Research Paper & Annotated Bibliography : Advertisement

1030

� 2010 by JOURNAL OF CONSUMER RESEARCH, Inc. ● Vol. 37 ● April 2011 All rights reserved. 0093-5301/2011/3706-0011$10.00. DOI: 10.1086/657430

The Self-Activation Effect of Advertisements: Ads Can Affect Whether and How Consumers Think about the Self

DEBRA TRAMPE DIEDERIK A. STAPEL FRANS W. SIERO

Comparing consumption with nonconsumption situations, we propose and test the self-activation effect of advertisements, which holds that attractiveness-relevant products in advertisements can increase consumer self-activation and lower con- sumer self-evaluation. Four experiments provide support for this effect by showing that after viewing advertised beauty-enhancing products, but not advertised prob- lem-solving products, thoughts about the self are more salient and self-evaluations are lower, compared with viewing the same products outside of an advertisement context. The findings hold for different products and different manipulations. We also present evidence for the mediating role of appearance self-discrepancy ac- tivation as a potential mechanism underlying the effect. The findings suggest that advertisements for attractiveness-relevant products may at times constitute social comparison standards, with which consumers compare themselves.

O ne of the signature strengths of the advertising industrylies in its ability to transform seemingly mundane ob- jects into highly desirable products. Products can change into different entities once they are placed in advertisements. In their everyday appearance, products are relatively distant, self-defining stimuli: shoes are for walking, and cars are for driving. However, the meaning that is conveyed by such

Debra Trampe is assistant professor of marketing at the Department of Marketing at the University of Groningen, P.O. Box 800 9700 AV Gron- ingen, the Netherlands ([email protected]). Diederik A. Stapel is professor of consumer science and director of the Tilburg Institute for Behavioral Economics Research (TIBER), Tilburg University, P.O. Box 90153, 5000 LE Tilburg, the Netherlands ([email protected]). Frans W. Siero is as- sociate professor of statistics and data analysis/social psychology at the University of Groningen, Grote Kruisstraat 2/1, 9712 TS Groningen, the Netherlands ([email protected]). This article is based on the first author’s dissertation, under supervision of the second and third authors. The authors are very grateful to Rik Pieters for his valuable comments on an earlier draft. They would also like to thank the editor, the associate editor, and the reviewers for their helpful feedback and support throughout the review process. The authors thank Reinder Dallinga, Judith Grob, Michael Häfner, Wander Jager, Jennifer Jordan, Marijke Leliveld, Karlijn Massar, Saskia Schwinghammer, Ilona Weydeveld, and Carina Wiekens for help in various stages of the research, and Gundola Kraus and Bernd Batke for designing

products can change dramatically once they are placed in an advertisement. In an advertisement, these products can become potentially self-relevant entities, things that can be bought and owned: shoes and cars are for buying. For ex- ample, the meaning signaled by a high-heeled shoe in an advertisement is radically different from the meaning that is conveyed when one encounters the exact same shoe out- side of its advertisement context. In the latter situation, the shoe is a relatively distant, neutral product. However, in the former advertisement context, the shoe is a desirable, po- tentially self-relevant product that one can buy and that suggests a possible self-image of beauty and attractiveness.

the materials used in the studies. This research was supported in part by grant no. T32 MH19728 from the Dutch Science Foundation (Nederlandse Organisatie voor Wetenschappelijk Onderzoek), awarded to Diederik A. Stapel and Frans W. Siero, and a “Pionier” Grant from the Dutch Science Foundation, awarded to Diederik A. Stapel. Parts of the research reported here were presented at the 2006 meeting of the Dutch Social Psychology Association.

John Deighton served as editor and Laura Peracchio served as associate editor for this article. Electronically published October 14, 2010

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1031

The present research looks at important but understudied consequences of the different meanings that advertised ver- sus nonadvertised products can convey. In four experiments, we address the hypothesis that the different meanings con- veyed by attractiveness-relevant products, once they are placed in an advertisement, can trigger important self-pro- cesses. Specifically, we suggest that advertised attractive- ness-relevant products, as opposed to nonadvertised attrac- tiveness-relevant products, can affect both the extent to which individuals think about the self and how they evaluate the self. We examine the moderating role of product type by distinguishing between beauty-enhancing and problem- solving products. In addition, we explore the mediating role of self-discrepancy activation.

MEANING CONVEYED WITHIN AND BY AN ADVERTISEMENT

Interestingly, the relevant literature has not yet explicitly ac- knowledged the idea that the meaning that is conveyed by advertised products may be very different from the meaning that is conveyed by nonadvertised products. Studies com- monly focus on how the content of advertisements can affect product perceptions, evaluations, and processing style. In other words, most consumer research has studied variations within consumer situations rather than variations between consumer and nonconsumer situations. Or, using Folkes’s (2002, 1) words, “Little consumer research directly compares consumption with non-consumption or customer with non- customer behavior.” The central aim of the current article is to fill a void in the existing literature by making comparisons between consumer and nonconsumer situations.

Despite the gap in the literature, there are a number of perspectives that are relevant for the current research. For example, research on the value of possessions has examined the role that acquiring and using consumer products can play in defining the self (Richins 1994). Also, advertising research has focused on how meaning changes as a function of the way products are represented within (rather than by, which is the present concern) the advertisement context (McQuarrie and Mick 1999; Peracchio and Meyers-Levy 2005; Phillips 1997; Scott 1994). Contemporary consumer research acknowledges that advertisements can possess sym- bolic features and that advertisements are often highly styl- ized representations that compel consumers to engage ad- vertisements as symbolic systems (Aaker 1999; Phillips and McQuarrie 2004; Scott 1994; Solomon 1983). That is, rhe- torical theory posits that advertisements can convey mean- ing that goes beyond their physical characteristics, and as a consequence, consumers must draw upon associations and cultural knowledge to grasp the implicit meaning the ad- vertisement conveys. For example, consider a recent ad- vertisement for Clinique cosmetics. In the advertisement, a bright red lipstick in a shiny silver case is situated next to a patent leather shoe with a high heel in the same bright red color as the lipstick. Both are placed on a shiny surface that subtly mirrors their silhouette. This image is likely to

signal a different meaning than an advertisement that places the same lipstick against a neutral white background so that the lipstick’s features are the most salient part of the ad- vertisement (Scott 1994). Specifically, this second adver- tisement displays the lipstick in a straightforward manner, illustrating what the lipstick looks like, whereas the first advertisement contains a product image that conveys mean- ing that transcends the product’s physical characteristics. That is, in the first advertisement, the implicit message (i.e., one of glamour and attractiveness) is being carried by the image alone. Related to this, the term “implicature” has been coined to refer to information that is implicitly communi- cated to an audience and that is inferred by readers to provide meaning for a message (Phillips 1997; Sperber and Wilson 1986). For example, a teddy bear that is displayed in an advertisement for fabric softener may elicit the implicature of “softness.” Similarly, the aforementioned lipstick that is displayed in the advertisement with the shiny red shoe may elicit the implicature of “attractiveness.”

Thus, existing research has focused mainly on aspects within advertisements (such as implicatures) rather than ex- amining the advertisement’s context per se. That is, much less attention has been given to the notion that an adver- tisement itself may be a symbolic system that may convey meaning that goes beyond its physical features. Instead of considering how implicatures in advertisements may help consumers give meaning to advertisements, in the current research, we suggest that in fact an advertisement itself may constitute an implicature. Similar to the idea that subtle changes within the advertisement can have important im- plications for the way the advertisement is perceived, read, and processed, we suggest that merely placing a product in an advertisement can transform relatively distant products into potentially self-relevant products, which, in turn, may cause dramatic changes in the effects these products exert on their perceivers.

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS

Investigating the effects of the ability of advertisements to change the meaning of the products they display is important since people are confronted with advertisements numerous times a day. Moreover, as Folkes (2002, 1) noted, exposure to advertisements may have dramatic effects on consumers’ perception and evaluation: “When situational cues lead peo- ple to perceive themselves as customers, they then interpret the world differently than when they do not perceive them- selves as customers.”

In the present research, we speculate that the consequences of the ability of advertisements to address people as consumers may be twofold. First, when individuals are confronted with an advertisement and thus are addressed as consumers, they may relate the advertisement to themselves; that is, they need to make consumer choices and decisions about whether they want or need the advertised product. In other words, one con- sequence of depicting products in an advertisement context may

1032 JOURNAL OF CONSUMER RESEARCH

be that advertisements gain in self-relevance: advertisements may cause consumers to think about themselves in relation to the advertised product. Thus, advertisements may affect the extent to which consumers think about themselves, that is, the extent to which the self is activated.

The notion that advertisements can trigger thoughts about the self is appreciated in studies focusing on self-referenc- ing. Consumer researchers in this stream have explored the variety of methods that advertisers can employ to prompt consumers to relate advertisements to themselves. The stud- ies in this literature have generally treated the extent to which research participants think about themselves as an independent variable, for example, by manipulating the type of thoughts people have about the self before advertisement exposure or by varying the advertisement copy (Burnkrant and Unnava 1995; Escalas 2007; Krishnamurthy and Sujan 1999; Meyers-Levy and Peracchio 1996). For example, Es- calas (2007) manipulated the text in an advertisement for running shoes in order to instigate either narrative self-ref- erencing (in which the advertisement copy encouraged par- ticipants to imagine themselves using the product) or ana- lytical self-referencing (in which the advertisement copy addressed participants directly with second-person pro- nouns; i.e., “you”). The main dependent variables of interest were the effects of self-referencing on the subsequent amount and type of thoughts participants had, their emo- tions, and advertisement effectiveness. Although these stud- ies provide valuable insight into how different types of ad- vertisements can elicit different types of self-referencing and how this subsequently affects persuasion, they tell us little about whether advertisements in and of themselves can elicit thoughts about the self in consumers. That is, to the best of our knowledge, no previous study has treated self-referenc- ing as a dependent variable that results from an advertising context.

A second effect that advertisements may exert on con- sumers is that they may affect how consumers think about themselves. We know from research in social cognition (Schwinghammer, Stapel, and Blanton 2006; Stapel and Tes- ser 2001) that increased activation of thoughts about the self leads to an enhanced tendency to incorporate external in- formation into one’s self-view. Also, self-awareness re- search has shown that activating self-constructs raises con- cern for the standards by which one’s actions can be evaluated (Duval and Wicklund 1972; Scheier and Carver 1983). What determines the direction of the outcome of this comparison process? Stapel and Koomen (2001) proposed that when self-related thoughts are activated, self-distinc- tiveness and how one is different from others are empha- sized. Subsequent self-evaluations are likely to contrast away from comparison standards. Indeed, Stapel and Koo- men (2001) showed that when words like “I” or “me” are activated in individuals’ minds, information about upward standards produced self-evaluations that were lower than when information about a downward comparison standard was presented.

Following these earlier findings, we propose that if ad-

vertisements address the self and thereby activate thoughts about the self, then advertisements may, in turn, be used as a standard against which consumers evaluate the self. That is, products in advertisements may instigate a comparison process, thereby causing consumers to evaluate the self in relation to the advertisement, and as a consequence self- evaluations will be lowered.

BOUNDARY CONDITIONS AND MECHANISM

In sum, we hypothesize that an advertisement context can provide consumers with symbolic meaning, leading them to relate the advertisement to themselves. We investigate this proposition in the context of physical attractiveness, as per- sonal care and fashion products are among the top five product categories with the highest advertising expenditures (A.C. Nielsen 2009). Also, physical attractiveness is an important source of female consumers’ self-esteem (Smeesters, Muss- weiler, and Mandel 2010; Thompson et al. 1999). We define attractiveness-relevant products as those products whose us- age affects perceptions of physical attractiveness of the user. Following Bloch and Richins (1992) and Bower and Landreth (2001), within the category of attractiveness-relevant prod- ucts, we distinguish between products that may be considered “enhancing”—used primarily to increase one’s beauty—and products that may be considered “problem-solving” products—used primarily to hide or fix beauty flaws. Bloch and Richins (1992) discuss in their work the functions of adornments and their link to perceptions of attractiveness. In doing so, these authors distinguish three ways in which adorn- ments may affect perceptions of physical attractiveness. Two of them (remedies, such as hair coloring, and camouflage, such as wearing vertical stripes in order to appear taller) seem later to be combined by Bower and Landreth (2001) to form the general category of problem-solving products. The third way in which attractiveness may be increased by the use of adornments, as noted by Bloch and Richins, is enhancers: adornments that directly improve a physical as- pect, such as make-up to accentuate beautiful eyes. This concept bears a direct link to Bower and Landreth’s enhancing products. In sum, based on earlier work, we define problem- solving products as products that may increase one’s attrac- tiveness by downplaying negative aspects of one’s physical attractiveness, while enhancing products serve aesthetic pur- poses by enhancing beauty; they may be beautiful in and of themselves (e.g., shoes, jewelry), and/or they may enhance beauty through their application to the user (eye shadow, perfume). We hypothesize that the type of attractiveness-rel- evant product (enhancing vs. problem-solving) interacts with the advertisement context to affect self-activation and self- evaluation, such that the self-activation effect of advertise- ments holds primarily for enhancing products.

A number of scholars have proposed that products may communicate information about the identity of their pos- sessors (Belk 1988; Richins 1994). The notion of “the ex- tended self ” holds that consumers tend to regard (some of)

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1033

the products they own as part of their self-concept. Thus, some products may, in and of themselves, contain some level of self-relevance for consumers. Enhancing products may suggest an ideal or optimal appearance to consumers, insofar as they embody an ideal or maximum end-state. Considering the importance of appearance in Western societies, products that help consumers attain this important goal are likely to be relatively self-relevant to consumers. The usage of prob- lem-solving products may bring consumers closer in line with their goals and aspirations regarding physical attrac- tiveness as well; however, problem-solving products pri- marily serve to camouflage negative aspects of consumers’ appearances, which essentially moves them away from an unwanted state. Beauty-enhancing products, on the other hand, directly enhance consumers’ attractiveness by bringing them closer to a desired end-state. Because enhancing prod- ucts directly move consumers toward their desired ideal, whereas problem-solving products focus more on masking defects, an advertisement context may make beauty-enhanc- ing products more self-relevant than problem-solving prod- ucts. In line with this reasoning, earlier research has consis- tently found that idealized images in advertising tend to affect consumers to a larger extent than more average depictions (Richins 1991; Smeesters and Mandel 2006; Trampe, Stapel, and Siero 2007). Thus, considering that enhancing products may suggest an ideal appearance and that consumers have been observed to be more readily affected by idealized im- ages, we suggest that enhancing products may be more likely to gain in self-relevance compared to products whose beauty- enhancing qualities are less direct and pronounced (i.e., prob- lem-solving products).

As outlined earlier, advertisements may transform self-de- fining products into desirable, potentially self-relevant prod- ucts that may present to consumers the suggestion of a “more attractive you.” Integrating this notion with the idea that en- hancing products suggest an ideal appearance, we hypothesize that beauty-enhancing products (rather than problem-solving products) may interact with an advertisement context to in- crease their self-relevance. Specifically, we hypothesize that when enhancing products are placed in an advertisement con- text, they will activate the self to a larger extent than when they are not placed in an advertisement context. We also expect this interaction to lead to lower self-evaluations. For problem-solving products, however, we hypothesize that an advertisement context will not affect self-activation and self- evaluation.

If beauty-enhancing products suggest to consumers an ideal appearance and an advertisement context also makes salient ideals and aspirations, then thoughts about consum- ers’ standing relative to this ideal should be likely to be activated in response to exposure to enhancing products in an advertisement context. Indeed, a discrepancy between one’s current appearance and how one would ideally like to be has been found to be related to lower self-evaluations (Halliwell and Dittmar 2006). On a broader level, the norm- setting ability of the mass media (O’Guinn and Shrum 1997; Pollay 1986) has been well documented. As the media, in

general, and product advertising, specifically, tend to be “peopled by fantastic images” (Scott 1994, 252), consumer reality (i.e., consumers’ self-images) has been found to be affected by these fantastic portrayals. For example, paral- leling the typical female body shape that is displayed in the media (the typical model is now as much as 20% under- weight) is the increase in body dissatisfaction among con- sumers (Dittmar, Halliwell, and Stirling 2009). Similarly, heavy media use has been found to be related to overesti- mations of average material wealth: when frequently watch- ing the rich and famous on television, one readily overes- timates the typicality of such lifestyles (O’Guinn and Shrum 1997). Also, media exposure to successful others has been found to instigate a preference to acquire luxury brands (Mandel, Petrova, and Cialdini 2006). The latter likely rep- resents an effort to attain a similar level of success, as prod- ucts communicate information about the identities of their owners (Belk 1988). Central to these findings is the notion that consumers’ actual selves (their appearance, their level of success) are discrepant from what they perceive as their (media-driven) ideal selves. Based on these notions, we hy- pothesize that the self-activation effect of advertisements we proposed earlier may be explained by the activation of an appearance-related self-discrepancy. Thus, advertisements displaying beauty-enhancing (rather than problem-solving) products are likely to remind consumers of their own short- comings. This increased salience of not living up to a stan- dard may lead consumers to evaluate themselves relatively more negatively.

OVERVIEW OF STUDIES In the present research, we make a number of unique pre- dictions that are based on the idea that advertisements can change the self-relational meaning of the attractiveness- relevant products they present. In four experiments, we explore the effect that advertised products versus nonad- vertised products can have on consumers, as well as the moderating role of product type (beauty-enhancing vs. problem-solving products).

In study 1, we address the hypothesis that advertised enhancing (but not problem-solving) products may affect whether consumers think about themselves. That is, en- hancing products within an advertising context may address people’s self-concept, translating into activation of the self. In study 2, we address the hypothesis that viewing advertised enhancing products in an advertisement context also affects how consumers think about themselves. That is, we examine whether enhancing products that appear in an advertisement (as compared to when the advertisement context is absent) exert a comparison effect, affecting the way consumers eval- uate themselves. In study 3, we investigate the role of self- activation as the unique factor associated with the effects of enhancing products in advertisements. Specifically, we test whether the self-evaluative effects of advertised en- hancing products also arise when self-activation is directly manipulated. The goals of study 4 are threefold: (1) to elim- inate potential confounds in the manipulation of the adver-

1034 JOURNAL OF CONSUMER RESEARCH

tisement context in the earlier experiments by employing a different manipulation of advertisement context, (2) to dem- onstrate the basic effect with a wider range of products, and (3) to address the role of self-discrepancy activation as a potential mechanism underlying the findings from the earlier experiments.

PILOT TEST: PRODUCT SELECTION

We first performed a pilot test in order to identify products that were suitable for the later experiments. As the current research focuses on attractiveness-relevant products (those products whose usage affects perceptions of physical attractiveness of the user), we sought to identify a number of products that consumers regard as such. In order to test our hypotheses, within these attractiveness-relevant products, we needed prod- ucts that were regarded primarily as enhancing products, as well as products that were predominantly viewed as problem- solving products. Following Bower and Landreth (2001, 2), we define problem-solving products as those that “serve to fix or hide beauty liabilities or flaws.” Also following these authors, we label beauty-enhancing products as products that serve more aesthetic purposes by enhancing beauty, rather than masking defects. These products may either be inherently beautiful (e.g., a bracelet) or may enhance beauty through their application on the user (e.g., eye shadow).

Method

Participants. Thirty-five female students participated in the study. They completed a questionnaire in small groups in classrooms in exchange for partial course credit.

Materials and Procedure. The instructions on the title page told participants that the study concerned people’s judgments of various products. Specifically, instructions in- formed participants that the researchers were interested in the extent to which people judged various products as beauty related, as well as whether people viewed products as prob- lem-solving or beauty-enhancing products. We briefly ex- plained the meaning of the latter two concepts based on Bower and Landreth (2001). On the next page, a list of 20 products was displayed. In order to determine to what extent products were regarded as attractiveness relevant, for each product, participants indicated on a Likert scale whether they believed that (1) usage of the product does not increase attractiveness to (9) usage of the product increases attrac- tiveness. In order to select products that were primarily seen as problem-solving products, as well as a number of products that were viewed as being primarily enhancing products, participants rated the 20 products on a scale ranging from 1 (problem-solving product) to 9 (enhancing product).

Results and Discussion

After visual inspection of the product means, 6 was chosen as a minimum score for the attractiveness-relevant item (mean scores for the 20 products ranged from 3.69 to 6.94), leaving

a set of 10 products. Within this selection of products, the four products with the lowest mean on the problem-solving versus enhancing dimension were chosen to exemplify the problem-solving products (deodorant, toothpaste, acne con- cealer, and facial wash). The four highest-scoring products on this dimension were selected to exemplify the enhancing products (eye shadow, high-heeled shoe, perfume, and a bracelet). These products constitute the basic set of products from which we chose products for the experiments.

In study 1, we include eye shadow as the enhancing prod- uct and deodorant as the problem-solving product. A paired samples t-test revealed that these two products do not differ with respect to their relevance for attractiveness (M p 6.77, SDeye shadow p 1.37; Mdeodorant p 6.17, SDeye shadow p 2.28; t(34) p 1.25, NS), while eye shadow is regarded as a sig- nificantly more enhancing product than deodorant (Meye shadow p 7.24, SDeye shadow p 2.51; Mdeodorant p 3.47, SDdeodorant p 2.64; t(33) p 4.72, p ! .001). In studies 2 and 3, we add high-heeled shoes as an enhancing product. A repeated mea- sures ANOVA showed that the three products (eye shadow, shoe, and deodorant) did not differ in the extent to which respondents believe that usage increases attractiveness (Mshoe p 6.66, SDshoe p 1.76; F(1, 34) p 1.55, p p .22). Further analyses indicated that the three products differed with re- spect to the extent to which they were regarded as problem- solving versus enhancing products (F(1, 33) p 22.27, p ! .001). Planned contrasts revealed that the deodorant (M p 3.47, SD p 2.64) was rated as significantly more problem solving than the eye shadow (M p 7.24, SD p 2.51; F(1, 33) p 22.27, p ! .001) and the shoe (M p 6.38, SD p 2.42; F(1, 33) p 16.85, p ! .001).

In study 4, we include all eight products described above. A factor analysis on the problem-solving versus enhancing measure confirmed that the four products labeled “enhancing products” had high factor loadings on a first factor, whereas the four products labeled “problem-solving” loaded on a sec- ond factor (the first factor had an eigenvalue of 5.03, ex- plaining 62.88% of the variance; the second factor had an eigenvalue of .88, explaining 11.04% of the variance). We subsequently computed an index for enhancing products (Cronbach’s a p .85), as well as one for the problem-solving products (Cronbach’s a p .89). A paired samples t-test con- firmed that the problem-solving products were rated as sig- nificantly more problem solving (M p 3.33, SD p 2.11) than the enhancing products (M p 6.96, SD p 2.09; t(33) p �5.38, p ! .001).

STUDY 1: ADVERTISEMENTS ACTIVATE THE SELF

The aim of study 1 was to address the hypothesis that an advertised beauty-enhancing product would increase thoughts about the self relative to an advertised problem- solving product. Based on the pilot test, eye shadow ex- emplified the enhancing product and deodorant exem- plified the problem-solving product.

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1035

Method

Participants and Design. Sixty-six female students par- ticipated in the study. Participants were invited to the lab and received course credit for their participation. They were randomly assigned to the conditions of a 2 (context: ad- vertised product vs. nonadvertised product) # 2 (product type: eye shadow vs. deodorant) between-participants design or to a control condition. We used the control condition to simply measure self-activation. We chose the eye shadow to exemplify the enhancing product, whereas the deodorant represented the problem-solving product. Based on one of social comparison theory’s (Festinger 1954) central prop- ositions, the so-called similarity hypothesis, it is assumed that social comparison processes arise in response to ex- posure to targets that are considered relevant rather than irrelevant comparison targets. In this study and the ones that follow, we included only women as research participants as the products included in the studies are primarily relevant for women.

Materials and Procedure. We told participants they would take part in two unrelated studies. A booklet contained all materials and measures. On the first page of the booklet, we informed participants that they would participate in several advertising studies. In the advertisement context conditions, we told respondents that the purpose of the study was to gain more insight into the workings of advertising and people’s responses to advertisements, and hence they would view an advertisement and then answer some questions about this ad. In the nonadvertised conditions, we told respondents that the purpose of the study was to gain more insight into the work- ings of imagery and people’s responses to imagery. We in- formed participants that they would first view an image and that they subsequently would answer some questions about this image. On the second page of the booklet, participants viewed either eye shadow or deodorant that appeared either in an advertisement (advertised product condition) or not (nonadvertised product condition). The advertisements were created especially for this study by a professional graphic designer. They were inspired by existing advertisements. We gave both the eye shadow and the deodorant the fictitious name of Kusco. Participants in the nonadvertised product con- dition viewed only the product depiction, cut out from the advertisement. Figure 1 depicts the materials used in this study. Participants in the control condition did not view prod- ucts. These participants started with the “second” study, which involved the measurement of self-activation.

Self-Activation Measurement. We assessed the level of self-activation that was hypothesized to vary as a function of condition by means of the “Wezwe task” (Dijksterhuis and Van Knippenberg 2000; Stapel and Tesser 2001). This task was contained in the second part of the booklet. We told participants that earlier research had shown that when reading a foreign language, people were sometimes able to guess the correct translation of pronouns. Participants be- lieved that they would read a short story in “Wezwe, a language spoken only in New Guinea.” Participants saw a

short story in a nonexistent language in which 15 pronouns were underlined. Participants’ task was to guess the correct translation of the 15 pronouns. Our main interest was in the number of first-person pronouns (I, me, my) that participants listed. Previous research has shown that heightened self- activation or self-focus leads participants to list relatively more first-person pronouns (Dijksterhuis and Van Knippen- berg 2000; Stapel and Blanton 2004; Stapel and Tesser 2001). After all participants had completed the study, the experimenter thanked them and handed them a debriefing explaining the true purpose of the study.

Results and Discussion

We predicted that exposure to an advertised product would enhance the cognitive accessibility of self-cognitions and that this effect would especially be the case for en- hancing products. Specifically, we predicted that after view- ing the eye shadow ad, participants would list more first- person pronouns than participants in the other conditions.

Because the design incorporated a full factorial with con- trol, we adopted the following analytical strategy. We first tested for main and interaction effects in the 2 (context) # 2 (product type) factorial analysis of variance (ANOVA), thus ignoring the control condition. Then, following the ad- vice of Jaccard (1998), we conducted single degree of free- dom contrasts comparing means in the factorial to the con- trol condition. In the first step, the ANOVA revealed a main effect of context (F(1, 49) p 5.86, p ! .05). The main effect for product type was not significant (F(1, 49) p 2.06, NS). More importantly, the expected context # product type in- teraction was significant (F(1, 49) p 5.14, p ! .05). As a second step, we conducted a one-way ANOVA on the five conditions. This yielded a significant effect of condition (F(4, 61) p 3.86, p ! .01). We then conducted a series of four planned comparisons testing for differences between the four experimental conditions and the control condition. Consistent with our expectations, and as displayed in figure 2, after exposure to the eye shadow advertisement (M p 5.54, SD p 1.33) self-activation was higher than in the control condition (M p 4.08, SD p 1.26; t(61) p �2.86, p ! .01), whereas the other conditions did not differ from the control condition (for the eye shadow image condition (M p 3.77, SD p 1.24; t(61) p .60, NS), the deodorant image condition, (M p 4.08, SD p 1.38; t(61) p 0, NS), and deodorant advertisement, (M p 4.00, SD p 1.30; t(61) p �.15, NS).

The results of study 1 provide evidence that advertised products can activate the self. Moreover, study 1 also sup- ports our prediction that the type of product displayed in the advertisement matters: compared to a no exposure con- trol condition, the self was activated to a larger extent after seeing an advertisement featuring eye shadow (a beauty- enhancing product) than after seeing an advertisement fea- turing deodorant (a problem-solving product).

FIGURE 1

MATERIALS USED IN THE STUDIES: (A) EYE SHADOW ADVERTISEMENT USED IN STUDIES 1 AND 2; (B) NONADVERTISED SHADOW USED IN STUDIES 1–3; (C) SHOE ADVERTISEMENT USED IN STUDIES 2 AND 3;

(D) NONADVERTISED SHOE USED IN STUDIES 2 AND 3; (E) DEODORANT ADVERTISEMENT USED IN STUDIES 1 AND 2; (F) NONADVERTISED DEODORANT USED IN STUDIES 1–3.

NOTE.—Color version available as an online enhancement.

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1037

FIGURE 2

MEAN NUMBER OF FIRST-PERSON PRONOUNS AS A FUNCTION OF CONDITION (STUDY 1)

STUDY 2: ADVERTISEMENTS AFFECT SELF-EVALUATION

We designed study 2 to address the consequences of the finding that advertised products can activate the self. We hypothesized that as a consequence of their ability to activate the self, advertised products, as opposed to nonadvertised products, may be used as a standard against which consum- ers evaluate the self. That is, advertisements may cause con- sumers to evaluate the self in relation to the advertisement. As a consequence, we expected self-evaluations to be lower after exposure to advertised enhancing products, relative to exposure to the same products when outside of an adver- tisement context. We expected no difference in self-evalu- ation for products that were more problem-solving in nature.

Method

Participants and Design. One hundred and eighty-three female students participated in the study for course credit. They were randomly assigned to the conditions of the 2 (context: advertised product vs. nonadvertised product) x 3 (product type: eye shadow vs. shoe vs. deodorant) between- participants design. Again based on the pilot test, the eye shadow exemplified the enhancing product, whereas the de- odorant represented the problem-solving product. Moreover, in study 2, we added the shoe as a beauty-enhancing product. By doing so, we had two distinct enhancing products that we each expected to interact with the advertisement context to affect self-evaluations.

Materials and Procedure. The cover story was similar to the cover story in study 1. We told participants they would participate in several unrelated studies. A booklet contained

all materials and measures. On the first pages of the booklet, we informed participants in the “advertised product” con- ditions they would participate in a consumer study that (os- tensibly) tried to gain insight into the way people responded to advertising. We told participants in the “nonadvertised product” conditions they would participate in a study that (ostensibly) tried to gain insight into the way people respond to various imagery. On the next page, participants viewed, depending on condition, one of three target products that appeared in an advertisement (advertised product condition) or an image of the product (nonadvertised product condi- tion). As in study 1, a professional graphic designer created all materials especially for this study. The materials were inspired by existing advertisements. As in study 1, in the advertised product conditions, we named all three products Kusco. Figure 1 displays all materials used in this study.

Self-Evaluation Measurement. After exposure to the ad- vertised product or nonadvertised product image, partici- pants read: “Previous research has established that human perception and evaluation are to a large extent influenced by people’s personality and attitudes. To be able to control for this influence, it is important that you answer the fol- lowing questions” (Trampe et al. 2007). We included this statement to further increase our confidence that partici- pants’ responses to our self-evaluation questions were only affected by the experimental condition they received and not by possible demand characteristics. Then participants answered, in the reported order, four questions that aimed to measure self-evaluation: “How attractive do you find yourself?” “How satisfied are you with your appearance?” “How satisfied are you with yourself?” and “How satisfied are you with your body?” All answers were given on a 9- point scale (1 p not at all; 9 p extremely). The items were

1038 JOURNAL OF CONSUMER RESEARCH

FIGURE 3

MEAN SELF-EVALUATION AS A FUNCTION OF CONTEXT AND PRODUCT TYPE (STUDY 2)

combined to constitute a self-evaluation index (Cronbach’s a p .90). After all participants had completed the study, they were thanked and received a debriefing explaining the true purpose of the study.

Results and Discussion

For both the eye shadow and shoe, we expected that ex- posure to the product advertisement would result in lower self-evaluations compared to exposure to the nonadvertised depictions. In contrast, we expected that viewing the deo- dorant in either the advertisement or as a nonadvertised depiction would not affect self-evaluations. To test these hypotheses, an ANOVA was performed on the composite self-evaluation measure. This analysis revealed a significant main effect of both context (F(1, 177) p 16.96, p ! .001) and product type (F(2, 177) p 3.90, p ! .05). More im- portantly, however, the expected interaction between context and product type was significant (F(2, 177) p 10.91, p ! .001). As shown in figure 3, consistent with our hypothesis, exposure to the eye shadow resulted in lower self-evalua- tions when it appeared in the advertisement (M p 5.13, SD p 1.15) than when participants were exposed to the non- advertised version of the eye shadow (M p 6.62, SD p 1.18; F(1, 182) p 25.74, p ! .001). Also, when participants were exposed to the shoe in the advertisement, self-evalu- ations were lower (M p 5.72, SD p .94) than when they were exposed to only the nonadvertised depiction of the shoe (M p 6.63, SD p .99; F(1, 182) p 8.66, p ! .01). However, as expected, exposure to the deodorant did not differ as a function of context (F ! 1.5, NS). That is, as expected, self-evaluations were not significantly different after exposure to the deodorant advertisement (M p 6.62,

SD p 1.24) than after the nonadvertised deodorant (M p 6.26, SD p 1.18).

In sum, the results of study 2 provide empirical evidence for our hypothesis that exposure to advertised products can affect consumers’ self-evaluations. Specifically, after seeing an advertisement featuring an enhancing product (eye shadow or a high-heeled shoe) consumers evaluated themselves less positively than after seeing these products when they appeared without the advertisement context, that is, when the brand name was omitted and they were depicted against a white background. As predicted, we found no such difference be- tween advertised products and nonadvertised products for a problem-solving product (i.e., deodorant). These findings sup- port the self-activation effect of ads: when enhancing (but not problem-solving) products appear in an advertisement con- text, they influence whether and how consumers think about the self.

STUDY 3: DIRECTLY MANIPULATING SELF-ACTIVATION

In study 3 we set out to gather further evidence for our hypothesis that the activation of the self is the mechanism underlying the self-evaluative effects we found in study 2. If the results of study 2 were driven by the ability of ad- vertisements to activate thoughts about the self, then directly activating the self should render results similar to the ad- vertised product condition. Specifically, exposure to (non- advertisement) depictions of the eye shadow and the shoe should result in lower self-evaluations when the self has been experimentally activated than when the self has not been activated. In contrast, exposure to the (nonadvertise- ment) image of deodorant should not affect self-evaluations,

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1039

FIGURE 4

MEAN SELF-EVALUATION AS A FUNCTION OF SELF-ACTIVATION AND PRODUCT TYPE (STUDY 3)

independent of whether or not the self has been activated. The goal of study 3 was to empirically test these predictions.

Method

Participants and Design. One hundred and ninety female students participated in the study for course credit. They were randomly assigned to the conditions of the 2 (self-activation: yes vs. no) # 3 (product type: eye shadow vs. shoe vs. deodorant) between-participants design.

Materials and Procedure. The procedure of study 3 was similar to that of study 2, with two exceptions. First, par- ticipants received a self-activation manipulation by means of a word search task. Second, we only used the nonad- vertised depictions of the products as stimulus materials and not the advertisements (see fig. 1). The self-activation task was modeled after a task used by Brewer and Gardner (1996). Participants read a paragraph describing a trip to the city. We instructed them to carefully circle all the pronouns that appeared in the text as part of a word search task. The text was varied so that the same materials were presented with almost all of the pronouns referring to I or me (self- activation condition) or he or his (no self-activation con- dition). This task has been found to be successful in eliciting responses associated with self-activation (Brewer and Gard- ner 1996). Moreover, several subsequent studies have used it to manipulate self-activation (Schwinghammer et al. 2006; Stapel and Koomen 2001; Stapel and Tesser 2001). After completing the task, participants viewed, depending on con- dition, a depiction of the eye shadow, the shoe, or the de- odorant. Next, we measured participants’ self-evaluations using the same items as in study 2. We combined the items to form a composite measure of self-evaluation (Cronbach’s

a p .92). After all participants had completed the study, they were thanked and received a debriefing explaining the true purpose of the study.

Results and Discussion

We expected that self-evaluations would be lower when participants had been exposed to the nonadvertised images of the eye shadow and the shoe when the self was activated compared to when the self was not activated. In contrast, we expected that self-evaluations after viewing the image of the deodorant would not be affected by whether or not the self was activated. To test these hypotheses, an ANOVA was performed on the composite self-evaluation measure. This analysis revealed a significant main effect of self-ac- tivation (F(1, 184) p 22.89, p ! .001). More importantly, however, the expected interaction between self-activation and product type was significant (F(2, 184) p 7.67, p p .001). As can be seen in figure 4, in support of our hy- pothesis, exposure to the image of eye shadow resulted in lower self-evaluations when the self was activated (M p 5.39, SD p 1.21) than when the self was not activated (M p 6.92, SD p .73; F(1, 189) p 30.15). Also, when par- ticipants were exposed to the image of the shoe, self-eval- uations were lower when the self was activated (M p 5.95, SD p 1.02) compared to when the self was not activated (M p 6.70, SD p .93; F(1, 189) p 6.27, p p .01). How- ever, also as expected, after exposure to the image of de- odorant, self-evaluations were not affected by whether the self had been activated (F ! 1, NS). That is, self-evaluations were not significantly different after exposure to the image of the deodorant when the self was activated (M p 6.51, SD p 1.39) compared to when the self was not activated (M p 6.52, SD p 1.20).

1040 JOURNAL OF CONSUMER RESEARCH

STUDY 4: ADVERTISEMENT CONTEXT

Study 4 built on studies 1–3 to increase support for the self- activation effect of ads. First, we added more products per category. Based on the pretest, we chose four products as problem-solving products and four as enhancing products. These products were included in study 4 to examine how they interacted with an advertisement context. In addition, we in- cluded a manipulation check to provide additional validation that the products categories were perceived by participants as intended. A second alteration in study 4 was that we employed a different manipulation of advertisement context. Recall that in the previous experiments, we presented participants with an advertisement or a mere depiction of a product outside of an ad context. This approach may have unintentionally given rise to potential confounds (e.g., including the brand name or not; background design; “drama,” such as spillage of the eye shadow). Therefore, more evidence is needed that it is the advertisement context per se and not some specific prop- erty of the advertisement design that is driving the effects. A third aim of study 4 was to explore a potential underlying process for the current findings. Specifically, we investigated the mediating role of appearance-related self-discrepancy ac- tivation in producing the impact of enhancing products in advertisements on self-evaluations.

Method

Participants and Design. One hundred and seventy-five female students participated in the study for course credit. They were randomly assigned to the conditions of the 2 (context: advertised product vs. nonadvertised product) # 2 (product type: problem-solving vs. enhancing products) between-participants design. The product type variable con- sisted of four products per level, that is, four problem-solv- ing products and four enhancing products.

Materials and Procedure. We told participants they would participate in two unrelated studies. A booklet con- tained all materials and measures. The first page of the booklet informed participants they would participate in several un- related studies. In the advertisement context conditions, we told respondents they would be viewing a product that was recently displayed in an advertising campaign. In the no ad- vertisement context conditions, we merely told participants they would be viewing a product; the statement about the product having featured in an advertisement campaign was omitted. On the second page of the booklet, participants viewed the advertised or the nonadvertised product. In order to maintain the cover story, participants answered a number of questions about the product. They then proceeded to the self-evaluation questions, ostensibly to allow the researchers to control for personality influences on their perceptions of the product. The “second” study was introduced as a study on “ideals and aspirations.” Here, participants responded to questions aimed at capturing self-discrepancies. Finally, par- ticipants rated the product they had viewed earlier on a num- ber of dimensions, including the extent to which they judged

it as being a problem-solving versus enhancing product. After all participants had completed the study, they were thanked and received a debriefing explaining the true purpose of the study.

Stimulus Materials. Based on the results of the pretest, eye shadow, a high-heeled shoe, perfume, and a bracelet exemplified the enhancing products, whereas deodorant, toothpaste, acne concealer, and facial wash exemplified the problem-solving products. Depictions of these products were constructed especially for this study. Product depic- tions were based on existing products; however, these prod- ucts are unavailable in shops in The Netherlands (where the study was conducted).

Self-Evaluation Measurement. Participants then com- pleted the same self-evaluation measure as in study 2. Again, participants provided answers on a 9-point scale (1 p not at all; and 9 p extremely). The items were combined to constitute a self-evaluation index (Cronbach’s a p .73).

Appearance-Related Self-Discrepancy Measurement. On a 9-point scale, participants responded to two items designed to capture self-discrepancies (“To what extent is your current appearance different from your ideal appearance?” and “How large is the difference between your current and your ideal appearance?”). These items were combined to consti- tute a self-discrepancy index (r p .68).

Manipulation Check. Participants rated the products on a scale ranging from 1 (problem-solving product) to 9 (en- hancing product). In order to examine potential confounds, participants also rated their familiarity with the product (1 p never seen it before; 9 p very familiar with it), the frequency with which they use the product (1 p never; 9 p every day), the importance of the product to them (1 p not at all im- portant; 9 p extremely important), and the importance of advertising in their purchase decision for this product (1 p not at all important; 9 p extremely important).

Results and Discussion

Manipulation Check. Consistent with the product survey, the four products that were chosen to exemplify the problem- solving products were also in this study viewed as more prob- lem solving (M p 2.46, SD p .89) than the four enhancing products (M p 7.02, SD p .94; t(173)p �32.95, p ! .001). Individual comparisons between the eight products yielded the same results. Also, familiarity with the product depic- tions was very low (M p 2.09, SD p 1.60) and did not differ between enhancing (M p 2.08, SD p 1.62) and problem-solving (M p 2.09, SD p 1.56) products (t ! 1, NS). Frequency of use did not differ between products (Mproblem-solving p 5.42, SDproblem-solving p 2; Menhancing p 5.17, SDenhancing p 2.18; t ! 1, NS). In addition, product importance did not differ between the two product categories (Mproblem-solving p 2.49, SDproblem-solving p 1.19; Menhancing p 2.56, SDenhancing p 1.05; t ! 1, NS). However, participants indicated that adver- tising plays a larger role in their purchase decision regarding

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1041

FIGURE 5

MEAN SELF-EVALUATION AS A FUNCTION OF CONTEXT AND PRODUCT TYPE (STUDY 4)

enhancing products (M p 5.61, SD p 1.19) compared to problem-solving products (M p 4.87, SD p 1.21; t(173)p �4.24, p ! .001). Further analyses revealed that this effect was driven by a significant difference between the high-heeled shoes (M p 5.77, SD p 1.27) and the deodorant (M p 4.60, SD p .88). Other comparisons between individual products were not significant. Although this suggests a potential con- founding issue, when the main analyses reported below were performed without the shoes and the deodorant, the findings were identical. Therefore, the analyses below include all four problem-solving products and all four enhancing products.

Self-Evaluation. We hypothesized that self-evaluations would be lower after exposure to the products that ostensibly had appeared in an advertisement compared to exposure to the same products that were not associated with advertise- ments. In addition, this difference was only expected for en- hancing products but not for problem-solving products. To test these hypotheses, a 2 (context: advertised product vs. nonadvertised product) # 2 (product type: problem-solving vs. enhancing products) ANOVA was performed on the com- posite self-evaluation measure. This analysis revealed a sig- nificant main effect of both context (F(1, 171) p 31.99, p ! .001), and product type (F(1, 171) p 58.33, p ! .001). More importantly, however, the expected interaction between con- text and product type was significant (F(1, 171) p 27.95, p p .001). As can be seen in figure 5, consistent with findings in study 3, when participants were exposed to an enhancing product that they believed had appeared in an advertisement, self-evaluations were lower (M p 5.11, SD p .48) compared to when the same product was not described as having ap- peared in an advertisement (M p 6.23, SD p .77; F(1, 171) p 61.67, p ! .001). For the problem-solving products, the difference between the advertisement context (M p 6.44,

SD p .71) and the nonadvertised condition (M p 6.48, SD p .73) was not significant (F ! 1, NS). Analyzing the data using a 2 (context: advertised product vs. nonadvertised product) # 8 (product type: eye shadow vs. shoe vs. per- fume vs. bracelet vs. deodorant vs. toothpaste vs. concealer vs. facial wash) design yielded the same pattern of results and significance levels.

Appearance-Related Self-Discrepancy Mediation. We ex- pected self-discrepancy to mediate the effect of the adver- tisement context and product type on self-evaluations. In order to test this hypothesis, we followed suggestions by Baron and Kenny (1986; see also Fennis, Janssen, and Vohs 2009). First, we regressed self-discrepancy on the interaction between advertisement context and product type. This anal- ysis showed that the interaction significantly predicted self- discrepancy (b p .31, t p 4.81, p ! .001). Second, we regressed the dependent variable on the independent vari- able; this analysis showed that the interaction significantly predicted self-evaluation (b p �.31, t p �5.29, p ! .001). As the third step, the regression analysis with the interaction and the mediator as predictors and self-evaluation as the dependent variable showed that both the interaction (b p �.11, t p �2.51, p ! .05) and self-discrepancy (b p �.63, t p �12.86, p ! .001) affect self-evaluation significantly. However, the influence of the interaction on self-evaluation decreased substantially when self-discrepancy was added as a mediator (b was reduced from �.31 to �.11). Indeed, a Sobel test for mediation revealed that this decrease was significant (Z p �4.64, p ! .001). Thus, the data suggest that self-discrepancy partially mediates the effect of the in- teraction between advertisement context and product type on consumer self-evaluations.

The results from study 4 are important in several ways.

1042 JOURNAL OF CONSUMER RESEARCH

First, the general pattern of studies 1–3 was replicated using a broader range of products. Second, potential confounds in the manipulation of the advertisement context used in the previous studies were eliminated by simply telling partici- pants that a product had appeared in an advertisement. Third, study 4 provides empirical evidence for a potential mech- anism underlying the self-activation effect of advertise- ments: the effect was partially mediated by an appearance- related self-discrepancy. Also, we included a manipulation check for product type as well as several potential con- founding variables, such as frequency of use, product im- portance, and the importance of advertising in a purchase decision. Of the three potential confounding variables, only the latter variable differed between enhancing and problem- solving products. However, further inspection of the data revealed that this finding was due to only two products and eliminating these products did not change the pattern of the findings.

GENERAL DISCUSSION

Comparing consumption with nonconsumption situations, in the current research we proposed and tested the self-acti- vation effect of advertisements, which holds that attractive- ness-relevant products in advertisements can affect both whether consumers think about the self and how they think about the self. We designed four experiments to empirically address this hypothesis, as well as examine the moderating role of product type. Taken together, the findings indicate that an advertisement context can interact with the product type (enhancing vs. problem-solving) to affect whether and how consumers think about the self. Using different products and paradigms, beauty-enhancing products in an advertise- ment context activated the self, as compared to the same beauty-enhancing product when it was devoid of its adver- tising context (study 1). Also, consumers used beauty-en- hancing products (as opposed to problem-solving products) in an advertisement context as a standard against which they evaluated themselves. Specifically, in study 2, individuals evaluated their self-image more negatively when they had viewed, for example, a high-heeled shoe (a beauty-enhanc- ing product) in an advertisement than when they were ex- posed to the exact same shoe outside of an advertisement context. Study 3 found that experimentally activating the self interacted with the nature of the product to produce effects on self-evaluation, similar to the effects produced by exposure to advertised products. Finally, study 4 replicated the previous findings by using a different manipulation for the advertisement context. Study 4 also provided prelimi- nary evidence for the role of appearance-related self-dis- crepancy as a partial mediator of the effect.

Contribution and Implications

By providing evidence that products in advertisements, as opposed to nonadvertised products, can affect whether and how consumers think about themselves, the research findings add a critical insight to our understanding of ad-

vertising dynamics. By demonstrating how advertised and nonadvertised products affect consumers differently, the cur- rent studies contribute to the literature on visual imagery in ads. This literature has revealed important knowledge about how consumers draw meaning from ads (McQuarrie and Mick 1999; Peracchio and Meyers-Levy 2005; Phillips 1997; Phillips and McQuarrie 2004), as well as the workings of imagery in advertising (MacInnis and Price 1987). The current investigation adds to this body of research by dem- onstrating that an advertisement context itself can serve as a “figure of thought” that should not be processed literally but rather by using implicit associations and knowledge. Moreover, the current studies show how seemingly minor modifications can transform ordinary products into adver- tisements that are filled with implicatures. Stated differently, the current research adds to the existing literature on rhe- torical theory by examining how people interpret themselves on the basis of advertisements, rather than looking at how people interpret the advertisement on the basis of themselves (i.e., their knowledge structures).

The finding that viewing advertisements can affect the way individuals evaluate themselves has implications for theory and research on social comparison processes. At first this may seem surprising, since social comparison is usually defined as using other people as a basis on which the self is evaluated (Festinger 1954; Richins 1991; Wood 1996). However, Trampe et al. (2007, 111) recently noted that “even objects may constitute relevant comparison tar- gets . . . [this] implies that comparison relevance is de- termined by the social nature of perceivers and not nec- essarily by the social nature of the target.” In line with this idea, the pattern of findings in the current studies 2 and 4 may be interpreted as a social comparison effect: exposure to beauty-enhancing products in an advertisement lowered consumers’ self-evaluations in much the same way as ex- posure to thin models in advertisements has been found to lower self-evaluations (Häfner and Trampe 2009; Smeesters and Mandel 2006). It should be noted that the contrast effect typically observed after exposure to thin models seems a more general phenomenon than the social comparison effects re- ported in the current studies. Social comparison effects in response to nonhuman targets seem to emerge primarily under very specific circumstances. In the current studies, they arose only when both the product and the context were suggestive of an ideal appearance but not in other situa- tions. This “limited effects” finding fits with recent social comparison studies observing that under specific circum- stances, self-evaluative effects may arise after exposure to nonhuman targets. For example, Stapel and Schwing- hammer (2004) showed that when the comparison rele- vance of the target was undisputed (but not when the com- parison relevance is clearly low), people may incorporate information about animals into their self-views. Similarly, Trampe et al. (2007) showed that women who were body dissatisfied (but not body-satisfied women) evaluated themselves more negatively after exposure to a vase that was suggestive of a thin body shape. The current studies

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1043

are among the few studies reporting social comparison effects after viewing nonhuman targets; however, a sys- tematic investigation into the conditions under which social comparison effects emerge in response to nonhuman tar- gets would constitute a fruitful avenue for future research.

Directions for Future Research

The contributions of the present studies to the literature are accompanied by a number of as-yet-unanswered ques- tions. It seems that a particularly promising and relevant direction for future research would be to examine the effect that self-activation and self-evaluation may exert on adver- tising effectiveness. After all, products are advertised be- cause advertisers hope for consumers to buy them. If ad- vertisements may elicit important psychological processes, what are the consequences for product evaluations? Existing research has focused either on the effect of displaying female models in advertisements on self-evaluations (Smeesters and Mandel 2006) or on the effect of female models on product evaluations (Häfner and Trampe 2009), but few studies ex- amine both self-evaluations and product evaluations (Bower 2001; Bower and Landreth 2001; Halliwell and Dittmar 2004). For example, recent research by Bower (2001) and Bower and Landreth (2001) suggests that when consumers compare themselves to the physically attractive models in an advertisement and as a consequence evaluate themselves negatively, advertising effectiveness suffers. However, it is as yet unclear how self-evaluative effects after viewing ad- vertisements that only feature a product and no models can influence product evaluations. When speculating about this, one could come up with competing hypotheses. On the one hand, one could hypothesize that product evaluations will follow self-evaluations: if advertisements give consumers a bad feeling about themselves, individuals may not partic- ularly like the stimulus that was responsible for this poor self-image, causing a decrease in liking for the product. However, on the other hand, one could also argue that neg- ative self-evaluations may serve as a trigger to restore one’s self-view, enhancing product evaluation in turn.

Although in the current studies the focus was on how advertisements address people as consumers, it is important to note that advertisements can increase the salience of other roles as well. For example, advertisements can speak to people in their capacity as citizen, parent, friend, and so forth. It seems likely that advertisements that display prod- ucts that help people attain or enhance these important roles may also increase the activation of the self. However, it is plausible that these types of advertisements produce positive rather than negative effects on self-evaluation. Also, viewing an advertisement for eye shadow may inspire a female con- sumer to invest in her appearance, which may increase her self-evaluation in this domain (Lockwood and Kunda 1997). Indeed, previous research in the social comparison literature has revealed that a host of factors determine whether social comparisons result in contrastive or assimilative effects on the self (Suls and Wheeler 2000). One of the determinants of the direction of the outcome of social comparison is the

method by which self-evaluation is measured. For example, Smeesters and Mandel (2006) showed that measuring self- evaluations with a Likert scale can produce results that are different from results obtained when a free response measure is used. Similarly, Stapel and Suls (2004) provided evidence that the outcome of a comparison process (contrast vs. as- similation) can be a function of whether the comparison is made on an implicit or an explicit level. It is important to keep this in mind when interpreting studies 2 and 4, in which self-evaluation was measured using a Likert scale.

Only recently has self-activation been introduced in re- search on social comparison (Schwinghammer et al. 2006; Stapel and Koomen 2001; Stapel and Tesser 2001; Trampe et al. 2007). These studies generally support the notion that self-activation increases social comparison (Stapel and Tes- ser 2001). However, the outcomes of these comparisons ap- pear to be mixed. For example, Schwinghammer et al. (2006) observed that self-activation promotes self-serving interpretations of social comparison information. In their study 1, self-evaluations were more positive after having been exposed to an upward comparison standard when the self was activated than when the self was not activated. Stapel and Koomen (2001) also report self-defensive re- sponses after self-activation. In contrast, in the current re- search, self-evaluations showed a contrast effect after par- ticipants viewed the beauty-enhancing products when the self was activated by the ad (study 2) or by direct manip- ulation (study 3). How can these seemingly contradictory findings be reconciled? Thus, under what circumstances do people treat social comparison information defensively and under what circumstances do people incorporate social com- parison information into their self-views? One factor that seems important is the comparison situation itself. In studies by Schwinghammer et al. (2006), as well as Stapel and Koo- men (2001), the upward comparison information that was presented to participants consisted of information about a specific individual. Schwinghammer et al. exposed study par- ticipants to a picture of an (un)attractive female student, and Stapel and Koomen presented participants with a short “news- paper” article about a psychology student from their own university. In the presentation of the standard, it is seemed clear to participants that the person was directly relevant as a comparison standard, as the upward comparison information was presented to people in a very direct, explicit manner. In the Schwinghammer et al. and Stapel and Koomen studies, the comparison information was presented to participants much more directly than in the present studies. This explicit presentation could have made it hard for their study partici- pants to discard the standard as not relevant for comparison. One possibility of thwarting the negative implications of the information that is left for participants is for them to respond defensively to questions about self-evaluations.

In contrast, in the studies reported in the present research, the comparison situation is far less direct: participants are pre- sented with an advertisement in which no human model is featured; the ad only displays the product against a relatively simple background (study 2), or participants are merely told a

1044 JOURNAL OF CONSUMER RESEARCH

product recently appeared in an advertisement (study 4). It is plausible that, as a consequence, in the current studies the need to maintain self-esteem was less salient than in the Schwing- hammer et al. and Stapel and Koomen studies and therefore participants may have been less inclined to process the infor- mation in a self-serving manner. That is, it seems plausible that the current experimental situation makes the threat to the self less strong and defensive-responding less urgent and therefore, less likely. These speculations could be fruitfully put to an empirical test in future research.

Recall that in the current research we defined attractiveness- relevant products as those products whose usage affect per- ceptions of physical attractiveness of the user. In the exper- iments, we used products that were rated by study participants as being attractiveness relevant, ranging from perfume to toothpaste. However, what exactly defines a product as at- tractiveness relevant may not be limited to the products we used. For example, one may argue that a (luxury) car or an iPhone may qualify as attractiveness relevant as well insofar as using them affects perceptions of one’s attractiveness. Also, for some products, it is not immediately clear whether they would be seen as problem-solving products or as beauty- enhancing products. For example, the current research pro- vided data suggesting that toothpaste is considered as a pri- marily problem-solving product; however, whitening toothpaste may be seen as a more beauty-enhancing product. Also, the same product may be regarded as a problem-solving product by one consumer while another consumer may view it as an enhancing product, depending on the consumer’s own standing on the relevant attractiveness dimension. Future re- search may investigate the self-activation effect of advertise- ments using a broader range of products as well as products that are more ambivalent with regard to their standing on the problem-solving versus enhancing attributes. Also, taking into account consumers’ attractiveness with regard to the dimen- sion the product relates to (e.g., in the case of toothpaste: consumers’ satisfaction with their teeth) would be a fruitful endeavor for future research.

Because in the current research we included products that were primarily relevant for women, we included only female participants. However, despite the distinct situation in which we tested our hypotheses, we expect similar effects to arise with men who would use products related to male grooming. The crucial factor for the observed effects to emerge is likely to be the relevance of the comparison domain; for men, male products are more relevant than female products, and for women, vice versa. As Festinger (1954) suggested, and sub- sequent studies confirmed, relevant comparison standards in- vite social comparison processes more readily than do less relevant standards. For example, Brown et al. (1992) showed that women’s self-evaluations were negatively affected by exposure to attractive women (relevant comparison standards) but not by exposure to attractive men (less relevant compar- ison standards). In a similar vein, the self-activation effect of advertisements may be most pronounced, or even limited to, situations in which comparison relevance is high.

In addition to product evaluations, the current studies also

bear a link to the important issue of consumer well-being. Recently, in discussions on the impact of media depictions of the female beauty ideal, concern was raised over how the mass media in general and the advertising industry in par- ticular promote messages that may be detrimental to peo- ple’s mental and physical health. As Lasch (as quoted by Pollay 1986, 26) stated as early as the 1970s, “Strictly con- sidered, however, modern advertising seeks to promote not so much self-indulgence as self-doubt. It seeks to create needs, not to fulfill them; it generates new anxieties instead of allaying old ones.” The current research suggests that the ways in which advertisements can make this “self-doubt” salient may not be limited to advertisements that display ultra-thin models. Rather, even seemingly simple advertise- ments that merely display the product and a brand name may trigger thoughts about one’s own shortcomings.

REFERENCES

Aaker, Jennifer L. (1999), “The Malleable Self: The Role of Self- Expression in Persuasion,” Journal of Marketing Research, 36 (February), 45–57.

A. C. Nielsen (2009), Media Expenditures, 2008, Amsterdam: Nielsen Co.

Baron, Reuben M. and David A. Kenny (1986), “The Moderator- Mediator Distinction in Social Psychological Research: Con- ceptual, Strategic, and Statistical Considerations,” Journal of Personality and Social Psychology, 51 (December), 1173–82.

Belk, Russell (1988), “Possessions and the Extended Self,” Journal of Consumer Research, 15 (September), 139–68.

Bloch, Peter H. and Marsha L. Richins (1992), “You Look ‘Mah- velous’: The Pursuit of Beauty and the Marketing Concept,” Psychology and Marketing, 9 (January), 3–15.

Bower, Amanda B. (2001), “Highly Attractive Models in Adver- tising and the Women Who Loathe Them: The Implications of Negative Affect for Spokesperson Effectiveness,” Journal of Advertising, 30 (Fall), 51–63.

Bower, Amanda B. and Stacey Landreth (2001), “Is Beauty Best? Highly versus Normally Attractive Models in Advertising,” Journal of Advertising, 30 (Spring), 1–12.

Brewer, Marilynn B. and Wendi Gardner (1996), “Who Is This ‘We’? Levels of Collective Identity and Self Representations,” Journal of Personality and Social Psychology, 71 (July), 83–93.

Brown, Jonathon D., Natalie J. Novick, Kelley A. Lord, and Jane M. Richards (1992), “When Gulliver Travels: Social Context, Psychological Closeness, and Self-Appraisals,” Journal of Personality and Social Psychology, 62 (May), 717–27.

Burnkrant, Robert E. and H. Rao Unnava (1995), “Self-Referenc- ing: A Strategy for Increasing Processing of Message Con- tent,” Personality and Social Psychology Bulletin, 15 (De- cember), 628–38.

Dijksterhuis, Ap and Ad van Knippenberg (2000), “Behavioral Indecision: Effects of Self-Focus on Automatic Behavior,” Social Cognition, 18 (Spring), 55–74.

Dittmar, Helga, Emma Halliwell, and Emma Stirling (2009), “Un- derstanding the Impact of Thin Media Models on Women’s Body-Focused Affect: The Roles of Thin-Ideal Internalization and Weight Related Self-Discrepancy Activation in Experi- mental Exposure Effects,” Journal of Social and Clinical Psy- chology, 28 (1), 43–72.

THE SELF-ACTIVATION EFFECT OF ADVERTISEMENTS 1045

Duval, Shelley and Robert A. Wicklund (1972), A Theory of Ob- jective Self-Awareness, New York: Academic Press.

Escalas, Jennifer Edson (2007), “Self-Referencing and Persua- sion: Narrative Transportation versus Analytical Elabora- tion,” Journal of Consumer Research, 33 (March), 421–29.

Fennis, Bob M., Loes Janssen, and Kathleen D. Vohs (2009), “Acts of Benevolence: A Limited-Resource Account of Compliance with Charitable Requests,” Journal of Consumer Research, 35 (April), 906–24.

Festinger, Leon (1954), “A Theory of Social Comparison Pro- cesses,” Human Relations, 7 (2), 117–40.

Folkes, Valerie (2002), “Presidential Address: Is Consumer Be- havior Different?” Advances in Consumer Research, 29 (1), 1–4.

Häfner, Michael and Debra Trampe (2009), “When Thinking Is Beneficial and When It Is Not: The Effects of Thin and Round Advertising Models,” Journal of Consumer Psychology, 19 (October), 619–28.

Halliwell, Emma and Helga Dittmar (2004), “Does Size Matter? The Impact of Models’ Body Size on Women’s Body-Focused Anxiety and Advertising Effectiveness,” Journal of Social and Clinical Psychology, 23 (February), 104–22.

——— (2006), “Associations between Appearance-Related Self- Discrepancies and Young Women’s and Men’s Affect, Body Satisfaction, and Emotional Eating: A Comparison of Fixed- Item and Participant-Generated Self-Discrepancies,” Person- ality and Social Psychology Bulletin, 32 (April), 447–58.

Jaccard, James J. (1998), Interaction Effects in Factorial Analysis of Variance, Sage University Papers Series on Quantitative Applications in the Social Sciences, 107–18, Thousand Oaks, CA: Sage.

Krishnamurthy, Parthasarathy, and Mita Sujan (1999), “Ret- rospection versus Anticipation: The Role of the Ad under Retrospective and Anticipatory Self-Referencing,” Journal of Consumer Research, 26 (June), 55–69.

Lockwood, Penelope and Ziva Kunda (1997), “Superstars and Me: Predicting the Impact of Role Models on the Self,” Journal of Personality and Social Psychology, 73 (July), 91–103.

MacInnis, Deborah J. and Linda L. Price (1987), “The Role of Imagery in Information Processing: Review and Extensions,” Journal of Consumer Research, 13 (March), 473–91.

Mandel, Naomi, Petia K. Petrova, and Robert B. Cialdini (2006), “Images of Success and the Preference for Luxury Brands,” Journal of Consumer Psychology, 16 (1), 57–69.

McQuarrie, Edward F. and David G. Mick (1999), “Visual Rhetoric in Advertising: Text-Interpretive, Experimental, and Reader-Re- sponse Analyses,” Journal of Consumer Research, 26 (June), 37–54.

Meyers-Levy, Joan and Laura A. Peracchio (1996), “Moderators of the Impact of Self-Reference on Persuasion,” Journal of Consumer Research, 22 (March), 408–23.

O’Guinn, Thomas C. and L. J. Shrum (1997), “The Role of Tele- vision in the Construction of Consumer Reality,” Journal of Consumer Research, 23 (March), 278–94.

Peracchio, Laura A. and Joan Meyers-Levy (2005), “Using Stylistic Properties of Ad Pictures to Communicate with Consumers,” Journal of Consumer Research, 32 (June), 29–40.

Phillips, Barbara J. (1997), “Thinking into It: Consumer Interpre- tation of Complex Advertising Images,” Journal of Advertis- ing, 16 (Summer), 77–87.

Phillips, Barbara J. and Edward F. McQuarrie (2004), “Beyond Visual Metaphor: A New Typology of Visual Rhetoric in Advertising,” Marketing Theory, 4 (March), 113–36.

Pollay, Richard W. (1986), “The Distorted Mirror: Reflections on the Unintended Consequences of Advertising,” Journal of Marketing, 50 (April), 18–36.

Richins, Marsha L. (1991), “Social Comparison and the Idealized Images of Advertising,” Journal of Consumer Research, 18 (June), 71–83.

——— (1994), “Valuing Things: The Public and Private Meaning of Possessions,” Journal of Consumer Research, 21 (Decem- ber), 504–21.

Scheier, Michael F. and Charles S. Carver (1983), “Self-Directed Attention and the Comparison of Self with Standards,” Jour- nal of Experimental Social Psychology, 19 (May), 205–22.

Schwinghammer, Saskia A., Diederik A. Stapel, and Hart Blanton (2006), “Different Selves Have Different Effects: Self-Acti- vation and Defensive Social Comparison,” Personality and Social Psychology Bulletin, 32 (January), 27–39.

Scott, Linda M. (1994), “Images in Advertising: The Need for a Theory of Visual Rhetoric,” Journal of Consumer Research, 21 (September), 252–73.

Smeesters, Dirk and Naomi Mandel (2006), “Positive and Negative Media Image Effects on the Self,” Journal of Consumer Re- search, 32 (March), 576–82.

Smeesters, Dirk, Thomas Mussweiler, and Naomi Mandel (2010), “The Effects of Thin and Heavy Media Images on Overweight and Underweight Consumers: Social Comparison Processes and Behavioral Implications,” Journal of Consumer Research, 36 (April), 930–49.

Solomon, Michael (1983), “The Role of Products as Social Stimuli: A Symbolic Interactionism Perspective,” Journal of Con- sumer Research, 10 (December), 319–29.

Sperber, Dan and Deirdre Wilson (1986), Relevance: Communi- cation and Cognition, Oxford: Blackwell.

Stapel, Diederik A. and Hart Blanton (2004), “From Seeing to Being: Subliminal Social Comparisons Affect Implicit and Explicit Self-Evaluations,” Journal of Personality and Social Psychol- ogy, 87 (October), 468–81.

Stapel, Diederik A. and Willem Koomen (2001), “I, We, and the Effects of Others on Me: How Self-Construal Moderates So- cial Comparison Effects,” Journal of Personality and Social Psychology, 80 (May), 766–81.

Stapel, Diederik A. and Saskia A. Schwinghammer (2004), “De- fensive Social Comparisons and the Constraints of Reality,” Social Cognition, 22 (February), 147–67.

Stapel, Diederik A. and Jerry Suls (2004), “Method Matters: Ef- fects of Deliberate versus Spontaneous Social Comparisons on Activation, Behavior, and Self-Views,” Journal of Per- sonality and Social Psychology, 87 (December), 860–75.

Stapel, Diederik A. and Abraham Tesser (2001), “Self-Activation Increases Social Comparison,” Journal of Personality and Social Psychology, 81 (October), 742–50.

Suls, Jerry and Ladd Wheeler, eds. (2000), Handbook of Social Comparison, New York: Kluwer.

Thompson, J. Kevin, Leslie J. Heinberg, Madeline Altabe, and Stacey Tantleff-Dunn (1999), Exacting Beauty: Theory, As- sessment, and Treatment of Body Image Disturbance, Wash- ington, DC: American Psychological Association.

Trampe, Debra, Diederik A. Stapel, and Frans W. Siero (2007), “On Models and Vases: Body Dissatisfaction and Proneness to Social Comparison Effects,” Journal of Personality and Social Psychology, 92 (January), 106–18.

Wood, Joanne V. (1996), “What Is Social Comparison and How Should We Study It?” Personality and Social Psychology Bul- letin, 22 (May), 520–37.

Copyright of Journal of Consumer Research is the property of Journal of Consumer Research, Inc. and its

content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's

express written permission. However, users may print, download, or email articles for individual use.

1. Cit p_1:1:
2. Cit p_3:1:
3. Cit p_4:1:
4. Cit p_9:1:
5. Cit p_10:1:
6. Cit p_11:1:
7. Cit p_14:1:
8. Cit p_15:1:
9. Cit p_16:1:
10. Cit p_17:1:
11. Cit p_18:1:
12. Cit p_22:1:
13. Cit p_23:1:
14. Cit p_24:1:
15. Cit p_25:1:
16. Cit p_26:1:
17. Cit p_27:1:
18. Cit p_29:1:
19. Cit p_31:1:
20. Cit p_32:1:
21. Cit p_33:1:
22. Cit p_34:1:
23. Cit p_35:1:
24. Cit p_36:1:
25. Cit p_37:1:
26. Cit p_38:1:
27. Cit p_39:1:
28. Cit p_40:1:
29. Cit p_42:1:
30. Cit p_43:1:
31. Cit p_44:1:
32. Cit p_45:1:
33. Cit p_46:1:
34. Cit p_49:1:
35. Cit p_50:1: