- Email: [email protected]
Self-Referent Processing in Perceptions of Verbal and Visual Commercial Information
JOURNAL OF CONSUMER PSYCHOLOGY, 1(I), 83-102 Copyright 9 1992, Lawrence Erlhaum Associates, Inc.
Self-Referent Processing in Perceptions of Verbal and Visual Commercial Information Kathleen Debevec Department of Marketing School of Management University of Massachusetts, Amherst Jean B. Romeo Department of Marketing Carroll School of Management Boston College, Chestnut Hill
We examined how visual information in an ad may interact with and influence processing of verbal information and facilitate or inhibit self-referent judgments. Self-referencing is viewed as a mediator between individuals' perceptions of verbal and visual stimuli in advertising and their subsequent attitudes and intentions. A 3 • 2 experimental design was used to examine the effects of three verbal copy strategies (self copy, product benefit copy, and typical user copy) and two visual image strategies (product featured or typical users featured). To enhance realism, the study was conducted in the context of testing a new experimental magazine. The verbal focus of an ad was shown to encourage varying levels of self-referencing and differential attitudes and intentions when a product visual was featured, but not when a slice-of-life setting was featured. Self copy (copy written in the second person vernacular) accompanied by the product visual was the most effective strategy in encouraging self-referencing and favorable attitudes and intentions.
Advertisements are designed to draw a n audience's attention a n d to encourage viewers to relate to some aspect of the p r o d u c t a n d / o r the ad. It is presumed that if consumers can relate commercial information to themselves, they will be more likely to process the information, thus e n h a n c i n g the effectiveness of Requests for reprints should be sent to Kathleen Dehevec, Schoolof Management, University of Massachusetts, Amherst, MA 01003.
84
DEBEVECAND ROMEO
an ad. This relationship between self-thoughts and information processing is supported by studies in cognitive psychology which show that individuals who self-reference information are better able to learn and recall the information than individuals who do not (Bellezza, 1981, 1984; Bower & Gilligan, 1979; Keenan & Baillet, 1980; Kendzierski, 1980; Kuiper & Rogers, 1979; Lord, 1980; Rogers, Kuiper, & Kirker, 1977). These studies conceptualize selfreferencing as a cognitive process in which individuals associate self-relevant incoming information with information previously stored in memory in order to give the new information meaning. Rogers (1981) suggested that the selfreferencing process also has an evaluative or affective quality. In an advertising context, self-referencing has been shown to play a role in persuasion. Shavitt and Brock (1986) examined subjects' self-relevant thoughts listed in response to a communication designed to encourage varying levels of self-relevant responses and found self-thoughts to be predictive of purchase intention. They attempted to manipulate self-relevant responding by instructing subjects to relate an ad to their own experience (instructions typical of self-referencing research in cognitive psychology), but the instructions themselves did not enhance self-relevant responding. Instead, through an analysis of subjects' cognitive responses, they found that the instructions resulted in counterarguing. Shavitt and Brock 0986) concluded that respondents appeared to self-reference the communications spontaneously, and it was this spontaneous self-relevant responding that was predictive of purchase intention. Yalch and Sternthal (1985) were able to encourage self-referencing when individuals were given instructions to relate message information to their own experiences. These instructions were more effective in encouraging self-referencing than instructions that encouraged people to simply focus on themselves or no instructions at all. Instructions to link message information to oneself enhanced attitudes and behaviors as well as attitude-behavior consistency. This effect is supported by past theorizing that the self evokes an elaborate network of associations in memory and by evidence provided by Cacioppo and Petty (1979) which suggests that such networks can affect individuals' responses to a message. Yalch and Sternthal (1985) acknowledged that these networks may also interfere with the processing of message information when individuals are encouraged simply to focus on themselves. They emphasized the importance of individuals' linking message information to themselves to enhance persuasion. In an actual advertising situation, instructing viewers to link the ad to their own experiences is unrealistic. Thus, it is important to find ways to encourage positive self-referencing which can easily be implemented in advertising. Only one strategy has been tested and found effective thus far. Debevec and Iyer (1988) were able to encourage varying levels of self-referencing by ads featuring traditional and progressive sex-role portrayals. This strategy may be appro-
IMAGERY AND SELF-REFERENCING
85
priate for advertisers in some instances, but a more generalizable strategy involving both the verbal and visual elements of an ad would be preferable. The purpose of our study is to examine an alternative approach to encouraging self-referencing which does not require experimental instructions and is not limited to gender role portrayals. The approach utilized here involves a simple manipulation of the wording of a communication and its accompanying visual image; thus, it should be quite easy to employ and generalize to a wide variety of communications. In addition, the visual and verbal strategies investigated have been used intuitively by advertisers, but have not been examined from a theoretical standpoint in terms of predicting why they should be effective in inducing self-referencing. We also measured self-referencing directly through scaled measures which were developed based on how self-referencing has been conceptualized in prior research. The scales were designed to be generalizable to a wide variety of visual images and copy strategies. Only one previous study attempted to measure self-referencing through scaled measures (Debevec & Iyer, 1988), but the scales were inappropriate for the visual image strategies employed in this study. Other studies have tended to attribute their results to the self-referencing effect without actually measuring the construct. In examining the effectiveness of an ad, it is also necessary to consider how the verbal and visual elements may interact with one another. It is generally thought that the visual elements in an ad dominate over the verbal ones when the ad is processed (Mitchell, 1986; Paivio, 1969, 1971; Paivio & Csapo, 1973), but research has not examined how the visual element may then influence processing of the verbal element. It is likely that some visual strategies will be more effective than others in enhancing processing of the verbal information. Therefore, the effect of visual cues on verbal ones is important and was examined in this study in order to assess the relative effectiveness of inducing self-referencing in verbal and visual modes.
VERBAL COPY STRATEGIES AND SELF-REFERENCING The way in which an ad is worded should affect the degree to which the message information is self-referenced and responded to. In writing advertising copy, experts intuitively suggest that to get and keep an audience interested, one should refrain from using the pronoun I and instead write in terms of you (Bovee & Arens, 1989). This strategy is consistent, in principle, with the way in which social psychologists have induced self-referencing in past research. Individuals were asked whether trait words were descriptive of themselves (Bower & Gilligan, 1979; Rogers et al., 1977; Wells, Hoffman, & Enzle, 1984), and those engaging in this self-reference task exhibited greater recall for the
86
DEBEVECAND ROMEO
words than those asked just the relative meaning of each word (Rogers et al., 1977) or those instructed to relate the word to another person (Bower & Gilligan, 1979). Self-referencing is induced because it prompts individuals to focus attention on themselves while providing a link with the stimulus. Thus, it should be possible to stimulate self-referencing in an advertising context by addressing individuals in the second person (you) or by name (in a direct marketing or personal selling context) while referring to a product, a need, or a lifestyle characteristic. Individuals should be more likely to engage in self-referencing when exposed to an ad written in the second person than to those written in the first (e.g., a testimonial) or third (most people) person. Writing advertising copy in the second person vernacular is a very simple linguistic manipulation and, therefore, has the potential of being a very powerful persuasive strategy. This strategy is thus deserving of extensive study. Three verbal information strategies have been selected for inclusion in the study to induce varying levels of self-referencing. These verbal strategies include copy focusing either on: (a) product benefits, (b) activities or characteristics of the typical user, or (c) the audience directly (further described as self-referent copy). These three copy forms vary in their focus, whether it be the product, typical users, or the self. Slice-of-life/typical user copy is written in the third person, such as "'people on the g o . . . " Self-referent copy is copy written in the second person and directed toward individuals in the target audience (the pronoun you is used liberally throughout). This strategy is employed to encourage an audience to become personally involved with the message and product and to make them feel that the communication is meant for them as individuals. Although you copy is recommended by advertising experts in order to get an audience's attention (Bovee & Arens, 1989), its effectiveness has not been reported in the literature, and the theoretical reason for its proposed effectiveness has not been investigated. These two copy forms are compared to straight product-oriented copy.
VISUAL IMAGERY AND SELF-REFERENCING Self-referencing may also be stimulated through visual imagery which has been shown to be a very powerful tool in enhancing memory and persuasion (Paivio, 1969, 1971; Paivio & Csapo, 1973). In print ads, advertisers often feature characters who are purportedly like the target audience in a slice-of-life portrayal. The goal of this strategy is for viewers to relate to others who have similar lifestyles or interests and to make the inference that the product will fit into their own lifestyles. However, audience members, in making a similarity judgment, may or may not relate to the characters or situation featured. It is possible that self-referent judgments may be negative and lead
IMAGERY AND SELF-REFERENCING
87
to unfavorable attitudes and intentions if audience members have difficulty relating to the characters or situations. A negative visual imagery effect was noted by Kisielius and Sternthal (1984) who discovered that a vivid picture can undermine persuasion when unfavorable information is recalled from memory when processing occurs. In their study, an ad with a visual image led to a less favorable attitude toward the product than verbal statements that provided product information only. They proposed the "availability valence hypothesis," suggesting that the relative favorableness of a judgment is dependent on whether information is recalled from memory when processing a stimulus ad and the valence of that information. In this study, we examined the effect of visual cues on self-referencing. Two visual conditions were tested and evaluated relative to one another and to verbal strategies designed to encourage varying levels of self-referencing. The two visual treatments included a slice-of-life scene and a product visual. Numerous studies have shown that pictorial stimuli and/or ads are more memorable than verbal ones (Childers & Houston, 1984; Mitchell, 1986; Paivio, 1969; Rossiter & Percy, 1980; Starch, 1966), suggesting a phenomenon termed thepicture superiority effect. Because of the superiority of pictures over words, it is necessary to consider how verbal information may interact with the ad's visual imagery. Other studies have examined the independent effects of visual and verbal cues in ads (Mitchell, 1986) or how verbal cues influence the cognitive processing of ads (Edell & Staelin, 1983). In this study, we examined the extent to which verbal information in ads is processed in response to visual cues. In summary, the purpose of this study was to examine the extent to which verbal and visual strategies encourage favorable self-referencing and, thus, influence attitudes and intentions. In addition, the effect of visual cues on verbal information processing and the interaction of visual and verbal cues was investigated.
SPECIFIC HYPOTHESES Self-Referencing Effect Up to this point, it has been suggested that the verbal and visual information strategies will encourage varying levels of self-referencing. Prior to examining the effect of these strategies, it is first necessary to establish that self-referencing influences evaluations of the ad and product as well as behavioral intentions. The extent to which an ad and/or product are self-referenced needs to be evaluated relative to individuals' attitudes and intentions. In addition, selfreferencing must be examined as a mediator between individuals' perceptions
88
DEBEVEC A N D R O M E O
of each ad and their subsequent attitudes and intentions. It is expected that the more an individual self-references information, the more effective that information will be. Consequently, the first hypothesis is that individuals who experience high levels of self-referencing will have more positive attitudes and intentions than those who experience lower levels, irrespective of the ad viewed. In addition, self-referencing is a mediator between perceptions of the verbal information and visual focus and attitudes and intentions.
Visual Influence on Verbal Information Processing Although there is considerable evidence that information presented visually is more memorable than information presented verbally (Lutz & Lutz, 1977; Mitchell & Olson, 1981; Paivio, 1969, 1971; Paivio & Csapo, 1973), and thus can dominate individuals' attention, it is likely that some visual images will be more conducive to processing verbal information than others. We expected that the visual focus of the ad will influence processing of the verbal information in the following way. When the product is the focal point of an ad's illustration, the audience's information-processing efforts are directed toward processing information about the product only, thus resulting in efficient and continued processing of the verbal information. Viewers may make inferences about the product's usefulness to themselves when encouraged to do so by the verbal copy. Ads that feature typical users of the product (individuals to whom a target audience can relate) may require a deeper level of processing than those that feature only the product. With a typical user visual, the audience may not only make an inference about the product's benefit to themselves, but they may also infer the degree to which they can relate to the characters and lifestyles featured in the ad. If they react negatively to some aspect of the typical user visual, they will likely engage in counterarguments and pay little attention to the accompanying verbal information. Thus, verbal copy should be most effective in encouraging viewers to self-reference when its processing is not inhibited by the visual imagery. In summary, a second hypothesis is that when the visual emphasis in the ad is on the product, individuals' level of self-referencing will vary with the verbal content of an ad, but when the visual emphasis in the ad is on typical users of the product, individuals" level of self-referencing will not vary with the verbal content of an ad. Given that verbal information processing differences are expected when the product visual is featured, it is important to consider how and why the verbal information will be processed differentially and the resultant effectiveness of each verbal strategy. The copy strategies were selected for inclusion in the study because they are believed to encourage varying levels of self-referencing. The more an audience is able to self-reference verbal information, the more effective the verbal copy is expected to be. Of the three copy strategies, the you copy is expected to induce the highest level of self-referencing and, there-
IMAGERY AND SELF-REFERENCING
89
fore, be the most effective. Product benefit copy and typical user copy are expected to result in comparable levels of self-referencing. Therefore, a third hypothesis is that when the visual emphasis in the ad is on the product, self copy will result in more self-referencing than product benefit copy and typical user copy. There should be no significant difference in individuals' self-referencing in response to the latter two copy forms. Self copy should encourage a self-perception process on the part of individuals. Information processed in relation to the self should be more effective than information processed relative to others (Rogers et al., -1977). Research suggests that the "self" is a much more organized structure than the "other"; therefore, less effort is required to process self-relevant information than otherrelevant information (Bower & Gilligan, 1979; Keenan & Baillet, 1980; Kuiper & Rogers, 1979; Rogers, 1981). The self is thought to be one of the richest schemata in memory, thereby prompting deep encoding of information (Kennan & Baillet, 1980). Copy strategies that encourage self-referencing should be superior to copy strategies that encourage person or object perception and lead to more positive attitudes and intentions. Therefore, the last hypothesis suggests that when the visual emphasis in the ad is on the product, self copy will result in more positive attitudes and intentions than product benefit copy and typical user copy. However, there should be no significant difference in individuals' attitudes and intentions in response to the latter two copy forms.
METHOD Sample and Stimuli Participants in the study were 158 junior and senior college students (100 women and 58 men) enrolled in undergraduate marketing courses. The product selected for this study was a soft drink given the hypothetical name, Twist. Undergraduate students purchase and consume this product class, and soft drinks are promoted by both men and women. Thus, the product is one to which study participants should be able to relate, regardless of their gender. The hypotheses were tested in a 3 • 2 factorial design in which there were three verbal copy strategies (self copy, product benefit copy, and typical user copy) and two visual image strategies (product featured and typical users featured). Print was selected as the experimental medium, and ads of approximately equal length were developed. The content of the ads was similar (i.e., all conveyed the same product information) with the exception of pronoun variations (you vs. people) or omission of pronouns (product benefit copy). For example, the self-referencing copy included the headline " W H E N YOU W O R K H A R D A N D PLAY H A R D YOU N E E D TWIST," whereas the body copy read: "TWIST will fit your active lifestyle. It will quench your
90
DEBEVEC A N D R O M E O
thirst with the great taste of real fruit juice. You'll appreciate the added vitamins, too. TWIST will give you the energy you need to keep going." The typical user copy replaced you with people (the headline read: " P E O P L E W H O W O R K H A R D A N D PLAY H A R D D R I N K TWIST"). The product benefit copy simply read, " I N T R O D U C I N G TWIST." The visual featured either a can of Twist beside an enticing glass of the beverage (product visual) or a split screen, with one scene showing students studying and another scene showing a group playing tennis (typical users visual). One of the tennis players held a can of Twist.
Procedure The study was conducted over a 2-week period within which students signed up for 1 of 12 sessions. All treatments were tested in each of the sessions. Subjects were instructed that they would be evaluating a sample of articles and ads for a new magazine designed to appeal to college students who plan to begin a career upon graduation. Each participant was then given a packet that contained an article followed by the experimental ad and a second article also followed by an ad. A series of questions also followed each article and ad, but only those questions following the experimental ad are discussed further. Subjects evaluated each of the two articles and ads so that their attention would not be focused solely on the experimental ad. Participants were instructed to read the first article and respond to the questions. They were then given 2 min to view the experimental ad, after which they responded to a series of questions using 7-point scales that assessed their level of self-referencing, their attitude toward the ad and product, and their behavioral intentions. The same procedure was followed for the second article and ad.
Self-Referencing Measures The measures selected to assess self-referencing reflect how it has been conceptualized and induced in prior research. Shavitt and Brock (1986) suggested that self-referencing is related to Krugman's (1965) notion of involvement in terms of "conscious bridging experiences" (Krugman, 1965, p. 354; Shavitt & Brock, 1986, p. 152). Self-referencing has been described by others as a cognitive process whereby individuals associate self-relevant stimulus information with information previously stored in memory in order to give the new information meaning (Bellezza, 1984; Kuiper & Rogers, 1979; Markus, 1977; Rogers, 1981). In past studies, subjects have been instructed to relate stimulus information to themselves (Bellezza, 1984; Lord, 1980), to think of personal experiences that relate to a stimulus (Bower & Gilligan, 1979), or to picture themselves relative to stimulus information (Shavitt & Brock, 1986; Yalch & Sternthal, 1985). The extent to which individuals self-reference incoming in-
IMAGERY AND SELF-REFERENCING
91
formation should be reflected by the personal relevance of the information, the product's usefulness, and whether the individuals could picture themselves with the product. In this study, the degree to which individuals self-referenced an ad or product was assessed in three ways. The degree to which the ad was selfreferenced was measured by asking subjects how personally relevant the ad washsubsequently referred to as self-referencing of ad. The degree to which the product was self-referenced was measured by asking subjects whether the product was personally useful and personally beneficial. The product selfreferencing measures were averaged to form a composite measure--subsequently referred to as self-referencing of product (Cronbach's alpha = .77). The third self-referencing measure included two questions which assessed whether individuals could picture themselves trying the product and serving the product to a friend as they viewed the ad. Once again, the questions were averaged to form a composite measure, and this measure is subsequently referred to as try/serve (Cronbach's alpha = .90).
Attitude and Intention Measures Composite measures of subjects' attitudes toward the ad and product and their behavioral intentions were developed as a result of principal component analyses with varimax rotation. Subjects' attitude toward the ad was measured through 12 bipolar adjective scales, which were reduced to an Affective and a Cognitive factor. The first factor, which represents atfective impressions of the ad (personal/impersonal, beneficial/not beneficial, impressive/unimpressive, appealing/unappealing, interesting/boring, entertaining/unentertaining, good/bad), accounts for 53.6% of the variance (eigenvalue = 6.97). The second factor, which represents cognitive impressions of the ad (believable/ unbelievable, realistic/unrealistic, informative/uninformative, clear/unclear, logical/illogical), accounts for 9.2% of the variance (eigenvalue = 1.2). The Cronbach's alphas for Affective and Cognitive factors were .93 and .78, respectively. Subjects' attitude toward the product was measured on two sets of scales. The first set of eight scales assessed subjects' ratings of the product. The scales were reduced to a Product Rating factor (unique/ordinary, good/bad, superior/inferior, appealing/unappealing, entertaining/unentertaining, useful/ useless) which accounts for 60.9% of the variance (eigenvalue = 5.48). The second set of scales assessed how subjects thought the product would make them feel. Feelings were measured on six semantic differential scales which were reduced to a Product Feeling factor (special, happy, important, concerned) which accounts for 49.3% of the variance (eigenvalue = 2.96). The Cronbach's alphas for the Product Rating and the Product Feeling factors were .81 a n d . 87, respectively. Finally, intentions were measured through three scales that included the
92
DEBEVECAND ROMEO
likelihood that subjects would try the product, purchase the product at a retail store, and purchase the product in a bar or restaurant if the product were available locally. These measures were averaged to form a composite measure of intention (Cronbach's alpha of .85).
RESULTS The first hypothesis examines the effect of self-referencing on subjects' attitudes and intentions and the mediating property of self-referencing. The more an individual self-referenced the ad a n d / o r product, the more positive their attitudes and intentions were expected to be. The hypothesis received strong support for each of the three self-referencing measures (all ps < .01, see Table 1). Each self-referencing measure was also treated as a covariate relative to the verbal and visual predictor variables, whereas respondents' attitudes and behavioral intentions served as the criterion measure. Self-referencing of the ad or product mediated the effect of the verbal and visual treatments on all dependent variables (allps < .01). Thus, self-referencing plays an integral role in the processing of information in ads. The second hypothesis suggests that the visual image featured in an ad will influence reactions to the verbal information presented. The extent to which viewers self-reference the ad and/or product is expected to vary with the verbal content in an ad when the visual emphasis is on the product, but not when the visual emphasis is on typical users of the product. This hypothesis was investigated by performing three 2 • 3 (Visual Image • Copy Strategy) analyses of variance (ANOVAs) using the self-referencing measures as the dependent measures. The results indicated a significant Verbal X Image interaction for the product self-referencing measure, F(2, 127) = 5.20, p < .01 and marginally significant interactions for the ad, F(2, 127) -----2.39, p ---- .06; and t r y / serve, F(2, 127) = 2.39, p < . 10, self-referencing measures. An investigation of the interactions depicted in Figure 1 indicates that the effectiveness of the verbal copy in inducing self-referencing varied more across the product visual than across the typical user visual. As expected, when the visual emphasis is on the product, self-copy encourages more self-referencing of the product (M = 4.8), ad (M = 4.8), and intentions (M = 4.8), than when the visual emphasis is on the typical user (M = 3.4, M = 4.1, and M = 3.1, respectively). Thus, when accompanied by the typical user visual, self copy ranges from being moderately effective in terms of encouraging self-referencing of the ad to quite ineffective in prompting self-referencing of the product and intentions. These results are discussed further in the investigation of the third hypothesis. Hypothesis 3 addresses the level of self-referencing encouraged by each
IMAGERY AND SELF-REFERENCING
93
TABLE 1 Mean Attitude and Intention Scores by High Versus Low Self-Referencing
Self-Referencing Measure (Dependent Variable) Self-referencing of ad Attitude toward ad Atfective Cognitive Attitude toward product Affective Cognitive Intention Self-referencing of product Attitude toward ad Affective Cognitive Attitude toward product Affective Cognitive Intention Try/serve Attitude toward ad Affective Cognitive Attitude toward product Affective Cognitive Intention
Level of Self-Referencing Low
High
Significance
2.7 4.1
4.6 4.9
< .01 < .0 I
1.8 3.8 2. I
3.0 4.7 4.4
< .01 < .0 i < .01
2.8 4.0
4.7 5.1
< .01 < .01
1.9 3.8 2.2
3.0 4.7 4.5
<.01 <.01 < .01
2.8 4.0
4.9 5.2
< .01 < .01
2.0 3.6 2.2
3.1 4.9 4.8
< .0l < .01 < .01
Note. Individuals were classified as low in self-referencing if they had a self-referencing mean score less than 4.0 and high in self-referencing if their mean score was greater than or equal to 4.0. v e r b a l i n f o r m a t i o n s t r a t e g y g i v e n t h e p r e s e n c e o f the p r o d u c t visual. Self c o p y is e x p e c t e d to p r o d u c e m o r e s e l f - r e f e r e n c i n g t h a n p r o d u c t benefit c o p y a n d t y p i c a l u s e r copy, b u t t h e l a t t e r t w o c o p y strategies s h o u l d p r o d u c e s i m i l a r r e s p o n s e s f r o m viewers. T e s t s o f significant effects w e r e c o n d u c t e d to assess this h y p o t h e s i s . G r o u p m e a n s a n d significant differences a r e r e p o r t e d in T a b l e 2. T h e results i n d i c a t e d t h a t self c o p y d i d e n c o u r a g e m o r e self-referencing o f t h e p r o d u c t , t h e ad, a n d w h e t h e r o r n o t subjects c o u l d see t h e m s e l v e s t r y i n g t h e p r o d u c t t h a n t h e t y p i c a l u s e r copy. T h e self c o p y was significantly m o r e effective t h a n t h e p r o d u c t benefit c o p y for s e l f - r e f e r e n c i n g o f t h e a d a n d o n l y m a r g i n a l l y significant for s e l f - r e f e r e n c i n g t h e p r o d u c t (p < . 10). A s e x p e c t e d , t h e r e w e r e n o significant differences b e t w e e n t h e typical u s e r c o p y a n d t h e
i
C~
i r
94
l
l
l
,
r
v
f
r
v
r
,
v
,
r
l
T
i
v
V
E
f I
0
I ~ l W r l l l , ~ , , J l l , , r ' v r ~
4
~
4
4
4
4
4
4
4
~
.
.
.
.
.
.
.
.
~
TILl rr
U_
95
96
DEBEVEC AND ROMEO TABLE 2 Verbal Information Effectiveness When Product Visual is Featured
Type of Copy Variable
Self-referencing Ad Product Try/serve Attitude toward ad Affeetive Cognitive Attitude toward product Affeetive Cognitive Intention
Self
Product Benefit
Typical User
4.8 (a,b) 4.8 (a) 4.8 (a)
3.3 (a) 3.9 4.2
3.2 (b) 3.6 (a) 3.6 (a)
4.6 (a,b) 5.2 (a)
3.9 (a) 4.9
3.5 (b) 4.4 (a)
3.4 (a,b) 4.8 (a) 4.8(a,b)
2.3 (a) 4.3 3.5 (a)
2.0 (b) 4.1 (a) 3.1 (b)
Note. (a,b) refer to groups that are significantly different from one another at the .05 level of significance. For example, for self-referencingthe ad, 3.3 and 4.8 are significantlydifferent (both designated with "a") as are 4.8 and 3.2 (both designated with "b").
product benefit copy. In addition, when the typical user visual was featured, the self-referencing measures did not vary across the copy strategies. This was evidenced by tests of simple effects (all p s > . 12). The final hypothesis proposes that the self-referencing effects expected would transfer to more positive attitudes and intentions to buy the product. This hypothesis was investigated by performing five 2 • 3 (Visual Image X C o p y Strategy) A N O V A s using the attitudes toward the product and ad and intention to buy as the dependent measures. The results indicated significant Verbal • Image interactions for cognitive attitudes toward the product, F(2, 127) = 4.18, p = .02; affective attitudes toward the product, F(2, 127) = 4.348, p = .02; and intention to buy, F(2, 127) = 3.27, p = .04; and a marginally significant interaction for affective attitude toward the ad, F(2, 127) = 2.5,p = .09. The interaction effect for cognitive attitude toward the product was not significant (p = . 19). A n investigation of the interactions (see Figure 2) indicated that attitudes and intentions varied more across the verbal copy strategies when the product visual was featured than when the typical user visual was featured. Tests of simple effects indicated that when the product visual was featured, attitudes toward the product and ad (both affective and cognitive components) and intentions were significantly higher for the self copy than for the typical user copy. Self copy was significantly more effective than product benefit copy for affective attitudes toward the ad and product and intention to buy (see Table 2). As expected, there were no differences in attitude and intention between the product benefit and typical user copy.
IMAGERY AND SELF-REFERENCING
9")'
Again, when the typical user visual was featured, there were no significant differences for the attitudes and intention measures across the copy strategies (all ps > .24).
DISCUSSION The purpose of this study was to examine the role of self-referent processing in evaluations of ads varying in their visual and verbal focus and the interaction of verbal and visual information. The influence of self-referencing was first assessed, and then the strategies were evaluated in terms of the level of selfreferent thinking encouraged by each strategy and viewers' resultant attitudes and intentions. It was hypothesized that self-referencing would mediate attitudes toward the ad and product and behavioral intentions. As expected, the more an individual self-referenced the product and/or ad, the more favorable were their attitudes and intentions. This suggests that in a persuasion context, self-referencing is a desirable outcome. Thus, it is important for advertisers to understand which visual and verbal strategies encourage self-referent judgments. The level of self-referent processing and subjects' attitudes and intentions were shown to vary with the verbal and visual strategies. The results of this study are significant because they suggest that a simple linguistic manipulation in which ad copy directs viewers' attention to themselves can influence selfreferencing and attitudes and intentions. An equally important and related finding was that the visual focus of the ad influenced processing of the verbal information and, particularly, the self copy. When the product visual was featured, self copy was significantly superior to typical user copy and generally superior to product benefit copy in inducing viewers to self-reference the communication and in gaining more favorable ad and product evaluations and behavioral intentions. In fact, the self copy paired with the product visual was the most effective verbal and visual strategy examined. The effectiveness of the self copy decreased dramatically, however, when it was paired with the typical user visual (as seen in Figures 1 and 2). One explanation for this result may be that subjects did not view themselves as similar to the typical users featured in the ad and made inferences about the users that were inconsistent with their own self-image. When self copy was paired with the product visual, viewers may have made inferences about themselves and the product that were consistent and unlikely to generate negative evaluations. Yalch and Sternthal's (1985) findings, which were reported earlier, may offer additional insight. The product visual accompanied by self copy may have encouraged subjects to link product information to themselves, thereby enhancing the persuasive power of this copy strategy. The typical user visual
o
0
0
0
0
0
0
0
0
0
O
0
0
0
0
0
0
2
J
l I 98
~
l
,
'
l
W
l
W
l
l
l
l
r
l
~
l
l
LU
i
v
~
w
99
100
DEBEVECAND ROMEO
accompanied by self copy may have led subjects to simply focus on themselves relative to the typical users portrayed. Thus, in the former instance, a network of associations to the self may have been formed which should enhance persuasion, whereas in the latter instance, this network of associations may be absent, thus inhibiting persuasion. Yalch and Sternthal concluded that encouraging individuals to link message information to themselves is critical in enhancing persuasion, as opposed to encouraging them to only focus on themselves. Another possible explanation for the findings in this study is that the typical user visual inhibited processing of the verbal information in the ads, whereas the product visual encouraged processing. Support for this explanation lies in the fact that none of the verbal copy strategies were superior to one another when the visual focus was of typical users; instead, all copy strategies were moderately effective. However, there were significant differences in the effectiveness of the copy strategies when the product visual was featured. An examination of the effectiveness of the three copy strategies when paired with each of the two visual strategies raises another potential explanation for some of the findings. Research has shown that learning is facilitated when a visual image and verbal information have figural unity (are congruent) than when they do not (Lippman & Shanahan, 1973). In our study, when the orientation of the copy is congruent with the visual focus of the ad, selfreferencing is sometimes enhanced. This occurred for the self copy and typical user copy but not for the product benefit copy. Self-referencing of the product was significantly higher when self copy was featured with the product visual than when it was featured with the endorser/user visual (M ----4.8 vs. 3.5) and when typical user copy was paired with the endorser/user visual than when paired with the product visual (M = 4.6 vs. 3.6). A significant increase was also noted in the intention self-referencing measure when the self copy was paired with the product visual (M = 4.8 vs. 3.2). However, product benefit copy was no more effective, and in some cases less effective, in enhancing self-referencing when featured with the product visual than when featured with the endorser/user visual--a result contrary to the congruence explanation. Similar results were found when examining the attitude and intention measures (as shown in Figure 2). In general, the effectiveness of the typical user copy and self copy tended to vary considerably with the visual focus of the ad but the product benefit copy did not. Self copy was more likely to be effective when paired with the product visual than when paired with the endorser/user visual, whereas typical user copy was more likely to be effective when paired with the endorser/user visual than with the product visual. Again, product benefit copy was no more effective when paired with the product visual than when paired with the endorser/user visual. In addition, if only congruence were operating, two strategies might be expected to be superior to all others--the product benefit copy accompanied
IMAGERY AND SELF-REFERENCING
1 01
by the product visual and the typical user copy accompanied by the typical user visual. These strategies were both less effective than the self copy paired with the product visual There are also no significant differences in the effectiveness of the copy strategies when the typical user is featured. Thus, the congruence explanation is consistent with some but not all of the results. Further research is warranted to investigate the feasibility of each of these potential explanations in order to yield a more complete conceptualization of self-referencing than is currently available. It is important to understand under what conditions self copy will be effective and when and why the visual image is likely to facilitate or inhibit self-referencing and the effectiveness of the copy strategies. An interesting finding was the absence of main effects for the verbal and visual strategies and the strong interactions noted across all dependent measures. Clearly, individuals are reacting to the visual focus of the ad. What is not so clear is what is going through their minds as they are reacting to the visual element. Are they noting the congruity or incongruity between the verbal and visual strategies? Are they linking message information to themselves, and, if so, is their reaction favorable or unfavorable? Are they making inferences about the product or typical users relative to their perception of themselves? Additional experimentation that partitions out these effects should be revealing, as would the collection of cognitive response data in which subjects' spontaneous thoughts about the product, the ad, and themselves could be assessed. Although this study raises a series of unanswered questions, it also provides a foundation for future research on strategies that may encourage self-referencing and may reveal how the verbal and visual elements in an ad interact to enhance or inhibit the self-referencing effect.
ACKNOWLEDGMENTS Research for this article was supported in part by a University of Massachusetts School of Management summer research grant. We gratefully acknowledge the assistance of Thomas Srull and three anonymous reviewers for comments on an earlier draft of this article.
REFERENCES Bellezza, F. S. (1981). Mnemonic devices: Classification, characteristics and criteria. Review of Educational Research, 51, 247-275. Bellezza, F. S. (1984). The self as a mnemonic device: The role of internal cues. Journal of Personality and Social Psychology, 47, 506-516. Bovee, C., & Arens, W. F. (1989). Contemporary advertising. Homewood, IL: Irwin.
102
DEBEVEC AND ROMEO
Bower, G. H., & Gilligan, S. G. (1979). Remembering information related to one's self. Journal of Research in Personality, 13, 420-432. Caeioppo, J., & Petty, R. (1979). Effects of message repetition and position on cognitive response, recall, and persuasion. Journal of Personality and Social Psychology, 37, 97-109. Childers, T., & Houston, M. J. (1984). Conditions for a picture-superiority effects in consumer memory. Journal of Consumer Research, 11, 643-654. Debevec, K., & Iyer, E. (1988). Self-referencing as a mediator of the effectiveness of sex-role portrayals in advertising. Psychology & Marketing, 5, 71-84. EdeU, J., & Staelin, R. (1983). The information processing of pictures in print advertisement. Journal of Consumer Research, 10, 45-61. Keenan, J.M., & Baillet, S.D. (1980). Memory for personally and socially significant events. In R.S. Nickerson (Ed.), Attention and performance (Vol. 8, pp. 651-659). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Kendzierski, D. (1980). Self schemata and scripts: The recall of self referent and scriptal information. Personality and Social Psychology Bulletin, 6, 23-29. Kisielius, J., & Sternthal, B. (1984). Detecting and explaining vividness effects in attitudinal judgments. Journal of Marketing Research, 21, 54-64. Krngman, H. (1965). The impact of television advertising: Learning without involvement. Public Opinion Quarterly, 29, 349-356. Kuiper, N. A., & Rogers, T. B. (1979). Encoding of personal information: Self--other differences. Journal of Personality and Social Psychology, 37, 499-512. Lippman, M. Z., & Shanahan, M. W. (1973). Pictorial facilitation of paired-associate learning: Implications for vocabulary training. Journal of Educational Psychology, 64, 216-222. Lord, C. G. (1980). Schemas and images as memory aids: Two modes of processing social information. Journal of Personality and Social Psychology, 35, 63-78. Lutz, K. A., & Lutz, R. J. (1977). Effects of interactive imagery on learning: Application to advertising. Journal of Applied Psychology, 62, 493-498. Markus, H. (1977). Self-schemata and processing information about the self. Journal of Personality and Social Psychology, 35, 63-78. Mitchell, A. (1986). The effect of verbal and visual components of advertisements on brand attitudes and attitude toward the advertisement. Journal of Consumer Research, 13, 12-24. Mitchell, A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitudes7 Journal of Marketing Research, 18, 318-332. Paivio, A. (1969). Mental imagery and associative learning and memory. Psychological Review, 76, 241-263. Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart & Winston. Pavio, A., & Csapo, K. (1973). Picture superiority in free recall: Imagery or dual coding. Cognitive Psychology, 5, 176-206. Rogers, T. B. (1981). A model of the self as an aspect of the human information processing system. In N. Cantor & J. F. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp. 193-214). HiUsdale, NJ: Lawrence Erlbaum Associates, Inc. Rogers, T. B., Kuiper, N. A., & Kirker, W. S. (1977). Self-referencing and the encoding of personal information. Journal of Personality and Social Psychology, 35, 677-688. Rossiter, J., & Percy, L. (1980). Attitude change through visual imagery in advertising. Journal of Advertising, 9, 10-16. Shavitt, S., & Brock, T. C. (1986). Self-relevant responses in commercial persuasion. In J. Olson & K. Sentis (Eds.), Advertising and consumer psychology (pp. 149-171). New York: Praeger. Starch, D. (1966). How does the shape of ads affect readership? Media~Scope, 10, 83-85. Wells, G., Hotfman, C., & Enzle, M. (1984). Self versus other-referent processing at encoding and retrieval. Personality and Social Psychology Bulletin, 10, 574--584. Yalch, R., & Sternthal, B. (1985). The effect of self-referencing strategies on messagepersuasiveness. Unpublished manuscript, University of Washington, School of Business Administration, Seattle.
Self-Referent Processing in Perceptions of Verbal and Visual Commercial Information Kathleen Debevec Department of Marketing School of Management University of Massachusetts, Amherst Jean B. Romeo Department of Marketing Carroll School of Management Boston College, Chestnut Hill
We examined how visual information in an ad may interact with and influence processing of verbal information and facilitate or inhibit self-referent judgments. Self-referencing is viewed as a mediator between individuals' perceptions of verbal and visual stimuli in advertising and their subsequent attitudes and intentions. A 3 • 2 experimental design was used to examine the effects of three verbal copy strategies (self copy, product benefit copy, and typical user copy) and two visual image strategies (product featured or typical users featured). To enhance realism, the study was conducted in the context of testing a new experimental magazine. The verbal focus of an ad was shown to encourage varying levels of self-referencing and differential attitudes and intentions when a product visual was featured, but not when a slice-of-life setting was featured. Self copy (copy written in the second person vernacular) accompanied by the product visual was the most effective strategy in encouraging self-referencing and favorable attitudes and intentions.
Advertisements are designed to draw a n audience's attention a n d to encourage viewers to relate to some aspect of the p r o d u c t a n d / o r the ad. It is presumed that if consumers can relate commercial information to themselves, they will be more likely to process the information, thus e n h a n c i n g the effectiveness of Requests for reprints should be sent to Kathleen Dehevec, Schoolof Management, University of Massachusetts, Amherst, MA 01003.
84
DEBEVECAND ROMEO
an ad. This relationship between self-thoughts and information processing is supported by studies in cognitive psychology which show that individuals who self-reference information are better able to learn and recall the information than individuals who do not (Bellezza, 1981, 1984; Bower & Gilligan, 1979; Keenan & Baillet, 1980; Kendzierski, 1980; Kuiper & Rogers, 1979; Lord, 1980; Rogers, Kuiper, & Kirker, 1977). These studies conceptualize selfreferencing as a cognitive process in which individuals associate self-relevant incoming information with information previously stored in memory in order to give the new information meaning. Rogers (1981) suggested that the selfreferencing process also has an evaluative or affective quality. In an advertising context, self-referencing has been shown to play a role in persuasion. Shavitt and Brock (1986) examined subjects' self-relevant thoughts listed in response to a communication designed to encourage varying levels of self-relevant responses and found self-thoughts to be predictive of purchase intention. They attempted to manipulate self-relevant responding by instructing subjects to relate an ad to their own experience (instructions typical of self-referencing research in cognitive psychology), but the instructions themselves did not enhance self-relevant responding. Instead, through an analysis of subjects' cognitive responses, they found that the instructions resulted in counterarguing. Shavitt and Brock 0986) concluded that respondents appeared to self-reference the communications spontaneously, and it was this spontaneous self-relevant responding that was predictive of purchase intention. Yalch and Sternthal (1985) were able to encourage self-referencing when individuals were given instructions to relate message information to their own experiences. These instructions were more effective in encouraging self-referencing than instructions that encouraged people to simply focus on themselves or no instructions at all. Instructions to link message information to oneself enhanced attitudes and behaviors as well as attitude-behavior consistency. This effect is supported by past theorizing that the self evokes an elaborate network of associations in memory and by evidence provided by Cacioppo and Petty (1979) which suggests that such networks can affect individuals' responses to a message. Yalch and Sternthal (1985) acknowledged that these networks may also interfere with the processing of message information when individuals are encouraged simply to focus on themselves. They emphasized the importance of individuals' linking message information to themselves to enhance persuasion. In an actual advertising situation, instructing viewers to link the ad to their own experiences is unrealistic. Thus, it is important to find ways to encourage positive self-referencing which can easily be implemented in advertising. Only one strategy has been tested and found effective thus far. Debevec and Iyer (1988) were able to encourage varying levels of self-referencing by ads featuring traditional and progressive sex-role portrayals. This strategy may be appro-
IMAGERY AND SELF-REFERENCING
85
priate for advertisers in some instances, but a more generalizable strategy involving both the verbal and visual elements of an ad would be preferable. The purpose of our study is to examine an alternative approach to encouraging self-referencing which does not require experimental instructions and is not limited to gender role portrayals. The approach utilized here involves a simple manipulation of the wording of a communication and its accompanying visual image; thus, it should be quite easy to employ and generalize to a wide variety of communications. In addition, the visual and verbal strategies investigated have been used intuitively by advertisers, but have not been examined from a theoretical standpoint in terms of predicting why they should be effective in inducing self-referencing. We also measured self-referencing directly through scaled measures which were developed based on how self-referencing has been conceptualized in prior research. The scales were designed to be generalizable to a wide variety of visual images and copy strategies. Only one previous study attempted to measure self-referencing through scaled measures (Debevec & Iyer, 1988), but the scales were inappropriate for the visual image strategies employed in this study. Other studies have tended to attribute their results to the self-referencing effect without actually measuring the construct. In examining the effectiveness of an ad, it is also necessary to consider how the verbal and visual elements may interact with one another. It is generally thought that the visual elements in an ad dominate over the verbal ones when the ad is processed (Mitchell, 1986; Paivio, 1969, 1971; Paivio & Csapo, 1973), but research has not examined how the visual element may then influence processing of the verbal element. It is likely that some visual strategies will be more effective than others in enhancing processing of the verbal information. Therefore, the effect of visual cues on verbal ones is important and was examined in this study in order to assess the relative effectiveness of inducing self-referencing in verbal and visual modes.
VERBAL COPY STRATEGIES AND SELF-REFERENCING The way in which an ad is worded should affect the degree to which the message information is self-referenced and responded to. In writing advertising copy, experts intuitively suggest that to get and keep an audience interested, one should refrain from using the pronoun I and instead write in terms of you (Bovee & Arens, 1989). This strategy is consistent, in principle, with the way in which social psychologists have induced self-referencing in past research. Individuals were asked whether trait words were descriptive of themselves (Bower & Gilligan, 1979; Rogers et al., 1977; Wells, Hoffman, & Enzle, 1984), and those engaging in this self-reference task exhibited greater recall for the
86
DEBEVECAND ROMEO
words than those asked just the relative meaning of each word (Rogers et al., 1977) or those instructed to relate the word to another person (Bower & Gilligan, 1979). Self-referencing is induced because it prompts individuals to focus attention on themselves while providing a link with the stimulus. Thus, it should be possible to stimulate self-referencing in an advertising context by addressing individuals in the second person (you) or by name (in a direct marketing or personal selling context) while referring to a product, a need, or a lifestyle characteristic. Individuals should be more likely to engage in self-referencing when exposed to an ad written in the second person than to those written in the first (e.g., a testimonial) or third (most people) person. Writing advertising copy in the second person vernacular is a very simple linguistic manipulation and, therefore, has the potential of being a very powerful persuasive strategy. This strategy is thus deserving of extensive study. Three verbal information strategies have been selected for inclusion in the study to induce varying levels of self-referencing. These verbal strategies include copy focusing either on: (a) product benefits, (b) activities or characteristics of the typical user, or (c) the audience directly (further described as self-referent copy). These three copy forms vary in their focus, whether it be the product, typical users, or the self. Slice-of-life/typical user copy is written in the third person, such as "'people on the g o . . . " Self-referent copy is copy written in the second person and directed toward individuals in the target audience (the pronoun you is used liberally throughout). This strategy is employed to encourage an audience to become personally involved with the message and product and to make them feel that the communication is meant for them as individuals. Although you copy is recommended by advertising experts in order to get an audience's attention (Bovee & Arens, 1989), its effectiveness has not been reported in the literature, and the theoretical reason for its proposed effectiveness has not been investigated. These two copy forms are compared to straight product-oriented copy.
VISUAL IMAGERY AND SELF-REFERENCING Self-referencing may also be stimulated through visual imagery which has been shown to be a very powerful tool in enhancing memory and persuasion (Paivio, 1969, 1971; Paivio & Csapo, 1973). In print ads, advertisers often feature characters who are purportedly like the target audience in a slice-of-life portrayal. The goal of this strategy is for viewers to relate to others who have similar lifestyles or interests and to make the inference that the product will fit into their own lifestyles. However, audience members, in making a similarity judgment, may or may not relate to the characters or situation featured. It is possible that self-referent judgments may be negative and lead
IMAGERY AND SELF-REFERENCING
87
to unfavorable attitudes and intentions if audience members have difficulty relating to the characters or situations. A negative visual imagery effect was noted by Kisielius and Sternthal (1984) who discovered that a vivid picture can undermine persuasion when unfavorable information is recalled from memory when processing occurs. In their study, an ad with a visual image led to a less favorable attitude toward the product than verbal statements that provided product information only. They proposed the "availability valence hypothesis," suggesting that the relative favorableness of a judgment is dependent on whether information is recalled from memory when processing a stimulus ad and the valence of that information. In this study, we examined the effect of visual cues on self-referencing. Two visual conditions were tested and evaluated relative to one another and to verbal strategies designed to encourage varying levels of self-referencing. The two visual treatments included a slice-of-life scene and a product visual. Numerous studies have shown that pictorial stimuli and/or ads are more memorable than verbal ones (Childers & Houston, 1984; Mitchell, 1986; Paivio, 1969; Rossiter & Percy, 1980; Starch, 1966), suggesting a phenomenon termed thepicture superiority effect. Because of the superiority of pictures over words, it is necessary to consider how verbal information may interact with the ad's visual imagery. Other studies have examined the independent effects of visual and verbal cues in ads (Mitchell, 1986) or how verbal cues influence the cognitive processing of ads (Edell & Staelin, 1983). In this study, we examined the extent to which verbal information in ads is processed in response to visual cues. In summary, the purpose of this study was to examine the extent to which verbal and visual strategies encourage favorable self-referencing and, thus, influence attitudes and intentions. In addition, the effect of visual cues on verbal information processing and the interaction of visual and verbal cues was investigated.
SPECIFIC HYPOTHESES Self-Referencing Effect Up to this point, it has been suggested that the verbal and visual information strategies will encourage varying levels of self-referencing. Prior to examining the effect of these strategies, it is first necessary to establish that self-referencing influences evaluations of the ad and product as well as behavioral intentions. The extent to which an ad and/or product are self-referenced needs to be evaluated relative to individuals' attitudes and intentions. In addition, selfreferencing must be examined as a mediator between individuals' perceptions
88
DEBEVEC A N D R O M E O
of each ad and their subsequent attitudes and intentions. It is expected that the more an individual self-references information, the more effective that information will be. Consequently, the first hypothesis is that individuals who experience high levels of self-referencing will have more positive attitudes and intentions than those who experience lower levels, irrespective of the ad viewed. In addition, self-referencing is a mediator between perceptions of the verbal information and visual focus and attitudes and intentions.
Visual Influence on Verbal Information Processing Although there is considerable evidence that information presented visually is more memorable than information presented verbally (Lutz & Lutz, 1977; Mitchell & Olson, 1981; Paivio, 1969, 1971; Paivio & Csapo, 1973), and thus can dominate individuals' attention, it is likely that some visual images will be more conducive to processing verbal information than others. We expected that the visual focus of the ad will influence processing of the verbal information in the following way. When the product is the focal point of an ad's illustration, the audience's information-processing efforts are directed toward processing information about the product only, thus resulting in efficient and continued processing of the verbal information. Viewers may make inferences about the product's usefulness to themselves when encouraged to do so by the verbal copy. Ads that feature typical users of the product (individuals to whom a target audience can relate) may require a deeper level of processing than those that feature only the product. With a typical user visual, the audience may not only make an inference about the product's benefit to themselves, but they may also infer the degree to which they can relate to the characters and lifestyles featured in the ad. If they react negatively to some aspect of the typical user visual, they will likely engage in counterarguments and pay little attention to the accompanying verbal information. Thus, verbal copy should be most effective in encouraging viewers to self-reference when its processing is not inhibited by the visual imagery. In summary, a second hypothesis is that when the visual emphasis in the ad is on the product, individuals' level of self-referencing will vary with the verbal content of an ad, but when the visual emphasis in the ad is on typical users of the product, individuals" level of self-referencing will not vary with the verbal content of an ad. Given that verbal information processing differences are expected when the product visual is featured, it is important to consider how and why the verbal information will be processed differentially and the resultant effectiveness of each verbal strategy. The copy strategies were selected for inclusion in the study because they are believed to encourage varying levels of self-referencing. The more an audience is able to self-reference verbal information, the more effective the verbal copy is expected to be. Of the three copy strategies, the you copy is expected to induce the highest level of self-referencing and, there-
IMAGERY AND SELF-REFERENCING
89
fore, be the most effective. Product benefit copy and typical user copy are expected to result in comparable levels of self-referencing. Therefore, a third hypothesis is that when the visual emphasis in the ad is on the product, self copy will result in more self-referencing than product benefit copy and typical user copy. There should be no significant difference in individuals' self-referencing in response to the latter two copy forms. Self copy should encourage a self-perception process on the part of individuals. Information processed in relation to the self should be more effective than information processed relative to others (Rogers et al., -1977). Research suggests that the "self" is a much more organized structure than the "other"; therefore, less effort is required to process self-relevant information than otherrelevant information (Bower & Gilligan, 1979; Keenan & Baillet, 1980; Kuiper & Rogers, 1979; Rogers, 1981). The self is thought to be one of the richest schemata in memory, thereby prompting deep encoding of information (Kennan & Baillet, 1980). Copy strategies that encourage self-referencing should be superior to copy strategies that encourage person or object perception and lead to more positive attitudes and intentions. Therefore, the last hypothesis suggests that when the visual emphasis in the ad is on the product, self copy will result in more positive attitudes and intentions than product benefit copy and typical user copy. However, there should be no significant difference in individuals' attitudes and intentions in response to the latter two copy forms.
METHOD Sample and Stimuli Participants in the study were 158 junior and senior college students (100 women and 58 men) enrolled in undergraduate marketing courses. The product selected for this study was a soft drink given the hypothetical name, Twist. Undergraduate students purchase and consume this product class, and soft drinks are promoted by both men and women. Thus, the product is one to which study participants should be able to relate, regardless of their gender. The hypotheses were tested in a 3 • 2 factorial design in which there were three verbal copy strategies (self copy, product benefit copy, and typical user copy) and two visual image strategies (product featured and typical users featured). Print was selected as the experimental medium, and ads of approximately equal length were developed. The content of the ads was similar (i.e., all conveyed the same product information) with the exception of pronoun variations (you vs. people) or omission of pronouns (product benefit copy). For example, the self-referencing copy included the headline " W H E N YOU W O R K H A R D A N D PLAY H A R D YOU N E E D TWIST," whereas the body copy read: "TWIST will fit your active lifestyle. It will quench your
90
DEBEVEC A N D R O M E O
thirst with the great taste of real fruit juice. You'll appreciate the added vitamins, too. TWIST will give you the energy you need to keep going." The typical user copy replaced you with people (the headline read: " P E O P L E W H O W O R K H A R D A N D PLAY H A R D D R I N K TWIST"). The product benefit copy simply read, " I N T R O D U C I N G TWIST." The visual featured either a can of Twist beside an enticing glass of the beverage (product visual) or a split screen, with one scene showing students studying and another scene showing a group playing tennis (typical users visual). One of the tennis players held a can of Twist.
Procedure The study was conducted over a 2-week period within which students signed up for 1 of 12 sessions. All treatments were tested in each of the sessions. Subjects were instructed that they would be evaluating a sample of articles and ads for a new magazine designed to appeal to college students who plan to begin a career upon graduation. Each participant was then given a packet that contained an article followed by the experimental ad and a second article also followed by an ad. A series of questions also followed each article and ad, but only those questions following the experimental ad are discussed further. Subjects evaluated each of the two articles and ads so that their attention would not be focused solely on the experimental ad. Participants were instructed to read the first article and respond to the questions. They were then given 2 min to view the experimental ad, after which they responded to a series of questions using 7-point scales that assessed their level of self-referencing, their attitude toward the ad and product, and their behavioral intentions. The same procedure was followed for the second article and ad.
Self-Referencing Measures The measures selected to assess self-referencing reflect how it has been conceptualized and induced in prior research. Shavitt and Brock (1986) suggested that self-referencing is related to Krugman's (1965) notion of involvement in terms of "conscious bridging experiences" (Krugman, 1965, p. 354; Shavitt & Brock, 1986, p. 152). Self-referencing has been described by others as a cognitive process whereby individuals associate self-relevant stimulus information with information previously stored in memory in order to give the new information meaning (Bellezza, 1984; Kuiper & Rogers, 1979; Markus, 1977; Rogers, 1981). In past studies, subjects have been instructed to relate stimulus information to themselves (Bellezza, 1984; Lord, 1980), to think of personal experiences that relate to a stimulus (Bower & Gilligan, 1979), or to picture themselves relative to stimulus information (Shavitt & Brock, 1986; Yalch & Sternthal, 1985). The extent to which individuals self-reference incoming in-
IMAGERY AND SELF-REFERENCING
91
formation should be reflected by the personal relevance of the information, the product's usefulness, and whether the individuals could picture themselves with the product. In this study, the degree to which individuals self-referenced an ad or product was assessed in three ways. The degree to which the ad was selfreferenced was measured by asking subjects how personally relevant the ad washsubsequently referred to as self-referencing of ad. The degree to which the product was self-referenced was measured by asking subjects whether the product was personally useful and personally beneficial. The product selfreferencing measures were averaged to form a composite measure--subsequently referred to as self-referencing of product (Cronbach's alpha = .77). The third self-referencing measure included two questions which assessed whether individuals could picture themselves trying the product and serving the product to a friend as they viewed the ad. Once again, the questions were averaged to form a composite measure, and this measure is subsequently referred to as try/serve (Cronbach's alpha = .90).
Attitude and Intention Measures Composite measures of subjects' attitudes toward the ad and product and their behavioral intentions were developed as a result of principal component analyses with varimax rotation. Subjects' attitude toward the ad was measured through 12 bipolar adjective scales, which were reduced to an Affective and a Cognitive factor. The first factor, which represents atfective impressions of the ad (personal/impersonal, beneficial/not beneficial, impressive/unimpressive, appealing/unappealing, interesting/boring, entertaining/unentertaining, good/bad), accounts for 53.6% of the variance (eigenvalue = 6.97). The second factor, which represents cognitive impressions of the ad (believable/ unbelievable, realistic/unrealistic, informative/uninformative, clear/unclear, logical/illogical), accounts for 9.2% of the variance (eigenvalue = 1.2). The Cronbach's alphas for Affective and Cognitive factors were .93 and .78, respectively. Subjects' attitude toward the product was measured on two sets of scales. The first set of eight scales assessed subjects' ratings of the product. The scales were reduced to a Product Rating factor (unique/ordinary, good/bad, superior/inferior, appealing/unappealing, entertaining/unentertaining, useful/ useless) which accounts for 60.9% of the variance (eigenvalue = 5.48). The second set of scales assessed how subjects thought the product would make them feel. Feelings were measured on six semantic differential scales which were reduced to a Product Feeling factor (special, happy, important, concerned) which accounts for 49.3% of the variance (eigenvalue = 2.96). The Cronbach's alphas for the Product Rating and the Product Feeling factors were .81 a n d . 87, respectively. Finally, intentions were measured through three scales that included the
92
DEBEVECAND ROMEO
likelihood that subjects would try the product, purchase the product at a retail store, and purchase the product in a bar or restaurant if the product were available locally. These measures were averaged to form a composite measure of intention (Cronbach's alpha of .85).
RESULTS The first hypothesis examines the effect of self-referencing on subjects' attitudes and intentions and the mediating property of self-referencing. The more an individual self-referenced the ad a n d / o r product, the more positive their attitudes and intentions were expected to be. The hypothesis received strong support for each of the three self-referencing measures (all ps < .01, see Table 1). Each self-referencing measure was also treated as a covariate relative to the verbal and visual predictor variables, whereas respondents' attitudes and behavioral intentions served as the criterion measure. Self-referencing of the ad or product mediated the effect of the verbal and visual treatments on all dependent variables (allps < .01). Thus, self-referencing plays an integral role in the processing of information in ads. The second hypothesis suggests that the visual image featured in an ad will influence reactions to the verbal information presented. The extent to which viewers self-reference the ad and/or product is expected to vary with the verbal content in an ad when the visual emphasis is on the product, but not when the visual emphasis is on typical users of the product. This hypothesis was investigated by performing three 2 • 3 (Visual Image • Copy Strategy) analyses of variance (ANOVAs) using the self-referencing measures as the dependent measures. The results indicated a significant Verbal X Image interaction for the product self-referencing measure, F(2, 127) = 5.20, p < .01 and marginally significant interactions for the ad, F(2, 127) -----2.39, p ---- .06; and t r y / serve, F(2, 127) = 2.39, p < . 10, self-referencing measures. An investigation of the interactions depicted in Figure 1 indicates that the effectiveness of the verbal copy in inducing self-referencing varied more across the product visual than across the typical user visual. As expected, when the visual emphasis is on the product, self-copy encourages more self-referencing of the product (M = 4.8), ad (M = 4.8), and intentions (M = 4.8), than when the visual emphasis is on the typical user (M = 3.4, M = 4.1, and M = 3.1, respectively). Thus, when accompanied by the typical user visual, self copy ranges from being moderately effective in terms of encouraging self-referencing of the ad to quite ineffective in prompting self-referencing of the product and intentions. These results are discussed further in the investigation of the third hypothesis. Hypothesis 3 addresses the level of self-referencing encouraged by each
IMAGERY AND SELF-REFERENCING
93
TABLE 1 Mean Attitude and Intention Scores by High Versus Low Self-Referencing
Self-Referencing Measure (Dependent Variable) Self-referencing of ad Attitude toward ad Atfective Cognitive Attitude toward product Affective Cognitive Intention Self-referencing of product Attitude toward ad Affective Cognitive Attitude toward product Affective Cognitive Intention Try/serve Attitude toward ad Affective Cognitive Attitude toward product Affective Cognitive Intention
Level of Self-Referencing Low
High
Significance
2.7 4.1
4.6 4.9
< .01 < .0 I
1.8 3.8 2. I
3.0 4.7 4.4
< .01 < .0 i < .01
2.8 4.0
4.7 5.1
< .01 < .01
1.9 3.8 2.2
3.0 4.7 4.5
<.01 <.01 < .01
2.8 4.0
4.9 5.2
< .01 < .01
2.0 3.6 2.2
3.1 4.9 4.8
< .0l < .01 < .01
Note. Individuals were classified as low in self-referencing if they had a self-referencing mean score less than 4.0 and high in self-referencing if their mean score was greater than or equal to 4.0. v e r b a l i n f o r m a t i o n s t r a t e g y g i v e n t h e p r e s e n c e o f the p r o d u c t visual. Self c o p y is e x p e c t e d to p r o d u c e m o r e s e l f - r e f e r e n c i n g t h a n p r o d u c t benefit c o p y a n d t y p i c a l u s e r copy, b u t t h e l a t t e r t w o c o p y strategies s h o u l d p r o d u c e s i m i l a r r e s p o n s e s f r o m viewers. T e s t s o f significant effects w e r e c o n d u c t e d to assess this h y p o t h e s i s . G r o u p m e a n s a n d significant differences a r e r e p o r t e d in T a b l e 2. T h e results i n d i c a t e d t h a t self c o p y d i d e n c o u r a g e m o r e self-referencing o f t h e p r o d u c t , t h e ad, a n d w h e t h e r o r n o t subjects c o u l d see t h e m s e l v e s t r y i n g t h e p r o d u c t t h a n t h e t y p i c a l u s e r copy. T h e self c o p y was significantly m o r e effective t h a n t h e p r o d u c t benefit c o p y for s e l f - r e f e r e n c i n g o f t h e a d a n d o n l y m a r g i n a l l y significant for s e l f - r e f e r e n c i n g t h e p r o d u c t (p < . 10). A s e x p e c t e d , t h e r e w e r e n o significant differences b e t w e e n t h e typical u s e r c o p y a n d t h e
i
C~
i r
94
l
l
l
,
r
v
f
r
v
r
,
v
,
r
l
T
i
v
V
E
f I
0
I ~ l W r l l l , ~ , , J l l , , r ' v r ~
4
~
4
4
4
4
4
4
4
~
.
.
.
.
.
.
.
.
~
TILl rr
U_
95
96
DEBEVEC AND ROMEO TABLE 2 Verbal Information Effectiveness When Product Visual is Featured
Type of Copy Variable
Self-referencing Ad Product Try/serve Attitude toward ad Affeetive Cognitive Attitude toward product Affeetive Cognitive Intention
Self
Product Benefit
Typical User
4.8 (a,b) 4.8 (a) 4.8 (a)
3.3 (a) 3.9 4.2
3.2 (b) 3.6 (a) 3.6 (a)
4.6 (a,b) 5.2 (a)
3.9 (a) 4.9
3.5 (b) 4.4 (a)
3.4 (a,b) 4.8 (a) 4.8(a,b)
2.3 (a) 4.3 3.5 (a)
2.0 (b) 4.1 (a) 3.1 (b)
Note. (a,b) refer to groups that are significantly different from one another at the .05 level of significance. For example, for self-referencingthe ad, 3.3 and 4.8 are significantlydifferent (both designated with "a") as are 4.8 and 3.2 (both designated with "b").
product benefit copy. In addition, when the typical user visual was featured, the self-referencing measures did not vary across the copy strategies. This was evidenced by tests of simple effects (all p s > . 12). The final hypothesis proposes that the self-referencing effects expected would transfer to more positive attitudes and intentions to buy the product. This hypothesis was investigated by performing five 2 • 3 (Visual Image X C o p y Strategy) A N O V A s using the attitudes toward the product and ad and intention to buy as the dependent measures. The results indicated significant Verbal • Image interactions for cognitive attitudes toward the product, F(2, 127) = 4.18, p = .02; affective attitudes toward the product, F(2, 127) = 4.348, p = .02; and intention to buy, F(2, 127) = 3.27, p = .04; and a marginally significant interaction for affective attitude toward the ad, F(2, 127) = 2.5,p = .09. The interaction effect for cognitive attitude toward the product was not significant (p = . 19). A n investigation of the interactions (see Figure 2) indicated that attitudes and intentions varied more across the verbal copy strategies when the product visual was featured than when the typical user visual was featured. Tests of simple effects indicated that when the product visual was featured, attitudes toward the product and ad (both affective and cognitive components) and intentions were significantly higher for the self copy than for the typical user copy. Self copy was significantly more effective than product benefit copy for affective attitudes toward the ad and product and intention to buy (see Table 2). As expected, there were no differences in attitude and intention between the product benefit and typical user copy.
IMAGERY AND SELF-REFERENCING
9")'
Again, when the typical user visual was featured, there were no significant differences for the attitudes and intention measures across the copy strategies (all ps > .24).
DISCUSSION The purpose of this study was to examine the role of self-referent processing in evaluations of ads varying in their visual and verbal focus and the interaction of verbal and visual information. The influence of self-referencing was first assessed, and then the strategies were evaluated in terms of the level of selfreferent thinking encouraged by each strategy and viewers' resultant attitudes and intentions. It was hypothesized that self-referencing would mediate attitudes toward the ad and product and behavioral intentions. As expected, the more an individual self-referenced the product and/or ad, the more favorable were their attitudes and intentions. This suggests that in a persuasion context, self-referencing is a desirable outcome. Thus, it is important for advertisers to understand which visual and verbal strategies encourage self-referent judgments. The level of self-referent processing and subjects' attitudes and intentions were shown to vary with the verbal and visual strategies. The results of this study are significant because they suggest that a simple linguistic manipulation in which ad copy directs viewers' attention to themselves can influence selfreferencing and attitudes and intentions. An equally important and related finding was that the visual focus of the ad influenced processing of the verbal information and, particularly, the self copy. When the product visual was featured, self copy was significantly superior to typical user copy and generally superior to product benefit copy in inducing viewers to self-reference the communication and in gaining more favorable ad and product evaluations and behavioral intentions. In fact, the self copy paired with the product visual was the most effective verbal and visual strategy examined. The effectiveness of the self copy decreased dramatically, however, when it was paired with the typical user visual (as seen in Figures 1 and 2). One explanation for this result may be that subjects did not view themselves as similar to the typical users featured in the ad and made inferences about the users that were inconsistent with their own self-image. When self copy was paired with the product visual, viewers may have made inferences about themselves and the product that were consistent and unlikely to generate negative evaluations. Yalch and Sternthal's (1985) findings, which were reported earlier, may offer additional insight. The product visual accompanied by self copy may have encouraged subjects to link product information to themselves, thereby enhancing the persuasive power of this copy strategy. The typical user visual
o
0
0
0
0
0
0
0
0
0
O
0
0
0
0
0
0
2
J
l I 98
~
l
,
'
l
W
l
W
l
l
l
l
r
l
~
l
l
LU
i
v
~
w
99
100
DEBEVECAND ROMEO
accompanied by self copy may have led subjects to simply focus on themselves relative to the typical users portrayed. Thus, in the former instance, a network of associations to the self may have been formed which should enhance persuasion, whereas in the latter instance, this network of associations may be absent, thus inhibiting persuasion. Yalch and Sternthal concluded that encouraging individuals to link message information to themselves is critical in enhancing persuasion, as opposed to encouraging them to only focus on themselves. Another possible explanation for the findings in this study is that the typical user visual inhibited processing of the verbal information in the ads, whereas the product visual encouraged processing. Support for this explanation lies in the fact that none of the verbal copy strategies were superior to one another when the visual focus was of typical users; instead, all copy strategies were moderately effective. However, there were significant differences in the effectiveness of the copy strategies when the product visual was featured. An examination of the effectiveness of the three copy strategies when paired with each of the two visual strategies raises another potential explanation for some of the findings. Research has shown that learning is facilitated when a visual image and verbal information have figural unity (are congruent) than when they do not (Lippman & Shanahan, 1973). In our study, when the orientation of the copy is congruent with the visual focus of the ad, selfreferencing is sometimes enhanced. This occurred for the self copy and typical user copy but not for the product benefit copy. Self-referencing of the product was significantly higher when self copy was featured with the product visual than when it was featured with the endorser/user visual (M ----4.8 vs. 3.5) and when typical user copy was paired with the endorser/user visual than when paired with the product visual (M = 4.6 vs. 3.6). A significant increase was also noted in the intention self-referencing measure when the self copy was paired with the product visual (M = 4.8 vs. 3.2). However, product benefit copy was no more effective, and in some cases less effective, in enhancing self-referencing when featured with the product visual than when featured with the endorser/user visual--a result contrary to the congruence explanation. Similar results were found when examining the attitude and intention measures (as shown in Figure 2). In general, the effectiveness of the typical user copy and self copy tended to vary considerably with the visual focus of the ad but the product benefit copy did not. Self copy was more likely to be effective when paired with the product visual than when paired with the endorser/user visual, whereas typical user copy was more likely to be effective when paired with the endorser/user visual than with the product visual. Again, product benefit copy was no more effective when paired with the product visual than when paired with the endorser/user visual. In addition, if only congruence were operating, two strategies might be expected to be superior to all others--the product benefit copy accompanied
IMAGERY AND SELF-REFERENCING
1 01
by the product visual and the typical user copy accompanied by the typical user visual. These strategies were both less effective than the self copy paired with the product visual There are also no significant differences in the effectiveness of the copy strategies when the typical user is featured. Thus, the congruence explanation is consistent with some but not all of the results. Further research is warranted to investigate the feasibility of each of these potential explanations in order to yield a more complete conceptualization of self-referencing than is currently available. It is important to understand under what conditions self copy will be effective and when and why the visual image is likely to facilitate or inhibit self-referencing and the effectiveness of the copy strategies. An interesting finding was the absence of main effects for the verbal and visual strategies and the strong interactions noted across all dependent measures. Clearly, individuals are reacting to the visual focus of the ad. What is not so clear is what is going through their minds as they are reacting to the visual element. Are they noting the congruity or incongruity between the verbal and visual strategies? Are they linking message information to themselves, and, if so, is their reaction favorable or unfavorable? Are they making inferences about the product or typical users relative to their perception of themselves? Additional experimentation that partitions out these effects should be revealing, as would the collection of cognitive response data in which subjects' spontaneous thoughts about the product, the ad, and themselves could be assessed. Although this study raises a series of unanswered questions, it also provides a foundation for future research on strategies that may encourage self-referencing and may reveal how the verbal and visual elements in an ad interact to enhance or inhibit the self-referencing effect.
ACKNOWLEDGMENTS Research for this article was supported in part by a University of Massachusetts School of Management summer research grant. We gratefully acknowledge the assistance of Thomas Srull and three anonymous reviewers for comments on an earlier draft of this article.
REFERENCES Bellezza, F. S. (1981). Mnemonic devices: Classification, characteristics and criteria. Review of Educational Research, 51, 247-275. Bellezza, F. S. (1984). The self as a mnemonic device: The role of internal cues. Journal of Personality and Social Psychology, 47, 506-516. Bovee, C., & Arens, W. F. (1989). Contemporary advertising. Homewood, IL: Irwin.
102
DEBEVEC AND ROMEO
Bower, G. H., & Gilligan, S. G. (1979). Remembering information related to one's self. Journal of Research in Personality, 13, 420-432. Caeioppo, J., & Petty, R. (1979). Effects of message repetition and position on cognitive response, recall, and persuasion. Journal of Personality and Social Psychology, 37, 97-109. Childers, T., & Houston, M. J. (1984). Conditions for a picture-superiority effects in consumer memory. Journal of Consumer Research, 11, 643-654. Debevec, K., & Iyer, E. (1988). Self-referencing as a mediator of the effectiveness of sex-role portrayals in advertising. Psychology & Marketing, 5, 71-84. EdeU, J., & Staelin, R. (1983). The information processing of pictures in print advertisement. Journal of Consumer Research, 10, 45-61. Keenan, J.M., & Baillet, S.D. (1980). Memory for personally and socially significant events. In R.S. Nickerson (Ed.), Attention and performance (Vol. 8, pp. 651-659). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Kendzierski, D. (1980). Self schemata and scripts: The recall of self referent and scriptal information. Personality and Social Psychology Bulletin, 6, 23-29. Kisielius, J., & Sternthal, B. (1984). Detecting and explaining vividness effects in attitudinal judgments. Journal of Marketing Research, 21, 54-64. Krngman, H. (1965). The impact of television advertising: Learning without involvement. Public Opinion Quarterly, 29, 349-356. Kuiper, N. A., & Rogers, T. B. (1979). Encoding of personal information: Self--other differences. Journal of Personality and Social Psychology, 37, 499-512. Lippman, M. Z., & Shanahan, M. W. (1973). Pictorial facilitation of paired-associate learning: Implications for vocabulary training. Journal of Educational Psychology, 64, 216-222. Lord, C. G. (1980). Schemas and images as memory aids: Two modes of processing social information. Journal of Personality and Social Psychology, 35, 63-78. Lutz, K. A., & Lutz, R. J. (1977). Effects of interactive imagery on learning: Application to advertising. Journal of Applied Psychology, 62, 493-498. Markus, H. (1977). Self-schemata and processing information about the self. Journal of Personality and Social Psychology, 35, 63-78. Mitchell, A. (1986). The effect of verbal and visual components of advertisements on brand attitudes and attitude toward the advertisement. Journal of Consumer Research, 13, 12-24. Mitchell, A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitudes7 Journal of Marketing Research, 18, 318-332. Paivio, A. (1969). Mental imagery and associative learning and memory. Psychological Review, 76, 241-263. Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart & Winston. Pavio, A., & Csapo, K. (1973). Picture superiority in free recall: Imagery or dual coding. Cognitive Psychology, 5, 176-206. Rogers, T. B. (1981). A model of the self as an aspect of the human information processing system. In N. Cantor & J. F. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp. 193-214). HiUsdale, NJ: Lawrence Erlbaum Associates, Inc. Rogers, T. B., Kuiper, N. A., & Kirker, W. S. (1977). Self-referencing and the encoding of personal information. Journal of Personality and Social Psychology, 35, 677-688. Rossiter, J., & Percy, L. (1980). Attitude change through visual imagery in advertising. Journal of Advertising, 9, 10-16. Shavitt, S., & Brock, T. C. (1986). Self-relevant responses in commercial persuasion. In J. Olson & K. Sentis (Eds.), Advertising and consumer psychology (pp. 149-171). New York: Praeger. Starch, D. (1966). How does the shape of ads affect readership? Media~Scope, 10, 83-85. Wells, G., Hotfman, C., & Enzle, M. (1984). Self versus other-referent processing at encoding and retrieval. Personality and Social Psychology Bulletin, 10, 574--584. Yalch, R., & Sternthal, B. (1985). The effect of self-referencing strategies on messagepersuasiveness. Unpublished manuscript, University of Washington, School of Business Administration, Seattle.