82, p = .02. As expected, simple effects tests showed that the effect ofad appeal type was significant forwomen, F (1, 58) = 10.87, p = .01, with comparative advertisingincreasing inferences ofmanipulative intent. In contrast, formen, the effect of ad appeal type was not significant, F(1, 52) = .08, p = .78. These results fully supported
H2. It is also important to note that the difference between women's and men's responses to comparative advertising approached the significance level, F(1, 54) = 3.74, p = .06, suggesting that women expressed relatively higher levels of manipulative intent inferences than men when exposed to comparative advertising.
As expected, theinteraction between gender and ad appeal type on ad attitudes was also significant, F(1, 110) = 11.80. Forwomen, simple effects tests indicated that theinfluence of ad appeal type was significant, F(1, 58) = 6.21, p = .02, and as predicted, attitudes were more favorable for noncomparative ads than for comparativeads. For men, the effect was also significant, F(1, 52) = 5.18, p = .03, with means in the opposite direction from those for the women. These results fully supported H3a. The interaction between gender and ad appeal type was significant for brandattitudes as well, F(1, 110) = 8.70, p = .01. Simple effects tests revealed that theeffect ofad appeal type was significant forwomen, F(1, 58) = 4.16, p = .05, with means in thepredicted directions. As expected, formen, the effect of ad appeal type was also significant, F(1, 52) = 4.21, p= .05, with means in the predicted directions. Therefore, H3b was also fully supported. For purchase intentions, as expected, the twoway interaction between gender and ad appeal type was significant, F(1, 110) = 7.90, p= .01. Simple effects tests showed that the effect of ad appeal type was not significant forwomen, F(1, 58) = 1.81, p = .18, even though the means were in the predicted directions. Formen, the effect ofad appeal type was significant, with
Comparative appeals generating higher ratings of purchase intentions, F(1, 52) = 6.30, p = .01. Even though the interaction was significant, results of simple effects tests supported H3c for men but not for women.
As expected, the threeway interaction among attribute alignability, gender, and ad appeal type was significant for brand-evaluation involvement, F(1, 110) = 5.36, p = .02. Consistent with expectations, the two-way interaction between alignability and ad appeal type was not significant for women, F(I, 58) = 1.34, p = .25, but was significant formen, F(I, 52) = 5.65, p = .02. Further simple effects tests (see Table 5) showed that comparative appeals improved men's brand evaluations when alignable attributes were featured, F(1, 15) = 19-79, p = .01, but not when nonalignable attributes were featured, F(I, 37) = .48,p = .49. Theresults fully supported H4. Also as expected, for inferences of manipulative intent, the three-way interaction a mong attribute alignability, gender, and ad appeal type was significant, F(1, 110) = 3-99,p = .05. For men, the two-way interaction between alignability and ad appeal type was not significant, F(I,
52) = .80, p = .38, which was as expected, but forwomen, the interaction only approached significance, F(1, 58) = 2.81, p = .10. However, simple effects tests indicated that comparative appeals increased women's perceptions of manipulative intent when alignable attributes were featured, F(1, 20) = 9-27, p = .01, but not when nonalignable attributes were featured, F(1, 38) = 1.09, p = .30. Therefore, even though the interaction forwomen was not significant, results of simple effects tests were consistent with expectations, providing partial support for H5. Forad attitudes, the three-way interaction was again significant, F(1, 110) = 8.01, p = .01.
Consistent with predictions, for women, thetwo-way interaction between alignability and ad appeal type was significant, F(I, 58) = 5.41, p = .02. Comparativead appeals adversely affected their ad attitudes when alignable attributes were featured, F(1, 20) = 6.89, p = .02, but not when nonalignable attributes were featured, F(1, 38) = .02, p = .89. For men, thetwoway interaction was not significant, F(1, 52) = 2.32, p = .13- However, further simple effects tests (justified by the theory-driven, a priori nature of the hypothesis) indicated that comparativead appeals improved men's ad attitudes when alignable attributes were featured, F(1,15) = 6.63, p = .02, but not when nonalignable attributes were featured, F(1, 37) = .40,p =.53. For women, both the two-way interaction and simple effects tests were consistent with expectations, whereas formen, thetwo-way interaction was not significant, even though the results ofsimple effects tests were as expected. Theresults thus provide only partial support forH6a.
With regard to brand attitudes, the three-way interaction among attribute alignability, gender, and ad appeal type was significant, F(1,110) = .86,p = .01. As expected, the two-way interaction between alignability and ad appeal type forfemale participants was significant, F(I, 58) = 5.60, p = .02. Simple effects tests revealed that comparativead appeals negatively influenced women's ad attitudes when alignable attributes were featured, F(1,20) = 7.33, p = .01, but not when nonalignable attributes were featured, F(I, 38) = .12, p = .74. Formen, the two-way interaction was not significant, F(1, 52) = 2.53, p = .12. However, simple effects tests showed that comparativead appeals marginally improved men's brand attitudes when alignable attributes were featured, F(I, 15) = 4.42, p = .06, but not when nonalignable attributes were featured, F(I, 37) = .13, p = .72. Similar to findings for H6a, for women, both the two-way interaction and simple effects tests were consistent with expectations, where as for men, the two-way interaction was not significant, even though the results of simple effects tests were as expected. Therefore, results provide only partial support for H6b. The three-way interaction was significant for purchase intentions as well, F(I, 110) = 4.03, p = .05. The two-way interaction between alignability and ad appeal type forfemale participants
was not significant, F(I, 58) = 2.46, p = .12. However, simple effects tests showed that, as expected, comparativead appeals reduced women's purchase intentions when alignable attributes were featured, F(I, 20) = 5.00, p = .04, but not when nonalignable attributes were featured, F(I, 38) = .01, p = .96. The two-way interaction was not significant formen either, F(I, 52) = 1.10, p = .30, but according to simple effects tests, comparativead appeals enhanced men's purchase intentions when alignable attributes were featured, F(1,15) = 5.99, p = .03, but not when nonalignable attributes were featured, F(I, 37) = 1.53, p = .22. Even though the two interactions formen and women were not significant, the results of simple effects analyses were consistent with expectations, thus providing partial support forHoc.
DISCUSSION
Gender and attribute alignability were examined in this study as two important variables that moderate the elaboration enhancing influence of comparative advertising appeals. Given that gender is one of themost important bases formarketing segmentation, the influence of gender on ad responses has been an important research topic. Relevant gender research can be classified as fitting into one oftwo paradigms. Thefirst research paradigm focuses on self congruency effects, showing that appeals that are gender-role congruent are more effective than gender-role incongruent appeals (Maldonado, Tansuhaj, and Muehling 2003; Prakash 1992; Wang et al. 2000). It is believed that socialization shapes the unique self-construals of men versuswomen, and thereby alters theway they interpret and respond to ad messages (Wang et al. 2000). The second research paradigm has been driven by theories regarding information-processing
differences between the sexes. The present study adds to this body of research. The findings suggest that women, who are thought to be comprehensive processors with low elaboration thresholds, engage in similar degrees of brand-evaluation involvement regardless of ad appeal type, but generate more inferences regarding manipulative intent when comparative appeals are used. On the other hand, comparativead appeals enhance brand-evaluation involvement formen, presumably because only such appeals exceed their higher attention thresholds. Appeal types, however, have no effects on men's inferences regarding manipulative intent
Most important, gender differences in information processing were linked for thefirst time to two important aspects of the process of responding to advertising-brand-evaluation involvement and inferences regarding manipulative intent which influence evaluations of ads and brands(favorably in the case of the former and unfavorably in thecase of thelatter).Consistent with past research, it was found that enhancing message elaboration can produce positive outcomes in some conditions but can backfire in others. These contradictory effects can be explained by variations in brand-evaluation involvement and inferences of manipulative intent. This study establishes these two dependent variables as important ad responses, and they deserve more research attention in the future. Findings ofthis study counter a generally accepted explanation for gender differences in responses to advertising, namely, that women are taught to be submissive and are there fore more subject to the influences of persuasion (Worchel and Cooper 1976). If this were the case, female participants should have perceived less manipulative intent, thus generating more favorable ad and brand evaluations across conditions, which did not happen. The new evidence suggests that themajor difference between men and women is how they process ad messages, which alters their sensitivity to different appeals and message complexity. This suggests that more research on gender differences in information processing is warranted.
Moreover, although prior work has suggested that learning about a new product is enhanced by greater attribute alignability (Zhang and Markman 1998), thepresent findings suggest that the effect ofalignability may depend on viewer gender. For men, who engage in comprehensive processing less than women, the presence of alignable attributes in comparativeads appears to increase comprehension of therelative advantages of thetarget
brand. On the other hand, the effort to process ads featuring nonalignable attributes appears to make brand-evaluation elaboration difficult, thus decreasing the persuasive effectiveness of such ads. In contrast, women appear to respond more favorably to comparativeads featuring nonalignable attributes than to those featuring alignable attributes, presumably because the complexity ofprocessing nonalignable attributes suppresses their inferences of manipulative intent. These findings may be valuable for advertisers when determining the appropriate product attributes to feature in comparativeads that target product users ofa specific gender.
A review of comparative advertisingresearch showed that the effectiveness of comparative advertising can depend on contingent factors (Grewal et al. 1997). The present study further
demonstrated that the effectiveness of comparative advertising hinges on thead response mechanism it triggers. When comparative advertising involves viewers in brand evaluation, its effectiveness is enhanced. In contrast, when comparative advertising triggers perceptions of manipulative intent, its effectiveness is reduced.
Product judgments usually take place in contexts rich with competing product information.
Thead-exposure contexts forboth television and magazine ads are typically "cluttered" with such information. Brand judgments also depend on context (Chang 2003; Yi 1990). For example, the same product attribute can be evaluated in a different light when contexts vary (Yi 1990). Evaluations ofa new brandcan also vary as a function of the presence of competing ads (Chang 2003). Since it is common for consumers to be exposed to magazine ads for competing brands with in the same context, the use ofsuch a common context in this study should improve the generalizability of its findings. In this study, however, a competing ad was always present. It is therefore difficult to know whether women would engage in involved brand evaluations if the competing ad were absent. Future research can directly test for effects of the presence of competing ads. How ad presentation order in a competitive context affects ad evaluations is another topic that deserves more research attention. In this study, participants always viewed thead for the known brand prior to the target ad. Therefore, when processing the target ad, product information for the known brandwas still accessible and could have served as a basis for comparison. If theorder was reversed and the target ad was presented first, however, would ad viewers reevaluate the target ad in line with the new information? Prior research suggests that presentation order does influence comparison strategy, with the most recently presented brandserving as the subject of comparison and the previously presented brandserving as the referent (Houston and Sherman 1995; Houston, Sherman, and Baker 1989). Thus, examining possible order effects within thetheoretical framework ofthis study is an area worthy of further research.
The findings ofthis study should be interpreted with an understanding of the following
limitations. First, this study was conducted in an East Asian country, in a region where
comparative advertisingis not widely used and is viewed less favorably by consumers
(Donthu 1998). In addition, participants were young consumers, who are more distrustful of advertisers and more cynical of advertising than older consumers (Zanot 1984). Therefore, special caution is necessary when generalizing these findings to other regions or other age groups. Third, other theories may also explain the gender differences found in this study. For example, the processing differences documented in this study may be caused by productinvolvement differences between men and women. Future research needs to be conducted to rule out this possibility. Another limitation concerns thedegree ofclutter. Thead-viewing context ofthis experiment is best described as a "low-cluttered" rather than a "high-cluttered" context. It is possible that when the number of competing ads increases, ad perceivers will adopt different processing strategies from those observed in this study. Fifth, this study only compares a new brandwith a high market share brand. Past research indicated that relative
market share of the target product and the compared product moderates the effectiveness of comparative appeals (Pechmann and Stewart 1990). Future research exploring the interaction between gender and comparative appeals can consider thepossible influence of relative market share. Moreover, only one product category was explored in this study.
Electronic dictionaries, which can be characterized as utilitarian products, may be relatively less interesting to consumers than products with hedonic value such as MP3 players or digital cameras. This may explain why some of themean responses in this study were below the midpoints of the scales. In addition, men and women may vary in their interests in this product category. Finally, thedependent measures were administered sequentially. This may have created demand or carryover effects. Also, in terms ofmeasures, future research should include more process type measures, such as cognitive responses and recall, which may better help assess the underlying psychological mechanisms.
Despite these limitations, the findings do shed light on the understanding of gender differences in strategies for processing advertising messages and are of practical value to advertisers interested in using comparative appeals to the utmost advantage. In addition, the same phenomenon documented in this study may emerge when attention-getting tactics are employed, such as rhetorical questions (Petty, Cacioppo, and Heesacker 1981) or audio beeps (Lord and Burnkrant 1993), or when other attention-getting ad appeals are used, such as mystery ads, borrowed-interest ads (Campbell 1995), or provocative appeals (Pope, Voges, and Brown 2004). Investigating these similarities is another possible direction for future research.
Sidebar
Theauthor gratefully acknowledges theEditor and three anonymous reviewers fortheir
valuable insights and constructive comments. Theresearch forthis paper was funded by a
grant from theNational Science Council in Taiwan.
References
REFERENCES
Ahluwalia, Rohini, and Robert E. Burnkrant (2004), "Answering Questions About Questions:
A Persuasion Knowledge Perspective forUnderstanding theEffects ofRhetorical Questions,"
Journal ofConsumer Research. 31 (1), 26-42.
Andersen, Rolph E., and Marvin A. Jolson (1980), "Technical Wording in Advertising:
Implications forMarket Segmentation, "Journal ofMarketing, 44 (1), 57-66.
Boush, David, Marian Friestad, and Gregory M. Rose (1994), "Adolescent Skepticism
Toward TV Advertisingand Knowledge ofAdvertiser," Journal ofConsumer Research, 21 (1),
165-175.
Campbell, Margaret C. (1995), "When Attention-Getting AdvertisingTactics Elicit Consumer
Inferences ofManipulative Intent: TheImportance ofBalancing Benefits and Investment,
"Journal ofConsumer Psychology, 4 (3), 225-254.
_____, and Amna Kirmani (2000), "Consumers' Use ofPersuasion Knowledge: TheEffects of
Accessibility and Cognitive Capacity on Perceptions ofan Influence Agent," Journal of
Consumer Research, 27 (1), 69-83.
Chang, Chingching (2001), "Processing Advertisingin a Competitive Context," paper
presented at theannual meeting of the Association forEducation in Journalism and Mass
Communication, Washington, DC, August.
_____ (2002), "Self-Congruency as a Cue in Different AdvertisingProcessing Contexts,"
Communication Research, 29 (5), 503-536.
_____ (2003), "Relative Judgments in Competitive Contexts," in Advances in Consumer
Research, vol. 31, Barbara E. Kahn and Mary Frances Luce, eds., Valdosta, GA:
Association forConsumer Research, 700-706.
Chung, Janne, and Gary S. Monroe (1998), "Gender Differences in Information Processing:
An Empirical Test of theHypothesis-Confirming Strategy in an Audit Context," Accounting
and Finance, 38 (2), 265-279.
_____, and _____ (2001), "A Research Note on theEffects ofGender and Task Complexity
on an Audit Judgment," Behavioral Research in Accounting, 13, 111-125.
Cialdini, Robert B., Richard E. Petty, and John T. Cacioppo (1981), "Attitude and Attitude
Change," Annual Review ofPsychology, 32, 357-404.
Cleveland, Mark, Barry J. Babin, Michel Laroche, Philipa Ward, and Jasmin Bergeron (2003),
"Information Search Patterns forGift Purchases: A Cross-National Examination ofGender
Differences," Journal ofConsumer Behavior, 3, 20-47.
Cotte, June, Robin A. Coulter, and Melissa Moore (2005), "Enhancing or Disrupting Guilt:
TheRole ofAd Credibility and Perceived Manipulative Intent," Journal ofBusiness Research,
58 (3), 361-368.
Darley, William K., and Robert E. Smith (1995), "Gender Differences in Information
Processing Strategies: An Empirical Test of theSelectivity Model in AdvertisingResponse,"
Journal of Advertising, 24 (1), 41-56.
Donthu, Naveen (1998), "A Cross-Country Investigation ofRecall ofand Attitude Toward
Comparative Advertising," Journal of Advertising, 17 (2), 111-122.
Droge, Cornelia, and Rene Darmon (1987), "Associative PositioningStrategies Through
Comparative Advertising: Attribute VersusOverall Similarity Approaches, "Journal of
Marketing Research, 24 (4), 377-388.
Duff, Sarah J., and Elizabeth Hampson (2001), "A Sex Difference on a Novel Spatial Working
Memory Task in Humans," Brain and Cognition, 47 (3), 470-493.
Erngrund, Karin, Timo Mantyle, and Lars-Goran Nissan (1996), "Adult Age Difference in
Source Recall: A Population-Based Study," TheJournal ofGerontology, 51 (B), 335-340.
Friestad, Marian, and Peter Wright (1994), "ThePersuasion Knowledge Model: How People
Cope with Persuasion Attempts, "Journal ofConsumer Research, 21 (1), 1-31.
_____, and _____ (1995), "Persuasion Knowledge: Lay People's and Researchers' Beliefs
About thePsychology of Advertising, "Journal ofConsumer Research, 22 (1), 62-74.
Grewal, Dhruv, Sumumar Kavanoor, Edward F. Fern, Carolyn Costley, and James Barnes
(1997), "Comparative VersusNoncomparative Advertising: A Meta-analysis,"Journal of
Marketing, 61 (4), 1-15.
Houston, David A., and Steven J. Sherman (1995), "Cancellation and Focus: TheRole of
Shared and Unique Features in theChoice Process," Journal ofExperimental Social
Psychology, 31 (4), 357-378.
_____, _____, and Sara M. Baker (1989), "Feature Matching, Unique Features, and the
Dynamics of theChoice Process: Pre-decision Conflict and Post-decision
Satisfaction,"Journal ofExperimental Social Psychology, 27 (5), 411-430.
Jain, Shailendra P., Bruce Buchanan, and Durairaj Maheswaran (2000), "Comparative
VersusNoncomparative Advertising: TheModerating Impact ofPrepurchase Attribute
Verifiability," Journal ofConsumer Psychology, 9 (4), 201-211.
Jain, Shailendra Pratap, and Steven S. Posavac (2004), "Valenced Comparison," Journal of
Marketing Research, 41 (1), 46-58.
Klinteberg, Britt A., Stem E. Levander, and Daisy Schalling (1987), "Cognitive Sex
Differences: Speed and ProblemSolving Strategies on Computerized Neuropsychological
Tasks," Perceptual and Motor Skills, 65, 683-697.
Krugman, Herbert E. (1966), "TheMeasurement of AdvertisingInvolvement," Public Opinion
Quarterly, 30 (4), 583-596.
Laczniak, Russell N., and Darrel D. Muehling (1993), "TheRelationship Between
Experimental Manipulations and Tests ofTheory in an AdvertisingMessage Involvement
Context," Journal of Advertising, 22 (3), 59-74.
Laroche, Michel, Gad Saad, Mark Cleveland, and Elizabeth Browne (2000), "Gender
Differences in Information Search Strategies fora Christmas Gift," Journal ofConsumer
Marketing, 17 (2), 500-524.
Leigh, James H., and Anil Menon (1987), "Audience Involvement Effects on theInformation
Processing ofUmbrella Print Advertising," Journal of Advertising, 16(3), 3-12.
Lord, Kenneth R., and Robert E. Burnkrant (1993), "Attention VersusDistraction: The
Interactive Effect ofProgram Involvement and Attentional Devices on Commercial
Processing, "Journal of Advertising, 22 (1), 47-60.
Madden, Thomas J., Chris T. Alien, and Jacquelyn L. Twible (1988), "Attitude Toward theAd:
An Assessment ofDiverse Measurement Indices Under Different Processing 'Sets,'" Journal
ofMarketing Research, 25 (3), 242-252.
Maldonado, Rachel, Patriya Tansuhaj, and Darrel D. Muehling (2003), "TheImpact ofGender
on Ad Processing: A Social Identity Perspective," Academy ofMarketing Science, 3, 1-15.
Meyers-Levy, Joan (1989), "Priming Effects on Product Judgments: A Hemispheric
Interpretation, "Journal ofConsumer Research, 16(1), 76-86.
_____, and Durairaj Maheswaran (1991), "Exploring Differences in Males' and Females'
Processing Strategies,"Journal ofConsumer Research, 18 (1), 63-70.
_____, and Brian Sternthal (1991), "Gender Differences in the Use ofMessage Cues and
Judgments, "Journal ofMarketing Research, 28 (1), 84-96.
Mitchell, Andrew A., and Jerry C. Olson (1981), "Are Product Attribute Beliefs theOnly
Mediator of AdvertisingEffects on BrandAttitudes?" Journal ofMarketing Research, 18 (3),
318-332.
Muehling, Darrel D., Jeffrey Stoltman, and Sanford Grossbart (1990), "TheImpact of
Comparative Advertisingon Levels ofMessage Involvement," Journal of Advertising, 19 (4),
41-50.
Pechmann, Cornelia, and S. Ratneshwar (1991), "The Use of Comparative Advertising for
Brand Positioning: Associative Versus Differentiation," Journal ofConsumer Research, 18(3),
145-160.
_____, and David W. Stewart (1990), "TheEffects of Comparative Advertisingon Attention,
Memory, and Purchase Intentions," Journal ofConsumer Research, 17(2), 180-191.
Pettit-O'Malley, Kathy L., and Mark S. Johnson (1992), "Differentiative Comparative
Advertising: Some Positive Results Revealed by Measurement ofSimultaneous Effects on
theAd-Sponsoring and Comparison Brands," Journal ofCurrent Issues and Research in
Advertising, 14 (1), 35-44.
Petty, Richard E., and John T. Cacioppo (1983), "Central and Peripheral Routes to
Persuasion: Application to Advertising," in Advertisingand Consumer Psychology, Larry
Percy and Arch Woodside, eds., Lexington, MA: Lexington Books, 3-23.
_____, _____, and Martin Heesacker (1981), "Effects ofRhetorical Questions on Persuasion:
A Cognitive Response Analysis," Journal ofPersonality and Social Psychology, 40 (3),
432^40.
Pfau, Michael (1994), "Impact ofProduct Involvement, Message Format, and Receiver Sex
on theEfficacy of Comparative AdvertisingMessages," Communication Quarterly, 42 (3), 244-
258.
Pope, Nigel K. L., Kevin E. Voges, and Mark R. Brown (2004), "TheEffect ofProvocation in
theForm ofMild Erotic on Attitude Toward theAd and Corporate Image: Differences Between
Cause-Related and Product-Based Advertising," Journal of Advertising, 33 (1), 69-82.
Prakash, Ved (1992), "Sex Roles and AdvertisingPreferences," Journal of Advertising
Research, 32 (3), 43-52.
Prasad, V. Kanti (1976), "Communication-Effectiveness of Comparative Advertising: A
Laboratory Analysis," Journal ofMarketing Research, 13(2), 128-137.
Putrevu, Sanjay, and Kenneth R. Lord (1994), "Comparativeand Noncomparative Advertising
: Attitudinal Effects Under Cognitive and Affective Involvement Conditions,"Journal of
Advertising, 23 (2), 77-91.
_____, Joni Tan, and Kenneth R. Lord (2004), "Consumer Responses to Complex
Advertisements: TheModerating Role ofNeed forCognition, Knowledge, and Gender," Journal
ofCurrent Issues and Research in Advertising, 26 (1), 9-24.
Seidlitz, Larry, and Ed Diener (1998), "Sex Differences in theRecall ofAffective Experiences,
"Journal ofPersonality and Social Psychology, 74 (1), 262-271.
Wang, Cheng Lu, Terry Bristol, John C. Mowen, and Goutam Charkraborty (2000),
"Alternative Modes ofSelf-Construal: Dimensions ofConnectedness-Separateness and
AdvertisingAppeals to theCultural and Gender-Specific Self," Journal ofConsumer
Psychology, 9(2), 107-115.
Wilson, R. Dale, and Aydin Muderrisoglu (1980), "An Analysis ofCognitive Responses to
Comparative Advertising," in Advances in Consumer Research, vol. 7, Jerry C. Olson, ed.,
Ann Arbor, MI: Association forConsumer Research, 566-571.
Winer, Benjamin J., Donald R. Brown, and Kenneth M. Michels (1991), Statistical Principles
in Experimental Design, 3rd ed., New York: McGraw-Hill.
Worchel, Stephen, and Joel Cooper (1976), Understanding Social Psychology, Homewood,
IL: Dorsey Press.
Wright, Peter L. (1974), "Analyzing Media Effects on AdvertisingResponses," Public Opinion
Quarterly, 38 (2), 192-205.
Yalch, Richard E, and Rebecca Elmore-Yalch (1984), "TheEffect ofNumbers on theRoute to
Persuasion," Journal ofConsumer Research, 11 (1), 522-527.
Yi, Youjae (1990), "TheEffects ofContextual Priming in Print Advertisements," Journal of
Consumer Research, 17 (2), 215-223.
Zanot, Eric J. (1984), "Public Attitudes Towards Advertising: TheAmerican Experience,"
International Journal of Advertising, 3d), 3-15.
Zhang, Shi, and Gavan J. Fitzsimons (1999), "Choice-Process Satisfaction: TheInfluence of
Attribute Alignability and Option Limitation," Organizational Behavior and Human Decision
Processes, 77 (3), 192-214.
_____, Frank R. Kardes, and Maria L. Cronley (2002), "Comparative Advertising: Effects of
Structural Alignability on Target BrandEvaluations," Journal ofConsumer Psychology, 12(4),
303-311.
_____, and Arthur B. Markman (1998), "Overcoming theEarly Entrant Advantage: TheRole of
Alignable and Nonalignable Differences," Journal ofMarketing Research, 35 (4), 413-426.
_____, and _____ (2001), "Processing Product Unique Features: Alignability and
Involvement in Preference Construction, "Journal ofConsumer Psychology, 11 (1), 13-27.
Zhang, Yong (1996), "Responses to Humorous Advertising: TheModerating Effect ofNeed for
Cognition," Journal of Advertising, 25 (1), 15-32.
AuthorAffiliation
Chingching Chang (Ph.D., University ofWisconsin-Madison) is a professor of advertising,
National Chengchi University, Taipei, Taiwan.
0 komentar:
Poskan Komentar