Advertising 7

The effect of exposure to sexual appeals in advertisements on memory, attitude, and purchase intention: a meta-analytic review

John G. Wirtz a, Johnny V. Sparksb and Thais M. Zimbresc

aCharles H. Sandage Department of Advertising, University of Illinois at Urbana-Champaign, Urbana, IL, USA; bDepartment of Journalism, Ball State University, Muncie, IN, USA; cDepartment of Communication, University of California, Davis, CA, USA

ARTICLE HISTORY Received 5 September 2013 Accepted 8 February 2017

ABSTRACT Researchers have often attempted to answer the question, ‘Does sex sell?’ In this article, we present a meta-analysis of studies that used an experiment to test the effect of sexual appeals in ads on memory, attitude, and purchase intention. Our analysis revealed a significant positive effect for sexual appeals on ad recognition and recall (weighted Cohen’s d = .38, p < .001), but the effect on brand recognition and recall was not significant (d = .09, p = .30). We also found that the effect of sexual appeals on attitude towards the ad was not significant (d = ¡0.07, p = .26); however, additional analysis showed that males (d = .27, p < .01) evaluate ads with sexual appeals significantly more positively than females (d = ¡.38, p < .001). Finally, we found a small significant negative effect on brand attitude (d = ¡.22, p < .05), but no effect on purchase intention (d = .01, p = .94).

KEYWORDS Sex in advertising; meta- analysis; sexual appeal; ad recognition and recall; brand recognition and recall; attitude toward ad; attitude toward brand; purchase intention

Introduction

The use of sexual appeals in advertisements has been popular historically and continues to be popular with advertisers in the United States, Europe, and Asia (Huang and Lowry 2012; Nelson and Paek 2008; Reichert and Carpenter 2004). That ads with sexual appeals are used so frequently is somewhat perplexing given that research suggests that ads with sexual content (e.g. nudity, explicit or implicit erotic behaviour, sexual innuendo) produce mixed results. For example, while some studies have found that ads and ad content are remembered more frequently when they feature sexual appeals (Davis and Welsch 1983; Furnham and Hiranandani 2009), other studies have found that including sexual appeals in ads produces no increase in, or in some cases, even lower recall of the ads and brands featured in the ads (Alexander and Judd 1978; Parker and Furnham 2007). Similarly, although a number of studies have found that ads with models who are nude or engaged in overt sexual behaviour are rated more negatively than ads with partial nudity or implied sexual behaviour (Peterson and Kerin 1977; LaTour and Henthorne 1994), more recent

CONTACT John G. Wirtz [email protected]

© 2017 Advertising Association

INTERNATIONAL JOURNAL OF ADVERTISING, 2018 VOL. 37, NO. 2, 168–198 https://doi.org/10.1080/02650487.2017.1334996

studies indicate that audience response to nudity and sexual behaviour may be more nuanced than earlier studies suggest (Black and Morton 2017; Wan, Luk, and Chow 2014).

Because of these conflicting patterns of results, it seems important for researchers to identify consistent findings and salient themes that emerge across the sex in advertising literature. Several scholars have begun to address these issues by providing narrative reviews of the literature (e.g. Percy and Rossiter 1992; Reichert 2002; Wilson and Moore 1979). However, in order to offer a more precise examination of the results of research in this area, a more systematic and quantitative analysis is needed. A relevant option here is a meta-analysis, as meta-analyses allow researchers to combine results from a sample of studies in order to test whether an average effect size differs significantly from zero (Lip- sey and Wilson 2001; Pigott 2012). Combining studies increases the statistical power of main effects tests, as well as facilitating analyses of potential moderator variables (Cohn and Becker 2003).

Therefore, the current article fills a gap in advertising research by presenting the results of a meta-analysis that tested whether exposure to ads with sexual appeals had a signifi- cant effect on ad recognition and recall, brand recognition and recall, attitude towards the ad or brand, and purchase intention.1 In addition to testing for a main effect of sexual appeals on those variables, we also examined the role of several potential moderators, including participant gender, product–sexual appeal congruity, gender of model/s in sex- ual appeals, the year a study was published, and the type of sample (i.e. college students, adults not in college). Knowing whether ads with sexual appeals produce significant posi- tive or negative main effects or conditional effects on memory, attitude, and purchase intention can help advertising researchers understand when, why, and how sexual con- tent in ads influences the persuasion process. Moreover, this information can be valuable to advertising practitioners, as it can help them determine the conditions under which ‘sex sells’ is a viable strategy for promoting a product or brand.

Sexual appeals and advertising

We define a sexual appeal as a persuasion attempt that uses words, images, and/or actions by models appearing in ads to deliver an explicit or implicit sexual message designed to evoke sexual thoughts, feelings, and/or arousal in a target audience (see Reichert, Heckler, and Jackson, 2001; Reichert 2002). One of the earliest examples in the United States of a brand using a sexual appeal occurred in 1871 when the Pearl Tobacco Company included an illustration of a partially nude woman on one of its cigarette packages (Reichert 2003). Since that time, brand managers and strategists have often used ads featuring sexual appeals to promote products and brands (Reichert and Carpenter 2004; Sivulka 2012). Because of the widespread use of sexual appeals, researchers have often focused their attention on studying the phenomenon of sex in advertising, and these efforts can be sep- arated into three broad research foci.

The first research focus consists of studies examining the type and frequency of sexual content appearing in ads (e.g. England and Gardner 1983; Soley and Reid 1988). These studies have established that females and heterosexual couples are the most common type of models used in ads with sexual appeals (Reichert et al. 1999; Soley and Kurzbard 1986), but also that the use of male models in sexual appeals, as well as lesbian and gay male models and erotic images is becoming more prevalent (Reichert and Carpenter

INTERNATIONAL JOURNAL OF ADVERTISING 169

2004). Research in this area also includes studies analysing the quantity and content of ads with sexual appeals appearing in countries in Europe, Asia, and Latin America (Frith, Ping, and Hong, 2005; Huang and Lowry 2012; Nelson and Paek 2008).

The second research focus comprises studies that explore the legal, regulatory, ethical, and cultural implications that ads featuring sexual appeals have on society (e.g. Boddewyn and Kunz 1991; Courtney and Whipple 1983; Gill 2009). This type of research is often informed by critical/cultural theories of race, gender, and sexuality, and scholars in this area sometimes combine research with advocacy for societal change. For example, Court- ney and Whipple (1983) refer to the way that ads with sex appeals perpetuate negative stereotypes about women as ‘The Problem.’ Cohan (2001) goes a step further and describes ads with sexual appeals as frequently ‘brutal and violent, rather than intimate and consensual’ (328), before urging consumers to boycott companies that treat women as sexual objects in their ads.

The third research focus encompasses studies that examine the effects that exposure to ads with sexual appeals have on audiences. Examples include studies presenting the results of surveys and interviews about attitudes towards the use of sexual content in ads (e.g. Christy and Haley 2008; Rouner, Slater, and Domenech-Rodr�ıguez 2003) and studies reporting the results of direct tests of the effects that sexual appeals have on individuals (e.g. Grazer and Kessling, 1995; Putrevu and Swimberghek 2013; Steadman 1969). This lat- ter group of studies not only comprises between- and within-subjects experiments, it also includes quasi-experiments testing whether individual difference variables influence responses to sexual appeals (e.g. sexual schema; Wyllie, Carlson, and Rosenberger 2015; Ye and Zhou 2007), research on context effects associated with the type of TV programme ads appear in (e.g. Bushman 2007; King, McClelland, and Furnham 2015), and studies test- ing whether reactions to ads with sexual appeals vary by country or geographic region (e.g. Cheung et al. 2013; Garcia and Yang 2006; Liu, Cheng, and Li 2009). We would also sit- uate narrative reviews of sex in advertising research here (e.g. Percy and Rossiter 1992; Reichert 2002; Wilson and Moore 1979), as well as empirical analyses, such as the meta- analysis presented in this article.

While each of these three research foci have made important contributions to our understanding of the prevalence and impact that sex in advertising has on society, studies that test and measure effects from exposure to ads with sexual appeals are most relevant to this article. Therefore, in the upcoming sections we provide descriptions and examples of common ways that sexual appeals have been operationalized in previous research. Our goal in these sections is to acquaint readers with concepts and variables that are relevant to the meta-analysis rather than to provide in-depth discussions of each topic, as that information can be found elsewhere (e.g. Reichert, 2002; 2003; Reichert and Lambiase, 2003).

Types of sexual appeals

Nudity

One of the most straightforward ways to operationalize a sexual appeal is to vary the amount of clothing that a model or models wear in an ad. Soley and Reid (1988) proposed a typology to categorize nudity in ads that has been used frequently as the basis for

170 J. G. WIRTZ ET AL.

developing experimental stimuli. The four categories of the typology are demure (e.g. ‘everyday dress’), suggestive (e.g. mini-skirts, short shorts), partially clad (e.g. bathing suits, lingerie), and nude (e.g. models who are wearing only a towel, silhouettes of nude models, models without any clothes), although as Reichert (2002) has observed, even ‘full nudity’ is usually manifested in models exposing only their buttocks and/or breasts but not pri- mary sex organs.

Examples of ways nudity has been operationalized with a female model include using a ‘scantily clad’ woman dressed in a bikini or wearing a bra but no blouse (Cui and Yang 2009; Liu, Cheng, and Li 2009); using a model whose naked breasts are shielded by a prod- uct (LaTour 1990); using a model whose breasts are fully exposed (Peterson and Kerin 1977; Dianoux and Linhart 2010); and using a model who is ‘completely nude’ (Alexander and Judd 1978; Dudley 1999). Similar operationalizations have been used with male– female couples (Belch et al. 1982; Tai 1999) and male models (Reidenbach and McCleary 1983; Simpson, Horton, and Brown 1996).

Explicit or implicit sexual behaviour

In addition to varying how much clothing models wear, researchers have also operational- ized sexual appeals by having a model or models engage in overt or implied behaviours that demonstrate sexual activity, interest, or availability (Reichert 2002). Examples of explicit sexual behaviours include having a woman touch her breasts in a sexual manner (Yan, Ogle, and Hyllegard 2010); having a male and female couple kiss or embrace (Garcia and Yang 2006; Huang 2004; Pope, Vogues, and Brown, 2004); and having a couple engage in sexual touching or using body positions to simulate sexual activity, such as coi- tus (Cheung et al. 2013; Severn, Belch, and Belch, 1990; So 1996). Examples of implicit sex- ual behaviours include having a male and female couple dress in revealing clothing (Reichert and Alvaro 2001); having a model dress or pose in a way that suggests sexual availability (Patzer 1980; Reichert, LaTour, and Kim 2007); and positioning a couple in a way that indicates a sexual encounter is imminent (Horton, Lieb, and Hewitt 1982; Mittal and Lassar 2000; Reichert and Fosu 2005).

Sexual innuendo

Sexual innuendo is a type of sexual appeal that uses words, images, and/or actions of models to deliver a message that can be interpreted as having a sexual or a non-sexual meaning (Reichert 2002). As a result, ads using sexual innuendo require the audience to recognize the sexual interpretation of the ad in order to be effective. A recent Burger King ad provides an example of sexual innuendo. The ad featured a fully clothed female model holding an oblong-shaped sandwich near her open mouth, which was accompanied by the headline, ‘It’ll blow your mind away’. The model’s pose combined with the headline could be interpreted as a reference to oral sex. Bello, Pitts, and Etzel (1983) operationalized sexual innuendo using an ad starring a young but fully clothed Brooke Shields sitting with her legs spread open and stating, ‘You want to know what comes between me and my Calvins? Nothing’. While there was nothing explicitly sexual in Shields’ words or actions, her body posture and the tone of her voice could easily be interpreted as an indication of sexual interest or availability.

INTERNATIONAL JOURNAL OF ADVERTISING 171

Sexual embeds

The final way to operationalize a sexual appeal that we consider is by using a sexual embed, which is a hidden word, picture, or symbol that conveys sexual information to an audience (Rosen and Singh 1992). Researchers who test for effects caused by exposure to sexual embeds hypothesize that people may be influenced by information in ads, even if they are not consciously aware that it exists (Trappey 1996). Sexual embeds have been operationalized using a hidden image of a naked woman in an ad (Rosen and Singh 1992); scratching words, such as ‘sex’ into the side of a pretzel or constructing words, such as ‘fuck’, out of cheese curls (Caccavale, Wanty, and Edell 1982); and as ‘phallic’ and ‘vagi- nal’ images or symbols hidden or partially hidden in ads (Ruth, Mosatche, and Kramer 1989; Ruth 1990).

Control and comparison conditions

Researchers who test for effects caused by exposure to sexual appeals in ads typically include a control or comparison condition with one or more ads without sexual appeals.2

One common approach is to use an ad or ads without any humans, such as ads with only the product and ad copy (e.g. Dahl, Sengupta, and Vohs 2009; Severn, Belch, and Belch 1990) or ads with other types of images (e.g. ‘pastoral scene’; Alexander and Judd 1978). A second approach is to use an ad with human models who are fully clothed (Black, Organ, and Morton 2010) or who represent ‘low sex’ (e.g. Liu, Cheng, and Li 2009) or ‘mild sex’ sit- uations (e.g. Mittal and Lassar 2000). A third approach is to use ads that have been pro- duced for existing products that judges or ‘experts’ have rated as non-sexual (Parker and Furnham 2007; Leka, McClelland, and Furnham 2013) or that have been pretested using a sample of participants from the same population (Das, Galekh, and Vonkeman 2015; Reichert and Alvaro 2001).

Common outcome variables used in sex in advertising research

Before introducing the dependent variables analysed in this meta-analysis, we want to acknowledge that there are other variables we could have included but did not (e.g. emo- tional response, heart rate, sexual arousal). These variables were excluded not because they are unimportant, but because we chose to focus on variables commonly associated with cognitive responses to persuasion. As with the previous sections, our discussion here is brief, as more detailed discussions of these variables can also be found elsewhere (e.g. Eagly and Chaiken 1993; Singh, Rothschild, and Churchill 1988; Spears and Singh 2004; Tulving and Craik, 2000).

Effect of sexual appeals on memory

Ad recognition In ad recognition tests, researchers typically expose participants to an ad or series of ads, and then later in the study present participants with some combination of previously viewed ads or ad content and previously unseen ads or ad content (Du Plessis 2005). Par- ticipants are then asked to differentiate between what they have and have not seen.

172 J. G. WIRTZ ET AL.

For example, Chestnut, LaChance, and Lubitz (1977) showed study participants 50 maga- zine ads with or without ‘sexual or attractive’ models. Later, they showed participants a second set of ads consisting of 25 previously viewed ads and 25 new ads. Participants indi- cated which ads they had and had not seen, and Chestnut et al. compared the recognition scores for participants in the ‘decorative ad’ condition to the recognition scores for partici- pants in the control condition.

Aided and unaided ad recall The ad exposure procedure for ad recall is the same or very similar to the one used in ad recognition tests (Du Plessis 2005). However, rather than asking participants to differenti- ate between ads or ad content they have and have not seen, participants are asked to report what ads or ad content they remember viewing. When using aided or cued recall, researchers give participants some type of ‘cue’, which can prompt participants as they respond to questions. For example, Davis and Welsch (1983) had participants look at four print ads, and later in the study participants were given a list of 16 components of the ads and asked to place the components with the correct ad. In contrast, with unaided or free ad recall, researchers typically use broader questions, which ask participants to describe any ads they saw during a study or to write down anything they remember about those ads. For example, Parker and Furnham (2007) and Furnham and Mainaud (2011) used a similar unaided ad recall procedure in which participants watched a portion of a TV pro- gramme that contained a selection of ads – some with sexual appeals and some without. Immediately after watching the programme, participants were given a blank sheet of paper and asked to write everything they could recall about any of the ads they had seen. The authors then coded the information provided by participants to create ad recall scores for sexual and non-sexual ads.

While there is no singular theory guiding research in this area, one popular theoretical explanation for hypothesized differences between ads with and without sexual appeals is that sexual content attracts attention to ads with sexual appeals and that more cognitive resources are allocated to those ads. This results in higher levels of memory storage, which is reflected in greater recognition and recall of the ads with sexual appeals (Lang 2000). Even studies that do not refer directly to Lang’s (2000) limited capacity model, often use a similar conceptual framework that argues that sexual appeals will attract attention and will produce higher levels of ad recognition and recall (e.g. Ferguson et al. 2010; Reichert and Alvaro 2001).

Therefore, we propose the following hypothesis:

Hypothesis 1: There will be a significant positive effect for sexual appeals on memory, such that recognition and recall of ads with sexual appeals will be higher than recognition and recall of ads without sexual appeals.

Brand recognition and recall While the method and measures used for brand recognition and recall are virtually the same as those for ad recognition and recall, we created a separate section to highlight the fact that some sex in advertising researchers have hypothesized that brands featured in ads with sexual appeals will be recognized less accurately and recalled less frequently than brands featured in ads without sexual appeals (Bushman 2007; Alexander and Judd 1978;

INTERNATIONAL JOURNAL OF ADVERTISING 173

Parker and Furnham 2007; Reichert and Alvaro 2001). One hypothesis is that the increased attention and cognitive resources allocated to the content of the sexual appeal come at the expense of the brands featured in ads with sexual appeals (Alexander and Judd 1978; Judd and Alexander 1983; Reichert and Alvaro 2001). A similar but more narrow hypothe- sis is that the content of certain types of TV programmes (e.g. those with sexual themes or erotic content) may be so engaging that fewer cognitive resources will be available to pro- cess ads and that consequently brand recognition and recall will be lower (Bushman 2007; Parker and Furnham 2007). Although some scholars contend that ads and brands will ben- efit from sexual appeals (e.g. Ferguson et al. 2010; Lang et al. 2003), we offer this hypothesis:

Hypothesis 2: There will be a significant negative effect for sexual appeals on memory, such that recognition and recall of brands featured in ads with sexual appeals will be lower than recognition and recall of brands in ads without sexual appeals.

Effect of sexual appeals on attitude and intention

Attitude towards the ad MacKenzie, Lutz, and Belch (1986) define attitude towards the ad as a ‘predisposition to respond in a favorable or unfavorable manner to a particular advertising stimulus’ (130). Historically, this has been an important variable for advertising scholars, because of its role as a potential mediating variable between exposure to ads and attitude towards the brands featured in those ads (Brown and Stayman 1992; MacKenzie, Lutz, and Belch 1986). However, there is some disagreement about whether attitudes towards ads with sexual appeals will be more positive or negative than ads without sexual appeals.

One argument for proposing that attitude towards the ad will be higher for ads with sex- ual appeals uses the same logic underlying predictions of a positive effect on memory: sex- ual appeals attract attention and increased allocation of cognitive resources, which then are expected to result in more favourable evaluations of those ads (e.g. Reichert, Heckler, and Jackson, 2001). A similar argument is that favourable cognitive and affective responses to sexual appeals will be manifested in more positive attitudes towards the ads (Dudley 1999; Severn, Belch, and Belch, 1990). For example, Dudley (1999) argued that because people respond favourably to attractive models, who are frequently featured in sexual appeals, then ads with sexual appeals will be evaluated more favourably. LaTour (1990) makes a sim- ilar argument, suggesting that the positive emotional response and sexual arousal produced by sexual appeals should also generate more positive attitudes towards those ads.

At the same time, the LaTour (1990) study provides a good example of an argument for why ads with sexual appeals will be evaluated negatively, as LaTour hypothesize that males will have a positive emotional and sexual response to a sexual appeal, so only they will have more positive attitudes towards the ad. Numerous other conditions or variables have been hypothesized to generate negative or less positive ad attitudes, including when models are completely nude (Peterson and Kerin 1977), when a model in a sexual appeal is the same sex as the person evaluating the ad (e.g. Jones, Stanaland, and Gelb 1998), when a sexual appeal is perceived as gratuitous or morally unjust (Dahl, Sengupta, and Vohs 2009; Mittal and Lassar 2000), and when there is a lack of congruity between a product or brand and the perceived appropriateness of using a sexual appeal (e.g. Chang and Tseng 2013; Peterson and Kerin 1977).

174 J. G. WIRTZ ET AL.

We revisit conditional effects in the sections on moderator variables, but for the main effect of sexual appeals on attitude towards the ad, we propose a non-directional hypothesis:

Hypothesis 3: There will be a significant effect for sexual appeals on attitude toward the ad, such that evaluations of ads with sexual appeals will differ significantly than evaluations of ads without sexual appeals.

Attitude towards the brand Spears and Singh (2004) define attitude towards the brand as a stable ‘summary evalua- tion of [a] brand that presumably energizes behavior’ (55). Because positive brand evalua- tions are associated with stronger purchase intentions (MacKenzie, Lutz, and Belch 1986), sex in advertising effects researchers often also include attitude towards the brand as a dependent variable along with attitude towards the ad (e.g. Black, Organ, and Morton 2010; Das, Galekh, and Vonkeman 2015; LaTour and Henthorne 1994). We also see the same or similar theories, hypotheses, and disagreements that we see with attitude towards the ad – some researchers argue that increased attention and positive cognitive and affective responses to sexual appeals will produce higher brand evaluations (Dudley 1999; Severn, Belch, and Belch, 1990), while others argue that brand evaluations will be conditional (Jones, Stanaland, and Gelb 1998; Mittal and Lassar 2000). Therefore, we also propose a non-directional hypothesis for attitude toward the brand:

Hypothesis 4: There will be a significant effect for sexual appeals on attitude toward the brand, such that evaluations of brands featured in ads with sexual appeals will differ signifi- cantly than evaluations of brands featured in ads without sexual appeals.

Purchase intention A purchase intention is ‘an individual’s conscious plan to make an effort to purchase a brand’ (Spears and Singh 2004, 56). Because intentions are positively associated with actual behaviour (Webb and Sheeran 2006), sex in advertising researchers often use pur- chase intention as a proxy for behaviour (e.g. Dudley 1999; Ferguson et al. 2010; LaTour and Henthorne 1994; Wyllie, Carlson, and Rosenberger 2015). Once again, we see similar theoretical arguments, hypotheses, and disagreements about the effect of sexual appeals in ads on purchase intention that we observed in the previous two sections (e.g. Dudley 1999; Mittal and Lassar 2000; Severn, Belch, and Belch, 1990). We also want to note that a number of researchers have measured attitude towards the ad, attitude towards the brand, and purchase intention, and then used the data to identify direct and indirect effects of sexual appeals on those variables (Black, Organ, and Morton 2010; Reichert et al. 2007; Yan, Ogle, and Hyllegard 2010). However, because the focus of this meta-analysis is testing for direct and conditional effects, we propose this hypothesis:

Hypothesis 5: There will be a significant effect for sexual appeals on purchase intention, such that the intention to purchase products featured in ads with sexual appeals differ significantly than the intention to purchase products featured in ads without sexual appeals.

Potential moderators of effects of sexual appeals

Participant gender

There are empirical and theoretical reasons to test participant gender as a potential mod- erating variable. Empirically, research indicates males have stronger reactions to sexual

INTERNATIONAL JOURNAL OF ADVERTISING 175

images (Murnen and Stockton 1997) and that on average males have more positive atti- tudes than females towards nudity, pornography, and sexual permissiveness (Oliver and Hyde 1993; Wright 2013). This empirical approach is represented in many of the early sex in advertisings studies, which did not provide a theoretical justification for examining whether males and females reacted to sexual appeals differently (e.g. Jones, Stanaland, and Gelb 1998; Peterson and Kerin 1977; Simpson, Horton, and Brown 1996. More recently, a number of theoretical reasons have been offered to explain why males and females may respond differently to sexual appeals, including sexual economics theory (Dahl, Sengupta, and Vohs 2009), the influence of culture and socialization on expecta- tions about gender and sexuality (Dianoux and Linhart 2010; Liu, Cheng, and Li 2009), and the ways that evolution has distinctively shaped what men and women value (Lull and Bushman 2015; Kyrousi, Panigyrakis, and Panopoulos 2016). Interestingly, some of the most recent studies in the meta-analysis have questioned whether males and females would respond differently to sexual appeals, given modern trends in both sexes towards more permissive sexual attitudes and behaviours (Ketelaar, Anschutz, and van Hemmen 2014; Wan, Luk, and Chow 2014).

Research question 1:Will males and females differ significantly in their responses to ads with sexual appeals, as measured by memory, attitude, and purchase intention?

Product–sexual appeal congruity

The second moderator variable that we tested was product–sexual appeal congruity, a variable that a number of sex in advertising researchers have already examined (Black and Morton 2017; Chang and Tseng 2013; Peterson and Kerin 1977; Simpson, Horton, and Brown 1996). Conceptually, the argument is that audiences have expectations for prod- ucts and brands and that those expectations extend to the perceived appropriateness of pairing a product or brand with a sexual appeal (Sherman and Quester 2005). A lack of congruity (or relevance or fit) between a product and a sexual appeal is then hypothesized to result in more negative attitudes towards the ad and brand and lower purchase inten- tions (Black and Morton 2017; Chang and Tseng 2013; Simpson, Horton, and Brown 1996). Examples of products tested as product–appeal congruent, include suntan lotion, body and massage oils and perfume, while examples of incongruent products include a wrench or tool set, baked goods, a USB drive, and a breakfast cereal (Black and Morton 2017; Chang and Tseng 2013; Peterson and Kerin 1977; Simpson, Horton, and Brown 1996).3

Research question 2: Will ads with product–sexual appeal congruity have a significantly dif- ferent impact on memory, attitude, and purchase intention than ads with product–sexual appeal incongruity?

Gender of models in sexual appeals

The third moderator that we tested was the gender of the models featured in a sexual appeal. The rationale for including this as a moderator variable is similar to the rationale for testing participant gender, except that the emphasis moves from the participant to the stimulus. Sciglimpaglia, Belch, and Cain (1978) is one of the earliest studies to include model gender as a variable, as they compared how participants responded to ads with a

176 J. G. WIRTZ ET AL.

female model, a male model, and a male–female couple. Subsequently, a few researchers have used model gender as a factor in their studies (Belch et al. 1982; Davis and Welch 1983; Jones, Stanaland, and Gelb 1998; Wan, Luk, and Chow 2014), although direct comparisons between conditions have been surprisingly rare (Jones, Stanaland, and Gelb 1998; Wan, Luk, and Chow 2014). One of the most interesting arguments for why model gender may be an important variable to consider is derived from evolutionary psychology. For example, Wan, Luk, and Chow (2014) hypothesized that females will respond more negatively to female nudity in sexual appeals, because of same-sex rivalry, while males will respond more posi- tively to female nudity because nudity can be a sign of sexual availability (see also Kyrousi, Panigyrakis, and Panopoulos 2016). We believe this is an interesting possibility, and the advantages of the meta-analytic method allow us to perform an exploratory analysis compar- ing responses to ads using females, males, and male–female couples as models.

Research question 3: Will exposure to sexual appeals with female models, male models or male-female couples have significantly different impact on memory, attitude, and purchase intention?

Publishing date and sample type

The final two moderators that we tested were the year a study was published and the type of sample that a researcher used. These variables are frequently tested as moderators in meta-analyses in order to examine whether subgroups of effect sizes vary significantly and whether the sample of studies in a meta-analysis represent different ‘populations’ of studies (Shadish and Sweeney 1991; Pigott 2012). Publication year and sample type are particularly relevant to this meta-analysis. For example, attitudes towards sex and expecta- tions of sex roles have changed significantly over the last 50 years (Kraaykamp 2002; Wells and Twenge 2005). Given that there is almost 50 years between when the first study in the meta-analysis was published (Steadman 1969) and the most recent study (Black and Morton 2017), it is worth examining year as a possible moderator. Similarly, although there is a reasonable argument to be made that testing psychological processes means that using samples of college students is appropriate (Anderson and Bushman 1997), it is still worthwhile to compare whether adult non-college participants have significantly different responses to sexual appeals.

Research question 4a: Will there be any significant differences in memory, attitude, and pur- chase intention between the earliest sex in advertising studies and those published more recently?

Research question 4b: Will there be any significant differences in memory, attitude, and pur- chase intention between studies using a sample of college students and studies using adult non-college student samples?

Method

Search procedures

We used three search methods to identify studies for inclusion in the meta-analysis. First, we searched the Academic Search Premier (EBSCO) and ProQuest online databases, which comprise aggregated content from other academic databases, including Communication

INTERNATIONAL JOURNAL OF ADVERTISING 177

& Mass Media Complete, Psychology & Behavioral Sciences Collection, PsycINFO, PsycArticles, Vocational & Career Collection, Academic Search Complete, and PsycCRITIQUES (excluding the ProQuest newspaper databases). We conducted two Boolean searches within each database using terms associated with sex in advertising research: [sex� AND advertising AND (attitude OR recall OR recognition OR nudity OR purchase intent� OR behavior)] and [nudity AND advertising AND (attitude OR recall OR recognition OR purchase intent� OR behavior OR sex� OR appeal)]. Second, we visited the websites of the Association for Edu- cation in Journalism and Mass Communication (AEJMC), the American Academy of Adver- tising (AAA), and the Association for Consumer Research (ACR), and searched for conference papers and abstracts containing the keywords ‘sex’ or ‘nudity’. When we located an abstract with relevant but incomplete information, we emailed the first author to request a copy of the paper.4 Third, we checked the reference lists of the sex in advertis- ing review articles (i.e. Lull and Bushman 2015; Percy and Rossiter 1992; Reichert 2002; Wil- son and Moore 1979) and the reference lists of publications as they were added to the meta-analysis. These searches produced a combined sample of 2285 journal articles, con- ference papers and abstracts, book chapters, and dissertations and theses that were evalu- ated for possible inclusion in the meta-analysis.

Criteria for inclusion and examples of excluded studies

We identified three criteria for inclusion that reflected the conceptual scope of the meta- analysis (Matt and Cook 1994). The first criterion for inclusion was that a study had to report the results of an experiment with a factor that included one or more treatment lev- els of exposure to an ad or ads with sexual appeals and a non-sexual control or compari- son condition (e.g. mild sexual appeal fully clothed couple; Black, Organ, and Morton 2010). Eligible study designs included between-subjects experiments (including quasi- experiments), within-subjects experiments, and mixed between- and within-subjects experiments. Sexual appeals could be operationalized using one or more of the following: level of nudity, explicit or implicit sexual behaviour, sexual innuendo, or sexual embeds. Excluded by this criterion were studies with physical attractiveness as the main or only independent variable, because our conceptual definition of a sexual appeal comprises more than physical attractiveness and because physical attractiveness has its own litera- ture (e.g. Eagly et al. 1991). Also excluded by this criterion were studies that described models in treatment ads as ‘decorative’ or ‘sexy’ but without providing additional informa- tion describing how those ads used sexual appeals that fit within our definition (e.g. Chestnut, LaChance, and Lubitz 1977; Reid and Soley 1981); studies that examined the effect of exposure to sexual appeals using other research methods, such as surveys, inter- views, or Q-sort (e.g. Lass and Hart 2004; Morrison and Sherman 1972; Wise, Ring, and Merenski 1974); and studies without a non-sexual control or comparison condition (e.g. Reichert, LaTour, and Kim 2007; Reichert and Walker 2005).

The second criterion for inclusion was that a study had to measure the effect of expo- sure to sexual appeals using one or more of the following dependent variables: ad recog- nition; aided or cued ad recall; unaided or free ad recall; brand recognition; aided or cued brand recall; unaided or free brand recall; attitude towards the ad; attitude towards the brand; or purchase intention. Excluded by this criterion were studies using exposure to ads with sexual appeals as an independent variable but that only measured other

178 J. G. WIRTZ ET AL.

dependent variables, such as physiological or psychophysiological responses to stimuli (e.g. Kilbourne, Painton, and Ridley 1985, Study 2; Sparks and Lang 2015).

The third criterion for inclusion was that a study had to report results in a manner that could be used to calculate or estimate a standardized mean difference (i.e. Cohen’s d). Excluded by this criterion were studies reporting the results of experiments as ‘statistically significant’ but with no additional information (e.g. Amyx and Amyx 2011; LaTour and Henthorne 1993); studies reporting only correlations or structural equation models (Huang 2004; LaTour, Pitts, and Snook-Luther 1990; Reichert and Fosu 2005); and studies that reported means for one or more relevant dependent variables but only in the form of charts or graphs (e.g. Vohs, Sengupta, and Dahl 2014; Weller, Sibley, and Neuhaus 1982).

When we found a published and an unpublished version of a study using the same sample, method, and results, we only used the published version in the meta-analysis. For example, Reichert et al. (2011) and Reichert, LaTour, and Ford (2011) have slightly different groups of authors, but otherwise the content and results are the same, so we only included the journal article in our study. By applying the criteria for inclusion and remov- ing duplicates, we obtained a final sample of 72 publications written or published between 1969 and 2017.5

Coding and data extraction

We used a two-step process to collect the data used in the meta-analysis. First, we coded descriptive variables for each publication and study.6 For each publication, we coded the first author’s last name, publication type (journal article, book chapter, conference paper or abstract, thesis or dissertation, other), year published or written, and number of studies relevant to the meta-analysis. For each study, we recorded the sample size, sample type (college or university students, adult non-college), per cent of female participants (recorded or calculated if only sample size and female N provided), number of levels of sexual appeal (including control or comparison condition), product/s appearing in stimu- lus ads (recorded, if provided), gender of model/s in treatment ads (female/s only, male/s only, male–female couple only, mixed males, females, couples, other), and control or com- parison image (non-human, non-sexual human/s, other).

Second, we collected the data needed for the main effects and moderator analyses. We applied two rules to maintain consistency across the sample of studies. First, when studies used more than two treatment levels of sexual appeal (e.g. Alexander and Judd 1978; Peterson and Kerin 1977), only data from the highest level of sexual appeal and the con- trol or comparison condition were coded. Second, when studies included another inde- pendent variable that also featured exposure to sexual content (e.g. Bushman 2007; Parker and Furnham 2007), we only coded data from the ‘low’ or non-sexual condition of the second variable. For example, Parker and Furnham (2007) conducted a 2 (ad type: sex- ual, non-sexual) x 2 (programme type: sexual, non-sexual) experiment, using an episode of ‘Sex in the City’ or ‘Malcolm in the Middle’ to operationalize programme type. In this case, only data from the non-sexual TV programme condition were coded. These rules helped us generate a more accurate estimate of the direct effect of exposure to ads with sexual appeals on memory, attitude, and purchase intention.7

As noted, the effect size used in the meta-analysis was a standardized mean difference (i.e. Cohen’s d), which we calculated by subtracting the mean of the control or comparison

INTERNATIONAL JOURNAL OF ADVERTISING 179

condition from the mean of the treatment condition and then dividing by the pooled standard deviation of the sample (Lipsey and Wilson 2001, 198). Whenever possible, we recorded the means, standard deviations, and cell sizes for those two conditions. When those data were not provided by the authors, we recorded other information that could be used to calculate or estimate a standardized mean difference (e.g. univariate F value, t-test, frequency table, p values; see Lipsey and Wilson 2001, 172–206).8 In cases where exact cell sizes were not provided, we estimated the cell size by dividing the total sample by the relevant number of experimental conditions. For example, Simpson, Horton, and Brown (1996) report that 341 participants were randomly assigned to one of eight treat- ment conditions, but they do not state how many were in each cell. Therefore, we divided 341 by 8 and used n = 42 for the eight treatment conditions.

Because we were interested in differentiating between ad recognition and recall and brand recognition and recall, we coded an outcome as brand recognition or recall only when authors explicitly stated that brand was the target. Outcomes were coded as ad rec- ognition or recall when ads or ad content (e.g. products, brands, image) was the target. For example, Grazer and Keeling (1995) state ‘respondents were asked [to report] what brand names they could recall’ (49). The authors then provide the means and percentages of the brands recalled by participants; this result was coded as unaided brand recall. Con- versely, Furnham and Mainaud (2011) asked participants ‘to write down everything they could remember about the advertisements’ (593) immediately after they had watched a TV programme featuring ads with and without sexual appeals. This was coded as unaided ad recall.

For the moderator analyses, when authors reported results separated by participant gender (e.g. Black, Organ, and Morton 2010; Jones, Stanaland, and Gelb 1998; Simpson, Horton, and Brown 1996) or by product-appeal congruity condition (e.g. Black and Morton 2017; Peterson and Kerin 1977; Simpson, Horton, and Brown 1996), we collected those data. Additionally, we coded outcomes as congruent or incongruent for studies that held the product in stimulus ads constant across treatment and control or comparison condi- tions. For example, Cui and Yang (2009) used a digital camera in all stimulus ads, so all out- comes reported in that study were coded as ‘not congruent’. To determine whether a product was congruent, we used a checklist that we developed based on previous research in this area (e.g. Black and Morton 2017; Chang and Tseng 2013; Reichert 2002; Simpson, Horton, and Brown 1996).9 Studies that mixed congruent and incongruent prod- ucts across sexual appeal conditions were not included in the product-appeal congruity analysis. For example, Ferguson et al. (2010) used ads featuring congruent products (e.g. Victoria’s Secret, Axe body spray) with ads featuring incongruent products (e.g. Microsoft Office) in the sexual and non-sexual conditions of the study, so it was excluded from that analysis.

Intercoder reliability

Our efforts to establish intercoder reliability were informed by Riffe, Lacy, and Fico (2014) and Cooper (2010). Two graduate students were trained over the course of approximately one month on content analysis. One of the authors and the graduate students then inde- pendently coded a sample of five publications. Results were compiled, compared, and dis- cussed. Categories and variable definitions were then refined, and another small sample

180 J. G. WIRTZ ET AL.

of publications was coded independently by the three coders. This process was repeated several times until coding categories were clear, and coders were consistent in their appli- cation of those categories.

A random sample of 14 publications that had not been coded previously was then selected, and the three coders independently coded all descriptive variables. Krippen- dorf’s alpha was calculated using the KALPHA SPSS macro (Hayes and Krippendorf 2007). The results indicated high levels of agreement between the three coders (i.e. between .73 and 1.00).10 The remaining publications were then coded independently with each person coding approximately 30%. For the effect size and moderator data, we adapted Cooper’s (2010) ‘double coding’ procedure in which two coders code all of the data independently and discuss the results and Albarrac�ın et al.’s (2001) iterative coding procedure in which coding categories are continually refined during the coding process. To establish inter- coder reliability, two of the authors coded a sample of 24 studies independently. Krippen- dorf’s alpha ranged from .76 to 1.0 for all variables, except ad and brand recognition and recall.11 Data were compared, and all discrepancies were discussed and resolved. This pro- cedure was repeated until all studies had been coded.

Missing data analysis

Data from each coder were combined into two omnibus data files – one for the descrip- tive analysis and one for the main effects and moderator analyses. We then conducted missing data analyses. For the publication and study descriptive variables, we verified that any missing data had not been provided by the authors (e.g. cell size, per cent of female participants). For the effect size data, we found 17 studies that reported means but no standard deviations or standard errors (e.g. Grazer and Keesling 1995; Alexander and Judd 1978). Of that number, 11 studies provided a univariate F value for sexual appeals, so we calculated the pooled standard deviation for the study sample using the procedure out- lined by Lipsey and Wilson (2001, 182–184). For the six studies that did not provide stan- dard deviations or univariate F values, we imputed the mean of the standard deviations from all other studies reporting that outcome variable and that also used the same type of scale (Furukawa et al. 2006). For example, Putrevu and Swimberghek (2013, Study 2) used seven-point response scales to measure attitude towards the ad, attitude towards the brand, and purchase intention. The authors report means but no standard deviations, so we imputed the mean standard deviation from all studies using a seven-point scale to measure the relevant variables.

Next, we created a data file for each dependent variable, so that the effect of exposure to ads with sexual appeals could be analysed separately. This allowed us to avoid potential problems associated with violating the assumption of independent effects (Hedges and Olkin 1985),12 but we also believe that advertising researchers are more likely to be inter- ested in the effect of sexual appeals on attitude towards the ad and attitude towards the brand rather than the effect of sexual appeals on ad and brand attitude combined. Cate- gorical variables were added to each file to represent the moderator variables using rele- vant data from the descriptive and effects coding files (e.g. female only data coded as ‘0’, male only data coded as ‘1’). A median split was used for the variable ‘year published’, with all studies at or below the median coded as ‘0’, and all studies above the mean coded as ‘1’.

INTERNATIONAL JOURNAL OF ADVERTISING 181

Meta-analytic procedure

We used a random-effects model when we tested for the main effect of sexual appeals in ads on memory, attitude, and purchase intention (Raudenbush 1994; Schmidt, Oh, and Hayes 2009). Random-effects models assume there is a distribution of ‘true’ effect sizes rather than one true effect that is the same across all studies, as is assumed in a fixed effect model (Borenstein et al. 2010). Therefore, we report an estimate of the mean of the distribution of effects for each dependent variable tested against the null hypothesis that the mean of the true effects is 0. One advantage of random-effects models is that they assume effect sizes vary within and between samples of studies, so estimates of standard errors are larger and confidence intervals are wider, which decreases the likelihood of making a Type 1 error (Borenstein et al. 2010; Schmidt, Oh, and Hayes 2009). A second advantage is that the assumptions underlying random-effects models make it appropriate to generalize that results of our meta-analysis beyond the sample of studies in our meta- analysis (Schmidt, Oh, and Hayes 2009).

We used a mixed-effects model for each of the moderator analyses, treating the effects of the moderator variables as fixed and the effect sizes within each level or subgroup of the moderator as random (Lipsey and Wilson 2001). By treating moderator variables as fixed, we can only make inferences about the categories or levels of the moderators that are included in our meta-analysis (Overton 1998). This is appropriate given the nature of the moderator variables included in the study (e.g. participant gender, sample type). For analyses of the effect sizes within each subgroup of the moderators (e.g. female partici- pants, male participants), the assumptions and advantages of random-effects models dis- cussed previously are applicable.

We used the Q statistic to test whether effect sizes vary significantly within a sample of studies and between two or more subgroups of studies (Higgins and Thompson 2002). We use the notation Qw when reporting the results of tests for variance in effect sizes within a group or subgroup of effect, which is also known as a test of heterogeneity of effects. In random-effects models, the assumption is that effect sizes are homogenous, so a significant Qw statistic indicates that the effect sizes in the sample are dispersed more than we would expect due to random error (Borenstein et al. 2010). Because Qw is likely to be significant when large numbers of effect sizes are analysed (Higgins and Thompson 2002), we also report I2 for each estimated mean effect size. The I2 statistic represents an estimate of the per cent of variance in a sample of effect sizes that is due to random error, with a range from 0% (i.e. all variation due to random error) to 100% (i.e. all variation due to study characteristics; Higgins et al. 2003). The I2 statistic can be interpreted as a mea- sure of the amount of inconsistency in a sample of effect sizes, with I2 statistics greater than 75% indicating relatively high levels of variability and inconsistency in effect sizes. We use the notation Qb when reporting the results of moderator analyses or tests for between-group variance in effect sizes. For readers familiar with analysis of variance, the Qb statistic is analogous conceptually to the results of a between-subjects univariate F test (Lipsey and Wilson 2001). A significant Qb indicates there is significant variance between the means of the groups of effect sizes in an analysis and that the moderator variable accounts for some portion of that variance (Pigott 2012).13

All main effects and moderator analyses were conducted using the Comprehensive Meta- Analysis (2014) software. We use N (or n) to refer to the number of participants represented

182 J. G. WIRTZ ET AL.

in a sample of studies, and K (or k) to refer to the number of effect sizes or studies in a sam- ple. Standardized mean differences were calculated such that all ds > 0 represent a positive effect for sexual appeals (e.g. higher recognition or recall, more positive attitudes), and all ds < 0 represent a negative effect for sexual appeals (e.g. lower recognition or recall, more neg- ative attitudes). Prior to conducting inferential statistical tests, all Cohen’s d statistics were corrected for sampling bias and weighted by sample size (Lipsey and Wilson 2001). Therefore, a mean-weighted Cohen’s d is reported for each dependent variable, along with 95% confidence intervals. We then tested whether the mean-weighted Cohen’s d differed significantly from 0 using a z-test with a critical value of 1.96 (a = .05). Z-scores greater than 1.96 indicate a significant positive effect for sexual appeals and z-scores less than¡1.96 indi- cate a significant negative effect for sexual appeals.

Results

Descriptive analysis of studies

The meta-analysis is based on 303 relevant effect sizes representing 17,704 participants collected from 78 studies that were reported in the 72 publications meeting the criteria for inclusion. The median publication year was 2006, with a range of 48 years (min 1969, max 2017). The mean sample size was 228.91 (SD = 267.02) with a range of 1716 (min 30, max 1746) and a median sample size of 160 (see Table 1 for additional results).

Table 1. Descriptive characteristics of publications and studies. N Per cent

Publication type (K = 72) Journal article 62 86.1 Conference paper or abstract 7 9.7 Book chapter 1 1.3 Thesis or dissertation 2 2.8

Sample type (k = 79)a

College or university students 62 78.5 Adults non-college or university 17 21.5

Levels of sex appeal (including control; k = 78) Two levels 60 76.9 Three levels 12 15.4 Four levels 5 6.4 Five levels 1 1.3

Mode of stimulus delivery (k = 79)b

Print ads 64 81.0 Video or TV ads 10 12.7 Other (e.g. billboard, poster) 5 6.3

Gender of models in treatment ads (k = 89)c

Females only 26 29.2 Males only 10 11.2 Male–female couple only 27 30.3 Females, males, and couples mixed 17 19.1 Other (e.g. sexual words or images) 9 10.1

Type of image in control ads (k = 78) Non-human (e.g. product alone) 19 24.4 Non-sexual human model/s 41 52.6 Other (e.g. mixed human, non-human) 18 23.1

Mean per cent of female participants (k = 55)d 52.90 (SD = 9.44) aK = 79 because one study reported data from a college student and a non-college student sample. bK = 79 because one study use more than one mode of delivery as an independent variable. cK = 89 because multiple studies used gender of model as an independent variable. dK = 55 because multiple studies did not report data on the number or percent of female participant.

INTERNATIONAL JOURNAL OF ADVERTISING 183

Effect of sexual appeals on memory

Ad recognition and recall We identified 20 effect sizes for ad recognition (k = 7) and unaided ad recall (k = 13) repre- senting 1511 participants. The main effect of sexual appeals on ad recognition and recall was significant and a mean weighted Cohen’s d = .38 (95% CI [.16, .60]; z = 3.34; p < .001)14 (see Table 2). The same data were used to test whether effect sizes from studies published in 2006 and earlier (k = 8) differed significantly from effect sizes from studies published in 2007 and later (k = 12). The result was significant, Qb(19) = 8.71, p < .001’ however, only the mean for post-2007 studies was significant, d = .63 (95% CI [.48, .78]; z = 8.48.; p < .0001; I2 = 28.36). We also found 11 relevant effect sizes that were used to test product–sexual appeal congruity as a possible moderator variable (congruent, k = 7; incongruent, k = 4). The result was also significant, Qb(10) = 11.87, p < .001, and the mean effect size for congruent products, d = .45 (95% CI [.10, .80]; z = 2.48; p < .05; I2 = 80.26) and incongruent products, d = ¡.46 (95% CI [¡.84, ¡.08]; z = ¡2.39; p < .05; I2 = 60.72) were also both significant (Table 3). We were unable to test participant gender, sample type, or gender of models in stimulus ads, as insufficient data were available for one or more subgroup of these variables.

Brand recognition and recall We located 31 relevant effect sizes for brand recognition (k = 1), aided brand recall (k = 6), and unaided brand recall (k = 24) representing 2674 participants in the sample. The main

Table 2. Main effect of sexual appeals on memory, attitude, and purchase intention. Variable k N Mean d 95% CI z-score Qw (df) I

2

Ad recognition and recalla 20 1511 .38 .16–.60 3.35�� 107.60 (19)��� 82.34 Ad recognition 11 703 .35 .08–.62 2.50� 51.23 (10)��� 80.48 Ad recall 14 1145 .37 .10–.64 2.67�� 73.25 (13)��� 82.25

Brand recognition and recallb 31 2674 .09 ¡.08–.27 1.04 174.92 (30)��� 82.85 Brand recognition 4 368 .56 ¡0.15–1.27 1.54 44.53 (3)��� 93.26 Aided brand recall 6 295 .83 0.11–1.54 2.26� 46.75 (5)��� 89.30 Unaided brand recall 24 2171 ¡.01 ¡.18–.16 ¡0.12 105.47 (23)��� 78.19

Attitude towards the ad 100 11,497 ¡.07 ¡.20–.05 ¡1.13 1,282.51 (99)�� 92.28 Attitude towards the brand 47 4152 ¡.22 ¡.42—.02 ¡2.13�� 526.42 (46)�� 91.26 Purchase intention 54 6964 .01 ¡.17–.18 0.08 673.05 (53)��� 92.13 �p < .05; ��p< .01; ���p < .001. aThree studies provided effect sizes for multiple measures of ad memory, but only data from one outcome per study were used in the combined ad recognition and recall test.

bTwo studies provided effect sizes for multiple measures of brand memory, but only data from one outcome per study were used in the combined brand recognition and recall test.

Table 3. Summary of results of tests of moderator variables. Variable k N Mean d 95% CI z-score I2 Qb (df)

Ad recognition and recall 11.86 (10)��

Product–appeal incongruity 4 583 ¡.46 ¡.84–08 ¡2.39� 60.72 Product–appeal congruity 7 459 .45 .10–80 2.48� 80.26

Attitude towards the ad 18.83 (62)���

Female participants 35 3319 ¡.38 ¡0.60–0.15 ¡3.25�� 89.82 Male participants 28 2037 .27 0.11–1.54 2.26� 77.33

Purchase intention 4.14 (44)�

Female models only 20 3703 ¡.20 ¡52–12 ¡1.29 94.42 Male models only 5 581 ¡.28 ¡.48–08 ¡2.80�� 0.00 Male–female couple only 20 2741 .08 ¡.21–36 0.53 89.44

�p < .05; ��p< .01; ���p < .001.

184 J. G. WIRTZ ET AL.

effect of sexual appeals on brand recognition and recall was not significant, d = .09 (95% CI [¡.08, .27]; z = 1.04; p = .30). For the moderator year published (2006 and earlier, k = 16; 2007 and later, k = 15), the result was not significant, Qb(29) = 1.23, p < .05). We identified 11 effect sizes separated by participant gender (female k = 5; male k = 7), but the effect of participant gender was not significant, Qb(10) = 0.84, p = .36. The test for product–appeal congruity (not congruent, k = 7; congruent, k = 11) also was not significant, Qb(17) = 0.09, p = .77. Finally, we coded 15 effect sizes for gender of model (female only k = 7; male only k = 2; male–female couple k = 6). While the test for model type was significant, Qb(14) = 2.05, p < .05, none of the means were significant, all ps >.20. There was insufficient data to test sample types as a moderator.

Effect of sexual appeals on attitude

Attitude towards the ad We identified 100 effect sizes measuring attitude towards the ad representing a sample of 11,497 participants for the test for a main effect of sexual appeals. The result was not sig- nificant, d =¡.07 (95% CI [¡.20, .05]; z =¡1.13; p = .26). The moderator test for publication year (2006 and earlier, k = 40; 2007 and later, k = 60) also was not significant, Qb(99) = 0.02, p = .89. The test for sample type (college, k = 72; adult non-college, k = 28) was significant, Qb(99) = 5.49, p < .02, but only the within-group mean for the college student sample was significant, d =¡.17 (95% CI [¡.32, ¡.02]; z =¡2.24.; p< .05; I2 = 91.43). For the moderator test for participant gender, we identified 63 relevant effect sizes (female, k = 35; male, k = 28). The result revealed that the means of the two samples differed significantly, Qb(62) = 18.83, p < .001. Further, the within-group mean for females was significant, d = ¡.38 (95% CI [¡.60, ¡.15]; z = ¡3.25; p < .001; I2 = 89.22), as was the within-group mean for males, d = .27 (95% CI [.09, .45]; z = 2.88.; p < .01; I2 = 77.33). For product–appeal congruity, we coded 89 relevant effect sizes (not congruent, k = 51; congruent, k = 38), but the differ- ence between the means was not significant, Qb(88) = 0.54, p = .47. Finally, we coded 86 effect sizes for model gender (females only, k = 47; males only, k = 10; male–female couple only, k = 29). The comparison revealed that the three means did not differ significantly, Qb(85) = 4.94, p = .09, although it is worth noting that the within-group mean for male models was significant, d = ¡.34 (95% CI [¡.62, -.06]; z = ¡2.39.; p < .05; I2 = 71.30).

Attitude towards the brand There were 47 effect sizes representing 4152 participants to test for a main effect of sexual appeals on attitude towards the brand. The result revealed a significant negative effect, d =¡0.22; (95% CI [¡.42,¡.02]; z =¡2.15; p< .05). The same data were used to test publi- cation year (2006 and earlier, k = 21; 2007 and later, k = 26), but the means did not differ significantly, Qb(46) = 1.21, p = .27. Similar results were obtained for sample type (college, k = 41; adult non-college, k = 6), Qb(46) = 1.48, p = .22 and for product–appeal congruity (incongruent, k = 24; congruent, k = 22), Qb(45) = 0.46, p = .50. We found 18 effect sizes separated by participant gender (female, k = 10; males, k = 8), but the difference in means between the samples was not significant, Qb(17) = 0.10, p = .93. For gender of models, we located 45 effect sizes (females only, k = 24; males only, k = 8; couple, k = 13), but a com- parison of the means was not significant, Qb(44) = 1.87, p = .40, although once again the

INTERNATIONAL JOURNAL OF ADVERTISING 185

within-group mean for male models was significant, d = ¡.36 (95% CI [¡.60, ¡.12]; z = ¡2.89.; p < .01; I2 = 48.08).

Purchase intention We identified 54 relevant effect sizes for purchase intention representing 6964 partici- pants. The main effect of sexual appeals on purchase intention was not significant, d = 0.01 (95% CI [¡.17, .18]; z = 0.08; p = .94). The test for publication year (2006 and earlier, k = 22; 2007 and later, k = 32) found that the means of the two samples did not differ sig- nificantly, Qb(53) = 1.27, p = .26. The test for sample type (college, k = 41; non-college, k = 13) also was not significant, Qb(53) = 0.88, p = .77. There were only 16 effect sizes sepa- rated by gender (males, k = 8; female, k = 8), but the difference between the two samples was not significant, Qb(15) = 0.15, p = .70. For product–appeal congruity, there were 50 rel- evant effect sizes (not congruent, k = 19; congruent, k = 31), and the result revealed the means differed significantly, Qb(49) = 5.03, p < .05, however, only the within-group mean for non-congruent products was significant, d = ¡.24 (95% CI [¡.42, ¡.06]; z = ¡2.56.; p < .01; I2 = 75.13). For model type, we identified 45 effect sizes (females only k = 20; males only k = 5; male–female couple k = 20), and we found that the means differed signifi- cantly, Qb(44) = 4.14, p < .05. However, similar to the results for brand and ad attitude, only the within-group mean for males only was significant, d = ¡.28 (95% CI [¡.48, ¡.08]; z = ¡2.80.; p < .01; I2 = 0.0).

Discussion

We conducted a meta-analysis of studies reporting the results of experiments that tested the effect of exposure to sexual appeals in ads on memory, attitude, and purchase inten- tion. Although our primary interest was testing whether the main effect of sexual appeals in ads was significant, we included moderator analyses to explore whether the effects of sexual appeals differed significantly for important subgroups, such as male and female participants; ads for products that are congruent with audience expectations about using sexual appeals and ads for products that are not congruent with expectations; and stimu- lus ads featuring female models, male models, or male–female couples as models.

Effect of sexual appeals on memory

The first noteworthy result of the meta-analysis was that there was a significant positive main effect for sexual appeals on ad recognition and recall (H1), meaning that ads with sexual appeals and their content were remembered more than ads and ad content with- out them. This result is consistent with theories of attention and message processing that argue that increased attention and allocation of cognitive resources produce higher levels of encoding and message storage, which then facilitates message retrieval and more accurate recognition and recall of messages (Lang 2000). However, we did not find a posi- tive or negative effect for sexual appeals on brand recognition and recall (H2). A similar null effect for sexual appeals on brand memory was also reported by Lull and Bushman (2015). Taken together, these results suggest that hypotheses proposing that the content of sexual appeals will ‘distract’message recipients from remembering the brands featured in those ads should be re-evaluated or reformulated. At the same time, we did not find a

186 J. G. WIRTZ ET AL.

significant positive effect for sexual appeals on brand, which has also been hypothesized (Lang et al. 2003; De Pelsmacker 1996). We did find that the mean effect size for aided brand recall was positive and significant and that the mean effect size for brand recogni- tion suggests a positive effect for sexual appeals, so it is possible that the effect of sexual appeals on brand memory is significant, while also not strong enough to be captured using unaided brand recall (Leigh, Zinhan, and Swaminathan 2006).

With regard to the moderator analyses, only one result for ad and brand recognition and recall seems worth further consideration – the significant difference in ad recognition and recall between ads with and without product–appeal congruity. We found that when ads with sexual appeals featured congruent products (e.g. perfume, massage oils), they were recognized more accurately and recalled more frequently than when they featured incongruent products (e.g. bread, USB drive). This result can also be interpreted using an attention – cognitive resources allocation framework (Lang 2000), whereby product– appeal congruity facilitates message encoding and storage whereby product–appeal con- gruity requires fewer cognitive resources, when then facilitates message encoding and storage and thereby attenuates those processes. Although this interpretation is consistent with other recent results (Samson 2016; Sparks and Lang 2015), we also need to be cau- tious given the small number of effect sizes in each subgroup.

Effect of sexual appeals on attitude and purchase intention

A second interesting finding was that averaged across 100 effect sizes representing more than 11,000 participants we found attitudes towards ads with sexual appeals were no dif- ferent than attitudes towards ads without them (H3). Given the statistical power associ- ated with such a large sample of effect sizes (Cohn and Becker 2003), we were somewhat surprised at this result, although we did note that many researchers test conditional hypotheses when attitude towards the ad is a dependent variable (e.g. Chang and Tseng 2013; Jones, Stanaland, and Gelb 1998; Mittal and Lassar 2000). We also found that the effect of sexual appeals on purchase intention was not significant (H5), although we did find a small but significant negative effect for sexual appeals on attitude towards the brand (H4). The null effects on ad attitude and purchase intention and the negative effect on brand attitude suggest that more sophisticated hypotheses than ‘spillover effects’ from the physical attractiveness of models used in ads (Dudley 1999; Severn, Belch, and Belch, 1990) should be pursued.

With regard to moderator variables, a third noteworthy – albeit possibly unsurprising – finding was that participant gender moderated the effect of sexual appeals on attitude towards the ad. We use the phrase ‘possibly unsurprising’, because so many researchers have explored differences by gender or have hypothesized that males will like ads with sexual appeals more than females (e.g. Black, Organ, and Morton 2010; Jones, Stanaland, and Gelb 1998; LaTour 1990; Simpson, Horton, and Brown 1996; So 1996). One contribu- tion that our meta-analysis makes is that we now have estimates for mean effect sizes for males and females for attitude towards that ad. Interestingly, there was no difference between males and females on attitude towards the brand or purchase intention, although both of those tests were conducted with significantly fewer effect sizes and should be interpreted cautiously.

INTERNATIONAL JOURNAL OF ADVERTISING 187

As we began to consider this pattern of results – males liking ads with sexual appeals more than females but finding no difference between males and females on brand atti- tude or purchase intention – we were drawn back to the possibility that the evolutionary histories of males and females may provide an explanation for these results. For example, Saad (2004) argues that males like ads with sexual appeals, because they frequently fea- ture physically attractive females who often also signal sexual availability by their poses or actions. In contrast, females dislike ads with sexual appeals, because they interpret the female models (alone or in couples) as intra-sex mating competition (Kyrousi, Panigyrakis, and Panopoulos 2016). This would explain why attitude toward the ad would be different for males and females, and it is possible that the effect does not extend to brand attitude or purchase intention. While a number of researchers have recently considered how evo- lution may explain the effects of sexual appeals in ads (Lull and Bushman 2015; Wan, Luk, and Chow 2014), we consider this a potentially valuable tool for explaining seemingly con- founding patterns of results.

Moderator analyses and heterogeneity of effects

There are two additional findings that we want to highlight briefly before discussing man- agerial implications and limitations of the meta-analysis. First, one of the most consistent results in the moderator analyses emerged from testing the effect of the gender of models as a moderator. Although the result of the test for between-group variance was only sig- nificant for purchase intention, there was a significant negative effect for male models on attitude towards the ad, brand attitude, and purchase intention. Because the number of effect sizes was so small, we could not test whether a strong negative reaction from male participants explained these results (Jones, Stanaland, and Gelb 1998). Nevertheless, we hope this result generates additional research in the future.

Second, we found significant dispersion in the effect sizes when we tested for hetero- geneity of effects within almost all of the subgroups, as well as for the main effects tests. Beyond the value of identifying variables that help us understand the complexity of a phe- nomenon, such as sexual appeals in ads, one of the goals for testing moderators in meta- analysis is to explain heterogeneity of effects (Borenstein et al. 2010). While there is no exact threshold for ‘too much’ heterogeneity, by convention I2 statistics greater than 75 are interpreted as high or significant dispersion (Higgins and Thompson 2002), and for many of our results we reported I2 statistics were greater than 75. One implication is that we should be cautious in interpreting results with these levels of I2 statistics (Higgins et al. 2003), something we have tried to do throughout the article. A second implication is that an unidentified moderator or moderators may exist that explains variance of effect sizes within the sample of studies (Higgins et al. 2003). This happened with several of our sub- group analyses (e.g. product–appeal congruity for ad recognition and recall; male models for attitude towards ad, purchase intention); however, it is clear that additional modera- tors should be examined in future meta-analyses.

Managerial implications

Even though we recommend additional empirical analyses of the results of sex in advertis- ing research, there are several findings here that should be useful for advertising

188 J. G. WIRTZ ET AL.

practitioners. First, if your goal is to create awareness of an ad or product, then sexual appeals offer an advantage, as long as the product that you are advertising ‘fits’ into audi- ence expectations (e.g. perfume, fashion accessories). This also suggests that ads with sex- ual appeals that ‘break the mould’ or go against expectations actually will be less effective, an effect that extends to purchase intention. If your goal is increasing brand awareness or reinforcing a brand position, then sexual appeals offer no advantage, but they also offer no disadvantage. Second, males clearly like ads with sexual appeals more than females. So, if your goal is getting males to like an ad or separating males and females into two groups, then using sexual appeals should accomplish that goal. Further, even though brands in ads with sexual appeals may be evaluated more negatively, any negative effect does not extend to purchase intention – except when products do not match audience expectations about using sexual appeals in ads (e.g. tool set, USB drive, breakfast cereal), although we suspect that this latter observation may not be necessary. Finally, when creating ads with sexual appeals, using females or male–female couples as models is likely more effective than using only a male or males as models. We do not want to over-interpret this pattern of results, but the most consistent finding was that ads with sexual appeals featuring only male models produced significantly more negative atti- tudes and lower purchase intentions.

Study limitations and suggestions for future research

There are several limitations that we want to acknowledge and discuss, as well as to sug- gest ways some areas of future research. First, our meta-analysis relies heavily on research published in peer-reviewed journals. Although random-effects models assume that a meta-analysis represents a sample from the total population of studies (Schmidt, Oh, and Hayes 2009), we cannot rule out the possibility that unpublished data-sets exist and that their inclusion in our study would change the results. Similarly, we identified a few studies that may have been relevant to our study but that did not provide data in a useable for- mat (Vohs, Sengupta, and Dahl 2014; Weller, Sibley, and Neuhaus 1982). Although we con- tacted authors and requested data, we had limited success with this effort. Second, we only tested the highest level of sexual appeal against the control or comparison condition. Given that slightly more than three-quarters of the studies only used two levels for sexual appeal, it is unlikely that our results would differ significantly by including the other condi- tions. However, we also cannot rule out the possibility that the effects of exposure to sex- ual appeals are curvilinear (Reichert 2002) and that moderate sexual appeals (e.g. partial nudity, implicit sexual behaviour) are more effective than strong sexual appeals (e.g. full nudity, explicit sexual behaviour). This is another area for researchers to consider in the future. Third, because we used the narrative reviews of the sex in advertising literature (e.g. Percy and Rossiter 1992; Reichert 2002; Reichert and Lambiase 2003) as a guide as we were designing our meta-analysis, we ended up including a very broad range of studies in our sample. While our motivation to apply statistical analyses to the types of studies pre- sented in those articles may have been laudable, a better approach may have been to start conceptually with attitudes and purchase intention and to drill down on those varia- bles more deeply. Further, the decision to include such a large range of studies in our sam- ple almost certainly accounts for some or much of the heterogeneity of effects that we reported, which is the final limitation that we will mention. As the amount of dispersion in

INTERNATIONAL JOURNAL OF ADVERTISING 189

effect sizes increases, the amount of confidence that we can have in interpreting our results decreases (Higgins et al. 2003). That does not mean that we should have no confi- dence or that we cannot draw any conclusions; rather, it suggests that further research is needed to establish whether the nature of sexual appeals (e.g. nudity, sexual innuendo, sexual embeds) means that there are multiple ‘smaller populations’ of studies or if there are important unidentified moderators that help to explain the effects of exposure to ads with sexual appeals on memory, attitude, and purchase intention.

Conclusion

Frequently descriptive reviews have presented discussions of the sex in advertising litera- ture in an ‘on the one hand’, ‘on the other hand’, format. As a result, isolated empirical observations could be used to support any argument, but collectively to support none. While our study does not offer a definitive answer to the question, ‘Does sex sell?’, the results presented in this meta-analysis move advertising research and practices a step closer to achieving the goal of understanding the complexity of the effects of exposure to sexual appeals in advertisements.

Notes

1. Lull and Bushman (2015) recently published a meta-analysis that examines the effects of expo- sure to violent and sexual media content on individuals, which included some of the same vari- ables in this meta-analysis. While there is some overlap, our meta-analysis focuses more specifically on advertising research and outcomes specific to advertising, such as ad recall and recognition and attitude towards the ad.

2. For simplicity sake, we will use the term ‘control condition’ throughout the rest of the article to refer to the non-treatment sexual appeal condition. Some authors use the term ‘comparison condition’, and others refer to an ad with a pastoral scene as a control ad and an ad with fully clothed models in non-sexual behaviours as a comparison condition.

3. A number of studies have tested whether TV programme–sexual appeal congruity or incongru- ity influences ad and brand recall and recognition (e.g. Bushman 2007; Furnham and Mainaud 2011; Parker and Furnham 2007). While conceptually the same, these studies have generally not tested product–appeal congruity, which is the focus of our moderator test. See Lull and Bushman (2015) for a review and discussion of the programme–appeal congruity research.

4. Each website differed in the years that papers and abstracts were available online. For the AEJMC website, the available years were 1997–2016. For the ACR website, available years were 1974–2016. For the AAA website, the years checked were 2007–2016. We also conducted a hand search of printed abstracts from AAA conferences between 1979 and 2006. A majority of the abstracts and conference papers that we located had been published subsequently in a journal or book chapter. Additional information about this part of the search is available from the first author.

5. A complete list of the publications providing data for the meta-analysis appears in the Referen- ces section of the paper. Studies in the meta-analysis are preceded by �.

6. We defined a publication as a published or unpublished manuscript reporting the results of one or more sex in advertising studies. We defined a study as an experiment using a unique sample and method and reported as such by the authors (e.g. Putrevu and Swimberghek 2013; Dahl et al. 2009). When authors used the same method and stimulus materials to collect data from multiple samples, we counted this as a single study (e.g. Liu et al. 2009; Yan et al. 2010).

7. The issue of whether or not to include data from the “sexual programme” conditions in the TV programme–appeal congruity studies was one that we considered carefully. Bushman (2007)

190 J. G. WIRTZ ET AL.

hypothesized a negative effect on memory for ads featured in programmes with sexual themes and content. Our position is that by including only data from the low-sex programmes, we pro- vide a “cleaner” test of the direct effect of sexual appeals on memory. However, in light of the findings presented in Lull and Bushman (2015), we wanted to eliminate the possibility that this decision would cause us to overstate or overestimate any significant effects. Therefore, we con- ducted a second set of analyses with data from the sexual and non-sexual TV programme con- ditions. Those results are summarized in Note 14.

8. We used univariate F values only when there were two levels of sexual appeal (e.g. appeal type: sexual, non-sexual). This allowed us to compare treatment and control conditions directly rather than collapsing three or more levels of sexual appeals into one effect size (e.g. level of nudity: demure, partial nudity, full nudity). This approach was also consistent with the way we collected other data.

9. Products were coded as congruent if they fit in any of the following categories: (a) cosmetics or makeup; (b) perfume or body fragrance; (c) oils and lotions (e.g. suntan lotion, body, massage oil); (d) undergarments (e.g. bras, underwear); (e) luxury clothing and accessories (e.g. designer jeans, jewellery); (f) alcohol or liquor (e.g. beer, whiskey); (g) cigarettes; or (h) products or brands associated with fashion or sex (e.g. American Apparel, Details, Vogue, Playboy). All other prod- ucts were coded as not congruent.

10. The Krippendorf’s alpha statistics for the coding for descriptive variables were last name first author 1.0; publication type 1.0; publication year 1.0; studies per publication 1.0; sample size 1.0; sample type 1.0; per cent of female participants .91; levels of sex appeal .87; products appearing in ad .73; model/s in treatment ad 1.0; type of control/comparison .84.

11. The Krippendorf’s alpha statistics for the data for the main effects and moderator analyses for the first 24 studies were number of relevant outcome variables .92; separate data for male and females 1.0; means and standard deviations (combined across dependent variables) .96; other data used to calculate Cohen’s d (combined) .76; product–appeal congruity .88; ad memory (recognition, aided recall, unaided recall) .52; brand memory (recognition, aided recall, unaided recall) .44.

12. Violating the assumption of independent effects can produce inflated estimates of population effect sizes and biased results of tests for heterogeneity of effects, which can occur when data provided by the same participants on two or more dependent variables are combined (Hedges and Olkin 1985 Scammacca, Roberts, and Stubing 2014). To avoid introducing statistical depen- dence into the data, we used the ‘shifting unit of analysis’ approach recommended by Cooper (2010), which entails conducting separate analyses for each dependent variable. In situations where studies reported data on two memory outcomes (e.g. ad recognition, ad recall; Samson 2016), only the data from ad or brand recall was included in the combined test, and then the ad or brand recognition data were included in the subgroup analysis.

13. Moderator analyses for participant gender, product–appeal congruity, and model type were conducted using only the data that were available within each subgroup. A second approach is to include all data and to treat any data not fitting in a category as an additional category (e.g. female participants, male participants, other). We chose the first approach, because it fits the assumptions of a mixed-effects model and because it provides a direct test between each sub- group of a moderator variable. See Pigott (2012) for additional discussion on this topic.

14. The weighted mean Cohen’s d for ad recognition and recall was calculated after adding data from the sexual programme conditions from the relevant studies. There were k = 23 effect sizes, and the result indicated a positive significant result for sexual appeals, d = .37 (95% CI [.15, .58]; z = 3.34.; p < .001; I2 = 81.83). The weighted mean Cohen’s d for brand recognition and recall was calculated after adding data from the sexual programme conditions from the relevant stud- ies. There were k = 33 effect sizes, and the result indicated no significant effect for sexual appeals, d = .16 (95% CI [¡.02, .33]; z = 1.78.; p = .08; I2 = 84.02). Additional details about this analysis can be obtained from the first author.

INTERNATIONAL JOURNAL OF ADVERTISING 191

Acknowledgements

The authors would like to thank the three anonymous reviewers for their constructive feedback on the original manuscript and revised versions of the manuscript. The first author would also like to thank Daniel O’Keefe for comments on a previous version of the manuscript and Michael Borenstein for his good humour and advice on addressing significant heterogeneity of effects. The length of time between the initial submission and the submission of the final revision was due to personal cir- cumstances rather than problems with the research project itself.

Disclosure statement

There are no potential conflicts of interest.

ORCID

John G. Wirtz http://orcid.org/0000-0001-7370-5851

References �Indicates that article provided data for the meta-analysis. Albarrac�ın, D., B.T. Johnson, M. Fishbein, and P.A. Muellerleile. 2001. Theories of reasoned action and

planned behavior as models of condom use: A meta-analysis. Psychological Bulletin 127: 142–161. �Alexander, M.W., and B. Judd Jr. 1978. Do nudes in ads enhance brand recall? Journal of Advertising

Research 18, no. 1: 47–50. Amyx, D., and K. Amyx. 2011. Sex and puffery in advertising: An absolutely sensational and sexually

provocative experiment. International Business & Management 2, no. 1: 1–10. Anderson, C.A. and B.J. Bushman. 1997. External validity of ‘trivial’ experiments: The case of labora-

tory aggression. Review of General Psychology 1: 19–41. �Aylesworth, A.B., R.C. Goodstein, and A. Kalra. 1999. Effect of archetypal embeds on feelings: An

indirect route to affecting attitudes? Journal of Advertising 23, no. 3: 73–81. �Belch, M.A., B.E. Holgerson, G.E. Belch, and J. Koppman. 1982. Psychophysiological and cognitive

responses to sex in advertising. In Advances in consumer research Vol. 9, ed. A. Mitchell, 424–7. Ann Arbor, MI: Association for Consumer Research.

�Bello, D.C., R.E. Pitts, and M.J. Etzel. 1983. The communication effects of controversial sexual content in television programs and commercials. Journal of Advertising 12, no. 3: 32–42.

�Black, I.R., G.C. Organ, and P. Morton. 2010. The effect of personality on response to sexual appeals. European Journal of Marketing 44: 1453–77.

�Black, I.R., and P. Morton. 2017. Appealing to men and women using sexual appeals in advertising: In the battle of the sexes, is a truce possible? Journal of Marketing Communications 23: 1–20.

Boddewyn, J.J., and H. Kunz. 1991. Sex and decency issues in advertising: General and international dimensions. Business Horizons 34, no. 5: 13–20.

Borenstein, M., L.V. Hedges, J.P. Higgins, and H.R. Rothstein. 2010. A basic introduction to fix- ed�effect and random�effects models for meta�analysis. Research Synthesis Methods 1, no. 2: 97–111.

Brown, S.P., and D.M. Stayman. 1992. Antecedents and consequences of attitude toward the ad: A meta-analysis. Journal of Consumer Research 19, no. 1: 34–51.

�Bushman, B.J. 2007. That was a great commercial, but what were they selling? Effects of violence and sex on memory for products in television commercials. Journal of Applied Social Psychology 37: 1784–96.

�Caccavale, J.G., T.C. Wanty III, and J.A. Edell. 1982. Subliminal implants in advertisements: An experi- ment. In Advances in consumer research, Vol. 9, ed. A. Mitchell, 418–423. Ann Arbor, MI: Associa- tion for Consumer Research.

192 J. G. WIRTZ ET AL.

�Chang, C.T. and C.H. Tseng. 2013. Can sex sell bread? The impacts of sexual appeal type, product type and sensation seeking. International Journal of Advertising 32: 559–85.

Chestnut, R.W., C.C. LaChance, and A. Lubitz. 1977. The “decorative” female model: Sexual stimuli and the recognition of advertisements. Journal of Advertising 6, no. 4: 11–14.

Cheung, M.-C., A.S. Chan, Y.M. Han, S.L. Sze, and N.H. Fan. 2013. Differential effects of Chinese wom- en’s sexual self-schema on responses to sex appeal in advertising. Journal of Promotion Manage- ment 19: 373–91.

Christy, T.P., and E. Haley. 2008. The influence of advertising context on perceptions of offense. Journal of Marketing Communications 14: 271–91.

Cohan, J.A. 2001. Towards a new paradigm in the ethics of women’s advertising. Journal of Business Ethics 33, no. 4: 323–37.

Cohn, L.D., and B.J. Becker. 2003. How meta-analysis increases statistical power. Psychological Meth- ods 8, no. 3: 243–53.

Comprehensive meta-analysis (Version 2) [Computer software]. 2014. Englewood, NJ: Biostat. http:// www.comprehensive.com

Cooper, H.M. 2010. Research synthesis and meta-analysis: A step-by-step approach. 4th ed. Washing- ton, DC: Sage.

Courtney, A.E., and T.W. Whipple. 1983. Sex stereotyping in advertising. Lexington, MA: Lexington Books.

�Cui, G., and X. Yang. 2009. Responses of Chinese consumers to sex appeals in international advertis- ing: A test of congruency theory. Journal of Global Marketing 22, no. 3: 229–45.

�Dahl, D.W., J. Sengupta, and K.D. Vohs. 2009. Sex in advertising: Gender differences and the role of relationship commitment. Journal of Consumer Research 36: 215–31.

�Das, E., M. Galekh, and C. Vonkeman. 2015. Is sexy better than funny? Disentangling the persuasive effects of pleasure and arousal across sex and humour appeals. International Journal of Advertising 34: 406–20.

�Davis, R.H., and J.A. Welsch. 1983. A new viewpoint on nudes in advertising and brand recall. Inter- national Journal of Advertising 2, no. 2: 141–6.

De Barnier, V., and P. Valette-Florence. 2006. Provocative sexually appealing advertisements: The influence of embarrassment on attitude towards the ad. In Asia-Pacific advances in consumer research Vol. 7, eds. M.C. Lees, T. Davis, and G. Gregory, 215–222. Sydney: Association for Con- sumer Research.

�De Pelsmacker, and M. Geuens. 1996. The communication effects of warmth, eroticism and humor in alcohol advertisements. Journal of Marketing Communications 2, no. 4: 247–62.

�Dianoux, C., and Z. Linhart. 2010. The effectiveness of female nudity in advertising in three Euro- pean countries. International Marketing Review 27: 562–78.

�Dudley, S.C. 1999. Consumer attitudes toward nudity in advertising. Journal of Marketing Theory & Practice 7, no. 4: 89–96.

Du Plessis, E. 2005. The advertised mind. Sterling, VA: MillwardBrown. Eagly, A.H., R.D. Ashmore, M.G. Makhijani, and L.C. Longo. 1991. What is beautiful is good, but...: A

meta-analytic review of research on the physical attractiveness stereotype. Psychological Bulletin 110: 109–28.

Eagly, A.H., and S. Chaiken. 1993. The psychology of attitudes. Belmont, CA: Wadsworth/Thomson Learning.

England, P., and T. Gardner. 1983. Sex differentiation in magazine advertisements: A content analysis using log-linear modeling. Current Issues & Research in Advertising 6: 253–68.

�Ferguson, C.J., A.M. Cruz, D. Martinez, S.M. Rueda, and D.E. Ferguson. 2010. Violence and sex as advertising strategies in television commercials. European Psychologist 15: 304–11.

Frith, K., S. Ping, and C. Hong. 2005. The construction of beauty: A cross-cultural analysis of women’s magazine advertising. Journal of Communication 55: 56–70.

�Furnham, A., and M. Hiranandani. 2009. Sexual attitudes and erotophobia and the recall of sexual content on television. Communications: The European Journal of Communication Research 34, no. 1: 73–86.

INTERNATIONAL JOURNAL OF ADVERTISING 193

�Furnham, A., and L. Mainaud. 2011. The effect of French television sexual program content on the recall of sexual and nonsexual advertisements. Journal of Sex Research 48: 590–98.

Furukawa, T.A., C. Barbui, A. Cipriani, P. Brambilla, and N. Watanabe. 2006. Imputing missing standard deviations in meta-analyses can provide accurate results. Journal of Clinical Epidemiology 59: 7– 10.

�Garcia, E., and K.C. Yang. 2006. Consumer responses to sexual appeals in cross-cultural advertise- ments. Journal of International Consumer Marketing 19, no. 2: 29–52.

Gill, R. 2009. Beyond the ‘sexualization of culture’ thesis: An intersectional analysis of ‘sixpacks’, ‘mid- riffs’ and ‘hot lesbians’ in advertising. Sexualities 12, no. 2: 137–60.

�Grazer, W.F., and G. Kessling. 1995. The effect of print advertising’s use of sexual themes on brand recall and purchase intention: A product specific investigation of male responses. Journal of Applied Business Research 11, no. 3: 47–57.

Hayes, A.F., and K. Krippendorff. 2007. Answering the call for a standard reliability measure for cod- ing data. Communication Methods and Measures 1: 77–89.

Hedges, L.V., and I. Olkin. 1985. Statistical methods for meta-analysis. Orlando, FL: Academic Press. Higgins, J.P.T., and S.G. Thompson. 2002. Quantifying heterogeneity in a meta-analysis. Statistics in

Medicine 21: 1539–58. Higgins, J.P.T., S.G. Thompson, J.J. Deeks, and D.G. Altman 2003. Measuring inconsistency in meta-

analyses. British Journal of Medicine 327: 557–60. �Horton, R.L., L. Lieb, and M. Hewitt. 1982. The effects of nudity, suggestiveness, and attractiveness

on product class and name recall. In Developments in marketing science, ed. V. Kothari, 456–9. Nacogdoches, TX: Academy of Marketing Science.

Huang, M. 2004. Romantic love and sex: Their relationship and impacts on ad attitudes. Psychology & Marketing 21, no. 1: 53–73.

Huang, Y., and D.T. Lowry. 2012. An analysis of nudity in Chinese magazine advertising: Examining gender, racial and brand differences. Sex Roles 66: 440–52.

�Hyllegard, K., J. Ogle, and R. Yan. 2009. The impact of advertising message strategy – fair labour v. sexual appeal – upon Gen Y consumers’ intent to patronize an apparel retailer. Journal of Fashion Marketing & Management 13: 109–27.

�Ismail, A.R., and T.C. Melewar. 2014. Attitude of Muslim consumers toward sex appeal in advertising: A comparative study between subcultures in Malaysia. Journal of Promotion Management 20: 553–70.

�Jones, M.Y., A.J.S. Stanaland, and B.D. Gelb. 1998. Beefcake and cheesecake: Insights for advertisers. Journal of Advertising 27, no. 2: 33–51.

Judd, B. Jr., and M.W. Alexander. 1983. On the reduced effectiveness of some sexually suggestive ads. Journal of the Academic of Marketing Science 11: 156–168.

�Kelly, J.S. 1979. Subliminal embeds in print advertising: A challenge to advertising ethics. Journal of Advertising 8, no. 3: 20–4.

�Ketelaar, P.E., D. Anschutz, and S. van Hemmen. 2014. Sex in advertising: Do women care anymore? Journal of Euromarketing 23: 17–35.

�Kilbourne, W.E., S. Painton, and D. Ridley. 1985. The effect of sexual embedding on responses to magazine advertisements. Journal of Advertising 14, no. 2: 48–56.

�King, J., A. McClelland, and A. Furnham. 2015. Sex really does sell: The recall of sexual and non-sex- ual television advertisements in sexual and non-sexual programmes. Applied Cognitive Psychology 29: 210–6.

Kraaykamp, G. 2002. Trends and countertrends in sexual permissiveness: Three decades of attitude changes in the Netherlands 1965–1995. Journal of Marriage and Family 64, no. 1: 225–39.

Kyrousi, A.G., G.G. Panigyrakis, and A.P. Panopoulos. 2016. Attitudes toward ads portraying women in decorative roles and female competition: An evolutionary psychology perspective. International Journal of Advertising 35: 771–98.

Lang, A. 2000 The limited capacity model of mediated message processing. Journal of Communica- tion 50: 46–70.

�Lang, A., K. Wise, S. Lee, and X. Cai. 2003. The effects of sexual appeals on physiological, cognitive, emotional, and attitudinal responses for product and alcohol billboard advertising. In Sex in

194 J. G. WIRTZ ET AL.

advertising: Perspectives on the erotic appeal, ed. T. Reichert, J. Lambiase, 107–32. Mahwah, NJ: Lawrence Erlbaum.

Lass, P., and S. Hart. 2004. National cultures, values and lifestyles influencing consumers’ perception towards sexual imagery in alcohol advertising: An exploratory study in the U.K., Germany and Italy. Journal of Marketing Management 20: 607–23.

�LaTour, M.S. 1990. Female nudity in print advertising: An analysis of gender differences in arousal and ad response. Psychology & Marketing 7, no. 1: 65–81.

LaTour, M.S., and T.L. Henthorne. 1993. Female nudity: Attitudes towards the ad and the brand, and implications for advertising strategy. Journal of Consumer Marketing 10, no. 3: 25–32.

�LaTour, M.S., and T.L. Henthorne. 1994. Ethical judgments of sexual appeals in print advertising. Journal of Advertising 23, no. 3: 81–90.

LaTour, M., R. Pitts, and D.C. Snook-Luther. 1990. Female nudity, arousal, and ad response: An experi- mental investigation. Journal of Advertising 19, no. 4: 51–62.

�Lee, S.Y., and S. Lee. 2016. The influence of masculinity and femininity in response to sex appeal advertising. Journal of Global Fashion Marketing 7, no. 4: 266–77.

Leigh, J.H., G.M. Zinkhan, and V. Swaminathan. 2006. Dimensional relationships of recall and recogni- tion measures with selected cognitive and affective aspects of print ads. Journal of Advertising 35, no. 1: 105–22.

�Leka, J., A. McClelland, and A. Furnham. 2013. Memory for sexual and non-sexual television com- mercials as a function of context and viewer gender. Applied Cognitive Psychology 27: 584–92.

Lipsey, M.W., and D.B. Wilson. 2001. Practical meta-analysis. Thousand Oaks, CA: Sage. �Liu, F., H. Cheng, and J. Li. 2009. Consumer responses to sex appeal advertising: A cross-cultural

study. International Marketing Review 26: 501–20. �Liu, F., J. Li, and H. Cheng. 2006. Sex appeal advertising: Gender differences in Chinese consumers’

responses. Asia Pacific Journal of Marketing and Logistics 18, no. 1: 19–28. �Loroz, P.S., and J.G. Helgeson. 2013. Boomers and their babies: An exploratory study comparing psy-

chological profiles and advertising appeal effectiveness across two generations. The Journal of Marketing Theory and Practice 21, no. 3: 289–306.

Lull, R.B., and B.J. Bushman. 2015. Do sex and violence sell? A meta-analytic review of the effects of sexual and violent media and ad content on memory, attitudes, and buying intentions. Psycho- logical Bulletin 141: 1022–48.

�Lynn, A.L. 1995. The effects of female sexual images on persuasion. Ph.D. diss., University of Houston. �MacKay, N.J., and K. Covell. 1997. The impact of women in advertisements on attitudes toward

women. Sex Roles 36: 573–83. MacKenzie, S.B., R.J. Lutz, and G.E. Belch. 1986. The role of attitude toward the ad as a mediator of

advertising effectiveness: A test of competing explanations. Journal of Marketing Research 23, no. 2: 130–43.

Matt, G.E., and T.D. Cook. 1994. Threats to the validity of research synthesis. In The handbook of research synthesis, ed. H. Cooper, and L.V. Hedges, 503–20. New York: Russell Sage Foundation.

�Mittal, B., and W.M. Lassar. 2000. Sexual liberalism as a determinant of consumer response to sex in advertising. Journal of Business & Psychology 15: 111–27.

Morrison, B.J., and R.C. Sherman. 1972. Who responds to sex in advertising? Journal of Advertising Research 12, no. 2: 15–19.

Murnen, S.K., and M. Stockton. 1997. Gender and self-reported sexual arousal in response to sexual stimuli: A meta-analytic review. Sex Roles 37, no. 3/4: 135–53.

Nelson, M.R., and H.J. Paek. 2008. Nudity of female and male models in primetime TV advertising across seven countries. International Journal of Advertising 27: 715–44.

�O’Connor, P.J., A. Baher, B. Gong, and E. Kane. 1986. Recall levels of sexuality in advertising. In AMA educators’ proceedings, ed. T.A. Shimp, S. Sharma, J.G. John, A. Quelch, J.H. Lindgren, W. Dillon, M. P. Gardner, and R.F. Dyer, 2–5. Chicago, IL: American Marketing Association.

Oliver, M.B., and J.S. Hyde. 1993. Gender differences in sexuality: A meta-analysis. Psychological Bulle- tin 114: 29–51.

Overton, R.C. 1998. A comparison of fixed-effects and mixed (random-effects) models for meta-anal- ysis tests of moderator variable effects. Psychological Methods 3: 354–79.

INTERNATIONAL JOURNAL OF ADVERTISING 195

�Pan, P. 2014. Toward an integrated model of purchase intention of dietary supplements in sexually oriented advertising. Journal of Food Products Marketing 20, no. 2: 132–45.

�Parker, E., and A. Furnham. 2007. Does sex sell? The effect of sexual programme content on the recall of sexual and non�sexual advertisements. Applied Cognitive Psychology 21: 1217–28.

�Patzer, G.L. 1980. A comparison of advertisement effects: Sexy female communicator vs non-sexy female communicator. In Advances in consumer research, Vol. 7, ed. J.C. Olson, 359–64. San Fran- cisco, CA: Association for Consumer Research.

Percy, L., and J.R. Rossiter. 1992. Advertising stimulus effects: A review. Journal of Current Issues & Research in Advertising 14, no. 1: 75–90.

�Peterson, R.A., and R.A. Kerin. 1977. The female role in advertisements: Some experimental evi- dence. Journal of Marketing 41, no. 4: 59–63.

�Petrescu, M., P. Korgaonkar, and J. Gironda. 2016. Viral advertising: A field experiment on viral inten- tions and purchase intentions. Journal of Internet Commerce 14: 384–405.

Pigott, T.D. 2012. Advances in meta-analysis. New York: Springer. �Pope, N.K.L., K.E. Voges, and M.R. Brown. 2004. The effect of provocation in the form of mild erotica

on attitude to the ad and corporate image. Journal of Advertising 33, no. 1: 69–82. �Putrevu, S. 2008. Consumer responses toward sexual and nonsexual appeals: The influence of

involvement, need for cognition, and gender. Journal of Advertising 37, no. 2: 57–70. �Putrevu, S., and K. Swimberghek. 2013. The influence of religiosity on consumer ethical judgments

and responses toward sexual appeals. Journal of Business Ethics 115: 351–65. Raudenbush, S.W. 1994. Random effects models. In The handbook of research synthesis, ed. H. Cooper,

and L.V. Hedges, 301–21. New York: Russell Sage Foundation. Reichert, T. 2002. Sex in advertising research: A review of content, effects, and functions of sexual

information in consumer advertising. Annual Review of Sex Research 13, no. 1: 241–73. Reichert, T. 2003. The erotic history of advertising. Amherst, NY: Prometheus Books. �Reichert, T., and E.M. Alvaro. 2001. The effects of sexual information on ad and brand processing

and recall. Southwestern Mass Communication Journal 17, no. 1: 9–17. Reichert, T., and C. Carpenter. 2004. An update on sex in magazine advertising: 1983 to 2003. Jour-

nalism & Mass Communication Quarterly 81: 823–37. Reichert, T., and I. Fosu. 2005. Women’s responses to sex in advertising: Examining the effect of

women’s sexual self-schema on responses to sexual content in commercials. Journal of Promotion Management 11, no. 2/3: 142–53.

�Reichert, T., S.E. Heckler, and S. Jackson. 2001. The effects of sexual social marketing appeals on cog- nitive processing and persuasion. Journal of Advertising 30, no. 1: 13–27.

Reichert, T., and J. Lambiase, eds. 2003. Sex in advertising: Perspectives on the erotic appeal. Mahwah, NJ: Lawrence Erlbaum Associates.

Reichert, T., J. Lambiase, S. Morgan, M. Carstarphen, and S. Zavoina. 1999. Cheesecake and beefcake: No matter how you slice it, sexual explicitness in advertising continues to increase. Journalism & Mass Communication Quarterly 76, no. 1: 7–20.

�Reichert, T., M.S. LaTour, and J.B. Ford. 2011. The naked truth: Revealing the affinity for graphic sex- ual appeals in advertising. Journal of Advertising Research 51: 436–48.

Reichert, T., M.S. LaTour, and J. Kim. 2007. Assessing the influence of gender and sexual self-schema on affective responses to sexual content in advertising. Journal of Current Issues & Research in Advertising 29, no. 2: 63–77.

Reichert, T., and K.M. Walker. 2005. Sex and magazine promotion: The effects of sexualized subscrip- tion cards on magazine attitudes, interest, and purchase intention. Journal of Promotion Manage- ment 11, no. 2/3: 131–41.

�Reichert, T., and S. Zhou. 2007. Consumer responses to sexual magazine covers on a men’s maga- zine. Journal of Promotion Management 13, no. 1/2: 127–44.

Reichert, T., K. Yoo, H. Choi, M.S. LaTour, and J.B. Ford. 2011. Nudity in fragrance advertising: Assess- ing the influence of individual constructs in response to a graphic sexual appeal. Paper presented at the annual meeting of the Association for Educators in Journalism and Mass Communication, August 10–13, in St. Louis, MO.

196 J. G. WIRTZ ET AL.

Reid, L.N., and L.C. Soley. 1981. Another look at the ‘decorative’ female model: The recognition of visual and verbal ad components. Current Issues & Research in Advertising 4, no. 1: 123–33.

Reidenbach, R.E., and K.W. McCleary. 1983. Advertising and male nudity: An experimental investiga- tion. Journal of the Academy of Marketing Science 11: 444–54.

Riffe, D., Lacy, S., & Fico, F. G. (2014). Analyzing media messages: Using quantitative content analysis in research. 3rd ed. New York: Routledge.

�Rosen, D.L., and S.N. Singh. 1992. An investigation of subliminal embed effect on multiple measures of advertising effectiveness. Psychology & Marketing 9: 157–73.

Rouner, D., M.D. Slater, and M. Domenech-Rodr�ıguez. 2003. Adolescent evaluation of gender role and sexual imagery in television advertisements. Journal of Broadcasting & Electronic Media 47: 453–4.

�Ruth, W.J., H.S. Mosatche, and A. Kramer. 1989. Freudian sexual symbolism: Theoretical considera- tions and an empirical test in advertising. Psychological Reports 64: 1131–9.

�Ruth, W.J. 1990. Effects of Freudian sexual symbolism in advertising on self-reported purchasing tendencies: A preliminary intrabrand analysis. Psychological Reports 67: 1207–10.

Saad, G. 2004. Applying evolutionary psychology in understanding the representation of women in advertisements. Psychology & Marketing 21: 593–612.

�Sabri, O., and C. Obermiller. 2012. Consumer perception of taboo in ads. Journal of Business Research 65: 869–73.

�Samson, L. 2016. The effectiveness of using sexual appeals in advertising: Memory for sexual and non-sexual visual content across genders. Media Psychology. Advance online publication. doi: 10.1027/1864-1105/a000194

Scammacca, N., G. Roberts, and K.K. Stuebing. 2014. Meta-analysis with complex research designs: Dealing with dependence from multiple measures and multiple group comparisons. Review of Educational Research 84: 328–64.

Schmidt, F.L., I.S. Oh, and T.L. Hayes. (2009). Fixed- versus random-effects models in meta-analysis: Model properties and an empirical comparison of differences in results. British Journal of Mathe- matical and Statistical Psychology 62: 97–128.

�Sciglimpaglia, D., M.A. Belch, and R.F. Cain Jr. 1978. Demographic and cognitive factors influencing viewers’ evaluations of “sexy” advertisements. In Advances in consumer research, Vol. 6, ed. W.L. Wilke, 62–65. Provo, UT: Association for Consumer Research.

�Sengupta, J., and D.W. Dahl. 2008. Gender-related reactions to gratuitous sex appeals in advertising. Journal of Consumer Psychology 18: 62–78.

�Severn, J., G.E. Belch, and M.A. Belch. 1990. The effects of sexual and non-sexual advertising appeals and information level on cognitive processing and communication effectiveness. Journal of Advertising 19, no. 1: 14–22.

Shadish, W.R., and R.B. Sweeney. 1991. Mediators and moderators in meta-analysis: There’s a reason we don’t let dodo birds tell us which psychotherapies should have prizes. Journal of Consulting and Clinical Psychology 59: 883–93.

Sherman, C., and P. Quester. 2005. The influence of product/nudity congruence on advertising effec- tiveness. Journal of Promotion Management 11, no. 2/3: 61–89.

�Simpson, P.M., S. Horton, and G. Brown. 1996. Male nudity in advertisements: A modified replication and extension of gender and product effects. Journal of the Academy of Marketing Science 24, no. 3: 257–62.

Singh, S.N., M.L. Rothschild, and G.A. Churchill. 1988. Recognition versus recall as measures of televi- sion commercial forgetting. Journal of Marketing Research 25: 72–80.

Sivulka, J. 2012. Soap, sex, and cigarettes: A cultural history of American advertising. 2nd ed. Boston, MA: Wadsworth/Cengage Learning.

�So, S.L.M. (1996) Does sex sell in a Chinese society? An empirical study. Brain & Language 3, no. 1/2: 79–94.

Soley, L., and G. Kurzbard. 1986. Sex in advertising: A comparison of 1964 and 1984 magazine adver- tisements. Journal of Advertising 15, no. 3: 46–64.

Soley, L.C., and L.N. Reid. 1988. Taking it off: Are models in magazine ads wearing less? Journalism Quarterly 65: 960–6.

INTERNATIONAL JOURNAL OF ADVERTISING 197

Sparks, J.V., and A. Lang. 2015. Mechanisms underlying the effects of sexy and humorous content in advertisements. Communication Monographs 82, no. 1: 134–62.

Spears, N., and S.N. Singh. 2004. Measuring attitude toward the brand and purchase intentions. Jour- nal of Current Issues & Research in Advertising 26, no. 2: 53–66.

�Steadman, M. 1969. How sexy illustrations affect brand recall. Journal of Advertising Research 9, no. 1: 15–19.

�Tai, H.S. (1999) Advertising ethics: The use of sexual appeal in Chinese advertising. Teaching Business Ethics 3, no. 1: 87–100.

Trappey, C. 1996. A meta-analysis of consumer choice and subliminal advertising. Psychology & Mar- keting 13: 517–30.

Tulving, E., and F.I.M. Craik, eds. 2000. The Oxford handbook of memory. New York: Oxford University Press.

Vohs, K.D., J. Sengupta, and D.W. Dahl. 2014. The price had better be right: Women’s reactions to sex- ual stimuli vary with market factors. Psychological Science 25, no. 1: 278–83.

�Walker, P., and A. Zaid. 2011. The effects of sexualization in advertisements. MA thesis. Hanover Univ.

�Wan, W.W., C.L. Luk, and C.W. Chow. 2014. Consumer responses to sexual advertising: The intersec- tion of modernization, evolution, and international marketing. Journal of International Business Studies 45: 751–82.

Webb, T.L., and P. Sheeran. 2006. Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin 132: 249–68.

�Weller, R.B., C. Roberts, and C. Neuhaus. 1979. A longitudinal study of the effect of erotic content upon advertising brand recall. Current Issues & Research in Advertising 2, no 1: 145–62.

Weller, RB., S.D. Sibley, and C. Neuhaus. 1982. Experimental results concerning the effect of the female model in television commercials on product and brand recall. Developments in Marketing Science 5: 468–72.

Wells, B.E., and J.M. Twenge. 2005. Changes in young people’s sexual behavior and attitudes, 1943– 1999: A cross-temporal meta-analysis. Review of General Psychology 9, no. 3: 249–61.

Wilson, R.D., and N.K. Moore. 1979. The role of sexually-oriented stimuli in advertising: Theory and lit- erature review. In Advances in consumer research, Vol 6, ed. W.L. Wilkie, 55–61. Ann Arbor, MI: Association for Consumer Research.

Wise, G.L., A.L. Ring, and J.P. Merenski. 1974. Reactions to sexy ads vary with age. Journal of Advertis- ing Research 14, no. 4: 11–16.

Wright, P.J. 2013. U.S. males and pornography, 1973–2010: Consumption, predictors, correlates. The Journal of Sex Research 50, no. 1: 60–71.

�Wyllie, J., J. Carlson, and P.J. Rosenberger III. 2015. Does sexual-stimuli intensity and sexual self- schema influence female consumers’ reactions toward sexualised advertising? An Australian per- spective. Australasian Marketing Journal 23: 188–95.

�Yan, R., J.P. Ogle, and K.H. Hyllegard. 2010. The impact of message appeal and message source on Gen Y consumers’ attitudes and purchase intentions toward American apparel. Journal of Market- ing Communications 16: 203–24.

�Ye, Y., and S. Zhou. 2007. Is it the content or the person? Examining sexual content in promotional announcements and sexual self-schema. Journal of Promotion Management 13, no. 1/2: 55–73.

198 J. G. WIRTZ ET AL.

Copyright of International Journal of Advertising is the property of Routledge 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.