Abstract
Older people—particularly women—continue to be underrepresented in advertising. Some companies, however, have recently begun featuring more older women as decorative models in their advertisements. That practice goes against earlier research dating back to the late 1980s, which recommended that advertisers should avoid featuring older women in advertisements. The purpose of this article is to examine consumer responses to advertising featuring older female decorative models. In four experiments, the effects of advertisements of mostly female-gendered products, featuring older and younger female decorative models, are compared. The findings indicate positive effects of advertisements featuring older female decorative models for both younger and older female consumers, and that these effects can be explained by increased social connectedness with the decorative models.
- Older people in advertising
- older women in advertising
- underrepresentation, decorative models
- social effects of advertising
MANAGEMENT SLANT
The findings in this article suggest that advertisers could benefit from considering older women as decorative models in advertisements.
Advertisements featuring older female decorative models have positive effects on attitudes toward the advertisements in both younger and older female consumers.
For male consumers, responses do not differ between advertisements featuring older and younger female decorative models.
Advertisements featuring older female decorative models also have positive effects on social connectedness, brand attitudes, and purchase intentions.
INTRODUCTION
Most decorative models in advertisements are considerably younger than the average consumer (Phillips, 2022; Zhang, Harwood, Williams, Ylänne-McEwen, et al., 2006). A 2019 analysis by AARP Research showed that, although 46 percent of U.S. adults are ages 50 years or older, only 15 percent of the online media images include people of this age (Thayer and Skufca, 2019). This underrepresentation of older people is even more pronounced for older women, as images of young women continue to dominate advertising (Lewis, Medvedev, and Seponski, 2011; Shinoda, Veludo-de-Oliveira, and Pereira, 2021). Older women, in fact, are more underrepresented in advertising than older men (Baumann and de Laat, 2012; Eisend, 2022; Gurrieri, 2021; Phillips, 2022). Despite this continued underrepresentation, some model agencies have recently started to see a higher demand for their older models, and more advertisers now feature older female decorative models in their advertisements than before (Jefferson, 2018). Still, no recent research has empirically examined the specific effects of using older women as decorative models in advertisements on consumer responses to the advertising in a consistent way. Earlier research did, however, recommend that advertisers should avoid featuring older women in their advertising. One study (Greco, 1988), for example, argued that featuring older people in advertising risks the alienation of other consumer groups. Another (Saad, 2004) specifically recommended that advertisers avoid using older women as decorative models, arguing that the predominance of young, attractive women as decorative models in advertising represented an effective advertising strategy. Although these recommendations from research are some decades old, they have not yet been examined empirically in more recent research in a consistent way. In this article, the authors therefore examine the commercial and social advertising effects of advertisements featuring older female decorative models.
Although there is a lack of research about the specific effects of challenging the underrepresentation of older people in advertising, the positive effects of featuring members of other underrepresented groups in advertising, such as homosexual couples (Eisend, 2019; Eisend and Hermann, 2019; Northey, Dolan, Etheridge, et al., 2020), have been documented. The authors of the current research propose that these positive effects will also extend to advertisements featuring older women as decorative models. The purpose of this article is to empirically examine the effects of advertisements that challenge the underrepresentation of older women in advertising. To this end, the authors compared consumer responses to advertisements featuring older and younger female decorative models. Not only has previous research demonstrated that older women are still underrepresented in advertising (Baumann and de Laat, 2012; Eisend, 2022; Phillips, 2022; Shinoda et al., 2021), but some scholars have pointed out a lack of research on the commercial and social advertising effects of featuring older people in advertising, particularly older women (Eisend, 2022; Phillips, 2022). As a theoretical explanation for the effects of advertisements featuring older female decorative models, the current authors examined the mediating role of social connectedness with the advertisement models, a known social effect of advertising. The authors of this article thus add to the corresponding literature by studying a social effect of advertising (Eisend, 2019; Kim, Hayes, Avant, and Reid, 2014) in the context of the underrepresentation of older people in advertising, as recently called for by Eisend (2022). The authors of this article also offer an empirical examination of the predictions made in some older research concerning the potential negative effects of using older women as decorative models in advertisements (Greco, 1988; Saad, 2004). The article contributes to advertising practice by providing those advertisers that currently feature (or want to feature) older female decorative models in their advertisements with guidance on what effects to expect and why. Research in this area is urgently needed, as some prominent advertisers have already started to feature older women as decorative models in their advertisements (Jefferson, 2018).
Older Women in Advertising
Most of the previous research about older people in advertising has focused on content analyses of advertisements, documenting the prevalence and portrayals of older people in advertising (Eisend, 2022; Zhang et al., 2006). These studies have documented a persisting underrepresentation of older people in advertising in both television (Baumann and de Laat, 2012; Simcock and Sudbury, 2006) and print (Lewis et al., 2011; Shinoda et al., 2021) advertisements. This continued underrepresentation is discouraging, as media underrepresentation can be harmful to the social groups affected by it (Baumann and de Laat, 2012). Among the reasons for the underrepresentation of older people in advertising are the beliefs that older people are both less attractive and less likely to adopt new products and technologies, combined with a predominance of young employees and ageism in the advertising industry (Eisend, 2022). Some advertisers are also cautious of using older models in advertising because they fear that this may alienate younger advertising audiences (Szmigin and Carrigan, 2000). As mentioned earlier, older women are even more underrepresented than older men in advertising. Phillips (2022) recently reviewed previous studies about the representation of older people in advertising and found that most reported a continued underrepresentation of older women. The reviewed studies also recommended an increased use of older female decorative models in advertising.
A recent review of studies on the advertising effects of featuring older people in advertising (Eisend, 2022) showed that there have only been 14 research articles that examined the commercial advertising effects and that no previous research has studied any social advertising effects. The authors’ own analysis of the seven research articles in the review that included effects on attitudes toward the advertisements (See Appendix A) showed that several of these were published in research fields other than advertising and that most are several decades old. Several of the studies also focused on the advertising effects of older endorsers (e.g., Kwon, Saluja, and Adaval, 2015; Skupin, Beldad, and Tempelman, 2020). The previous research on the advertising effects of decorative models has, however, long differentiated between endorsers and decorative models. Whereas decorative models are anonymous and mainly selected for their physical appearance and characteristics, endorsers (or spokespersons, as they are sometimes also called) have a more active role in endorsing products and messages and thus act as stronger cues that can have a greater impact on message acceptance and persuasion (Bristol, 1996). Other studies in the review are less relevant because they focused on advertising elements other than photos of decorative models, such as illustrations of older people (e.g., Day and Stafford, 1997; Nelson and Smith, 1988). The studies that examined the effects of featuring photos of older
Challenging underrepresentation can have positive effects on how advertisements and brands are evaluated by consumers.
decorative models in advertising on advertisement attitudes indicated varying results that are difficult to generalize (Chevalier and Lichtlé, 2012; Rotfeld, Reid, and Wilcox, 1982). As Eisend (2022) notes, there is a lack of recent research about the advertising and social effects of older people in advertising, as the older studies were inconclusive and failed to establish any coherent theoretical explanations for the effects that they describe.
Indeed, when it comes to the representation of older women in advertising, some older research recommended that advertisers avoid featuring older female decorative models in advertisements (Greco, 1988; Saad, 2004). These recommendations and the continued underrepresentation of older women in advertising go against some recent advertising research indicating that challenging underrepresentation can have positive effects on how advertisements and brands are evaluated by consumers (Åkestam, Rosengren, and Dahlen, 2017; Eisend and Hermann, 2019; Northey et al., 2020). In line with these findings, the authors predicted that challenging the underrepresentation of older women in advertising will have positive effects on consumer responses to the advertisements. As a theoretical explanation for these effects, the authors offer social connectedness with the older decorative models featured in the advertisements.
Social Connectedness
Social connectedness can be defined as the extent to which a person feels connected to other people, including the people depicted in advertisements (Hutcherson, Seppala, and Gross, 2008). It is also one of several known social effects of advertising. A variety of social effects linked to advertising have been documented, including social connectedness (Liljedal, Berg, and Dahlen, 2020) and empathy (Åkestam et al., 2017; Escalas and Stern, 2003). Social effects have also been found to mediate the effects of advertisements featuring decorative models on more traditional advertising metrics. Social connectedness, for example, has been found to mediate the positive effects of advertisements challenging the underrepresentation of other groups on attitudes toward advertisements and brands. One study (Åkestam et al., 2017) demonstrated positive effects on social connectedness for advertisements featuring homosexual imagery, thus challenging the underrepresentation of homosexual couples in advertising. In line with this, the authors proposed that advertisements featuring older female decorative models will lead consumers to experience an increased social connectedness with the models featured in advertisements.
As an explanation for this positive effect, in line with previous research, the current authors proposed that advertisements featuring older female decorative models will challenge consumer expectations of advertising, indicating that the brand has taken the effort to feature atypical decorative models in their advertising. This propels consumers to think more about the decorative models featured in the advertisements. Previous research indicates that such an increase in social thoughts about people in advertisements can have positive effects on consumer perceptions of social connectedness with decorative models in advertisements (Åkestam et al., 2017; Liljedal et al., 2020). Most advertising, however, does not produce these effects, because it tends to portray what is considered typical in advertising (e.g., White, heterosexual, young people), thereby stripping such portrayals of meaning (Grier and Brumbaugh, 1999). Not dissimilar to the current development in brand activism, in which brands actively take a stand in sociopolitical issues through the use of advertising (Moorman, 2020; Vredenburg, Kapitan, Spry, and Kemper, 2020), atypical choices of decorative models and portrayals in advertising can lead to an increase in social thoughts in consumers, which, in turn, leads them to experience increased social connectedness with the portrayed people (Liljedal et al., 2020).
As older women are still underrepresented in advertising, the authors propose that advertisements featuring older women will be perceived as more atypical and lead consumers to have more social thoughts about the advertisements. Thus, the authors hypothesized that advertisements featuring older female decorative models will lead consumers to experience increased social connectedness with the models featured in the advertisements.
H1: Social connectedness with the decorative models featured in the advertisements will be higher for advertisements featuring older female decorative models than for advertisements featuring younger female decorative models.
Previous research has showed that social connectedness can mediate the positive advertisement effects of advertising that challenges the underrepresentation of other groups of people (Åkestam et al., 2017; Eisend and Hermann, 2019). In line with this previous research, the authors of the current article expected that the increased social connectedness with the decorative models featured in the advertisements (proposed in H1) will, in turn, lead to positive advertisement effects for advertisements featuring older female decorative models. More specifically, because of the increased feelings of social connectedness with the decorative models featured in the advertisements, the authors proposed that advertisements featuring older women will improve attitudes toward the advertisements, as people generally like and feel positive toward things with which they feel connected (Hutcherson et al., 2008; Jeong and Kim, 2021). Thus, the authors predicted that advertisements featuring older female decorative models will have positive effects on attitudes toward the advertisement, compared with advertisements featuring younger female decorative models.
H2: Attitudes toward advertisements featuring older female decorative models will be more positive than attitudes for advertisements featuring younger female decorative models.
The authors suggested that advertisements featuring older female decorative models will have a positive effect on social connectedness with the models featured in the advertisements, which, in turn, will have a positive effect on attitudes toward the advertisements. The authors further proposed that the positive effects of featuring older women in advertising on social connectedness with the decorative models in the advertisements (H1) will mediate the positive effects on attitudes for the advertisements (H2). It was hypothesized that:
H3: Social connectedness will mediate the positive effects of advertisements featuring older female decorative models (described in H2) on attitudes toward the advertisements.
The downstream effects of advertisements featuring older female decorative models on brand attitudes and purchase intentions were also examined in the current article. The authors proposed that the positive effects of advertisements featuring older female decorative models on attitudes for the advertisements will, in turn, lead to positive effects on brand attitudes and purchase intentions (MacKenzie, Lutz, and Belch, 1986). Positive effects of advertisements challenging the underrepresentation of other groups have previously been demonstrated on brand attitudes and purchase intentions (Åkestam et al., 2017; Bond and Farrell, 2020).
The current article also examined two potential moderators to the positive effects of advertisements featuring older female decorative models proposed in the hypotheses; namely, consumer gender and consumer age. As some previous research about consumer responses to decorative models has demonstrated different effects for women and men (Berg, 2015), the authors propose that consumer gender should be considered as a potential moderator for the positive effects of advertisements featuring older female decorative models. In line with calls for more research to examine the cross-gender effects of gender-targeted advertising (Åkestam, Rosengren, Dahlen, et al., 2021), the authors examine the potentially moderating effects of consumer gender. Another potential moderator for the positive effects of advertisements featuring older female decorative models is consumer age. As younger decorative models provide age-related social cues that can trigger an identity threat in older consumers (Amatulli, Peluso, Guido, and Yoon, 2018), it is possible that older women will respond more negatively to younger female decorative models. Younger consumers are also relatively immune to age-related social cues (Amatulli et al., 2018) and would likely not experience an identity threat from being exposed to older female decorative models. Negative responses to female decorative models in advertising have, however, been recorded in women of all ages (Bower, 2001; Dittmar and Howard, 2004). Considering these related (but contradictory) findings, the authors of the current research examine consumer age as a potential moderator.
OVERVIEW OF THE STUDIES
The research hypotheses are addressed in four studies, all designed as between-subjects experiments. The authors provide an overview of the purpose, design, and results of the empirical studies (See Table 1).
Overview of Studies
STUDY 1
Study 1 was designed as a between-subjects experiment to test H2, which proposed that advertisements featuring older female decorative models, and thus challenging the underrepresentation of older women in advertising, would lead to more positive attitudes toward the advertisements compared with advertisements featuring younger female decorative models.
Methodology
Stimuli
The advertisements featuring older and younger female decorative models were operationalized as photos of an older woman and a younger woman featured in two advertisements for a fashion brand (Kate Spade). Elderly women are still under-represented in fashion advertising (Lewis et al., 2011). The two advertisements used in Study 1 had previously been featured on the Kate Spade website, and each showed an anonymous female decorative model with a handbag from the brand in front of her. Apart from the ages of the models, the two advertisements were similar in terms of composition, brand logotype, background color, and product. Real advertisements were used to increase the realism and external validity of the study.
Participants and Procedure
Study 1 used a U.K. online convenience sample from Prolific (80 women, 18–78 years old, mean age = 35.73). Prolific is an online research platform often used to recruit participants for academic research (Peer, Brandimarte, Samat, and Acquisti, 2017). Participants completed an online questionnaire and were subsequently compensated with 1 British pound sterling (£1 GBP) each. Participants were randomly assigned to the two experimental groups, and each participant viewed and answered questions about one advertisement only. As recommended by Peer, Vosgerau, and Acquisti (2014), attention check questions were used to ensure active participation.
Measurements
The first question measured attitudes toward the advertisements as the dependent variable. As Study 1 used a real advertisement from an established brand, the authors also measured brand familiarity, advertisement recognition, and advertisement model recognition. At the end of the questionnaire, Study 1 included two manipulation control measurements about the age of the model to ensure that the manipulation was successful. The authors present all scale items and reliability measures (See Appendix B).
Analysis and Results
To address H2, a t test was used to compare attitudes for the two advertisements, showing more positive attitudes toward the advertisement featuring an older decorative model (Molder = 5.47, SD = 1.14; Myounger = 4.77, SD = 1.45), t(78) = 2.40, p < .05. H2 was thus supported.
Additional analyses showed that there was no difference in brand familiarity between the experiment groups (Molder = 2.75, SD = 1.55; Myounger = 2.71, SD = 1.73), t(78) = .13, p = .897; and that few participants claimed to recognize either the advertisements (0 percent) or the models (1.3 percent). As expected, participants guessed the age of the older model to be higher than the age of the younger model (Molder = 70.16, SD = 7.19; Myounger = 22.67, SD = 2.44), older younger
Regardless of gender, all consumers may purchase gendered products, either for themselves or as gifts, or they may act as advisors to other consumers.
t(78) = 40.36, p < .001. Similarly, they rated the older model to be older on the age measurement scale (Molder = 5.62, SD = .82; Myounger = 1.81, SD = .72), t(78) = 22.23, p < .001.
Discussion
The results of Study 1 showed that attitudes toward the advertisement were more positive for the advertisement featuring an older female decorative model than for the advertisement featuring a younger female decorative model. The manipulation control questions confirmed that the decorative model in this advertisement was perceived as older.
STUDY 2
Study 2 was designed to test H2 using a between-subjects experiment and an online sample of both male and female consumers. The purpose of the study was to replicate the findings from Study 1 and to examine whether consumer gender moderates the positive effects of advertisements featuring older female decorative models. The study also used a thought protocol.
Methodology
Stimuli
Study 2 used the same two advertisements for a fashion brand (Kate Spade) previously used in Study 1. The advertisements were not specifically targeted to men but are, nevertheless, relevant to this consumer group. Consumers of all genders are exposed to most advertising, and the current study is concerned with consumer responses to the advertisements rather than responses to the advertised products. Further, although advertisements for handbags may be primarily targeted to female consumers, they are still relevant to consumers of all genders. Regardless of gender, all consumers may purchase gendered products, either for themselves or as gifts, or they may act as advisors to other consumers (Åkestam et al., 2021).
Participants and Procedure
Study 2 used an online U.S. convenience sample of 120 participants from MTurk (56 women, 64 men, aged 24–73 years, mean age = 40.28). Participants were each compensated with US$1 and randomly assigned to experiment groups, so that each participant viewed and answered an online questionnaire about one of the advertisements. All participants had an MTurk approval rate of at least 98 percent, and attention check questions were used.
Measurements
The same measurement as in Study 1 was first used for attitudes toward the advertisements. A thought protocol was then used to capture consumer responses to the advertisements and decorative models more fully. In the discussion leading up to H1, it was suggested that the positive effects of advertisements featuring older female decorative models on social connectedness may stem from an increase in social thoughts. This aspect was captured by the thought protocol in Study 2. For the thought protocol, participants were asked to write down as many thoughts as they liked, directly after having viewed the stimuli advertisement (Åkestam et al., 2017; Liljedal et al., 2020). Both authors then coded the resulting thoughts independently. Finally, Study 2 also included the same manipulation control questions used in Study 1 about model age, model age measured on a scale, brand familiarity, advertisement recognition, and model recognition. The authors present all scale items, reliability measures, and interrater agreement metrics (See Appendix B).
Analysis and Results
A two-way analysis of variance (ANOVA) was performed to examine the effects of model age (older model versus younger model) and consumer gender (male versus female) on advertisement attitudes. The overall test was significant, F(3, 116) = 3.04, p < .05, η2 = .073; and so was the main effect of model age, F(1, 116) = 10.02, p <.05, η2 = .032 (Molder = 5.47, SD = 1.65; Myounger = 4.93, older SD = 1.65), offering support for H2. The main effect of consumer gender (p = .340) was not significant, but the interaction effect between model age and consumer gender was significant, F(1, 116) = 12.96, p < .05, η2 = .041. An analysis of contrasts was used to compare attitudes toward the advertisements among men and women, revealing more positive attitudes in female consumers for the advertisement with an older female decorative model (Molder = 5.96, SE = .31; Myounger = 4.73, SE = .31), F(1, 116) = 8.17, p < .005. For male consumers, however, there was no difference in attitudes (Molder = 5.02, SE = .29; Myounger = 5.10, SE = .28), F(1,116) = .04, p = .845. H2 was thus supported for female consumers but not for male consumers.
An additional analysis of the thought protocol showed significantly more social thoughts related to the older model compared with the younger model (Molder = 1.01, SD = 1.18; Myounger = .68, SD = .83), t(151) = 2.09, p < .05. In further investigation of the thought protocols for men and women, respectively, the data showed that there were more age-related thoughts for the advertisement featuring an older female decorative model among both male consumers (Molder = .65, SD = .75; Myounger = .05, SD = .21), t(84) = 5.07, p < .001; and female consumers (Molder = .95, SD = .85; Myounger = .17, SD = .45), t(75) = 4.87, p < .001. These thoughts were generally positive for the older model (e.g., “I like there being an older lady in the [advertisement]”), compared with the more neutral age thoughts about the younger model (e.g., “young girl”), and there was no discernable difference between consumer genders for these comments. Finally, only participants who saw the advertisement featuring an older female decorative model had thoughts relating to the representation of older women in advertisements, and this held true for both for male consumers (Molder = .30, SD = .56; Myounger = .00, SD = .000), t(84) = 3.55, p < .001; and female consumers (Molder = .26, SD = .59; Myounger = .00, SD = .000), t(75) = 2.64, p < .01. Examples included “I like that they thought an older woman was ‘worthy’ for their [advertisement]” and “it’s not the age you usually see in these [advertisements].” As previously, there were no discernable differences between consumer genders in the comments.
Additional analyses indicated that only a small number of the participants claimed to recognize the advertisements (0.8 percent) or the models (6.7 percent). The guessed identities of the models varied but were all incorrect. The older model was judged to be older in years (Molder = 71.49, SD = 6.02; Myounger = 22.25, SD = 3.52), t(118) = 54.93, p < .001. A two-way ANOVA examining the effects of model age and consumer gender on perceived model age in years, F(3, 116) = 1,003.08, p < .001, η2 = .963; showed a significant main effect of model age, F(3, 116) = 3,001.33, p < .001, η2 = .963; but not for consumer gender (p = .608), and the interaction effect between model age and consumer gender was not significant (p = .246). The older model was also rated as older on the age scale (Molder = 6.19, SD = .74; Myounger = 1.57, SD = .81), t(118) = 32.64, p < .001. A two-way ANOVA examining the effects of model age and consumer gender on perceived model age rated on a scale, F(3, 116) = 361.82, p < .001, η2 = .903; showed a significant main effect of model age, F(3, 116) = 1,084.85, p < .001, η2 = .903; but no significant main effect for consumer gender (p = .795) or interaction effect between model age and consumer gender (p = .056). There was no difference in the familiarity with the brand between the participants who viewed the two advertisements (Molder = 4.17, SD = 1.88; Myounger = 4.07, SD = 1.75), t(118) = .30, p = .767. A two-way ANOVA examining the effects of model age and consumer gender on familiarity with the brand, F(3, 116) = 2.49, p = .064, η2 = .061; showed a significant main effect of consumer gender, F(3, 116) = 7.14, p < .01, η2 = .058. The main effect of model age (p = .817) and the interaction effect between model age and consumer gender (p = .652) were not significant.
Discussion
Study 2 replicated the positive effects of the advertisement featuring an older female decorative model on attitudes toward the advertisement found in Study 1, this time with a U.S. sample. Study 2 also indicated that consumer gender moderated the positive effects of the advertisements featuring older female decorative models. The positive effects on attitudes toward advertisements featuring older female decorative models thus seem to be limited to female consumers. It should be noted that there were no negative effects for male consumers. Study 2 merely indicated that, for the male participants, there was no difference in attitudes toward the advertisements featuring older and younger female decorative models.
STUDY 3
Study 3 was designed as a between-subjects experiment to test H1, H2, and H3 using fictive skin-care advertisements. Additionally, Study 3 used a larger and stratified online sample of female consumers from the United Kingdom that comprised two equally sized groups of younger women (up to 25 years old) and older women (50 years old and over). This sampling technique was used to ensure that there were enough participants from both age groups in the sample to allow further examination of whether the effects of advertisements featuring older female decorative models differed between these age groups. As Study 1 and Study 2 used real advertisements, there were some minor differences between the advertisements apart from the age differences between the decorative models. The older model, for example, wore sunglasses, and the two models wore different clothes. To minimize such differences, Study 3 used fictive advertisements designed to be as similar as possible.
Methodology
Stimuli
Two fictive advertisements for an anonymous skincare brand were created by the authors for Study 3. Skin care was selected because beauty advertisements typically feature younger models (Lewis et al., 2011). Skin care is also a category in which a vast majority of both younger and older women purchase and use products regularly (Statista, 2017). Both stimulus advertisements displayed a photo of a woman, a product photo of a skin-care product, and a blurred brand. All photos were purchased from stock photo provider Shutterstock. The brand was blurred to avoid any confounding effects of previous exposure to a brand. The only difference between the two advertisements was in the photos of the female decorative models: one model was older, and one was younger. The photos of the two female decorative models were selected to be similar in terms of composition, pose, and facial expression (See Figure 1).
Stimuli Advertisements for Skin Care Used in Study 3 and Study 4 Showing an Advertisement Featuring an Older Female Decorative Model (Left) and an Advertisement Featuring a Younger Female Decorative Model (Right). (Mock advertisements with photos from Shutterstock.)
Pretests
To verify the suitability of these two advertisements, two pretests were conducted. The first was conducted among four advertising researchers who were asked to rate the similarity of the two photos of decorative models. Similarity was measured to avoid other potential confounding effects due to differences between the photos other than the ages of the models. The selected photos were deemed similar on the basis of the resulting mean value of similarity (M = 5.50).
The second pretest was carried out with a U.K. Prolific sample similar to (but different from) the sample used in Study 3 (60 women, ages 15–81 years, mean age = 41.07 years). All participants viewed both advertisements but in a randomized order. The measures included model attractiveness and the degree to which the advertisement was perceived as realistic, as well as the same measurements for model age as in Study 1 and Study 2. A paired-samples t test indicated that the older model was perceived to be older (Molder = 56.58, SD = 7.50; Myounger = 32.82, SD = 5.52), t(59) = 24.91, p < .001. There were no statistically significant differences between the older and younger decorative models in terms of how attractive the participants perceived them to be (Molder = 6.07, SD = .79; Myounger = 5.91, SD = .98), t(59) = 1.63, p = .108; or perceived realism between the two advertisements (Molder = 4.39, SD = 1.51; Myounger = 3.92, SD = 1.54), t(59) = 1.94, p = .057.
Finally, an open-ended question was included to assess responses not captured by the other questions: “Please consider the [advertisement]. What are your thoughts on it?” Comments indicate that participants thought about the ages of the models, especially for the older model. Participants made general observations about the younger model (e.g., “blue eyes and blonde hair, typical model, nice teeth, happy”), whereas comments on the older model most often mentioned her age (e.g., “older woman but very good skin still; minimal lines and good teeth”). As in Study 2, the thoughts were generally positive toward the older model (e.g., “the more mature lady is beautiful”) compared with the younger model (e.g., “another youthful looking lady with clear skin”). Comments relating to the older model also included many notes on the underrepresentation of older women in advertising (e.g., “I think it shows a representation of older women [that] is normally not seen in [advertisements]”).
Participants and Procedure
Study 3 was designed as a between-subjects experiment and used a U.K. online convenience sample of 157 female participants from Prolific. To enable a comparison between the responses of different age groups, a filter provided by Prolific was used to recruit a stratified sample that comprised two different age groups. The authors recruited 80 participants who were 50 years or older (ages 50–77 years, mean age = 61.20), and 77 participants who were 25 years or younger (ages 18–25 years, mean age = 21.79). The participants were each compensated with £1.25 GBP and randomly assigned to experiment groups so that each participant viewed and answered an online questionnaire about one of the advertisements.
Measurements
In Study 3, the same measurement as that in Study 1 and Study 2 was first used to measure attitudes toward the advertisements. Then, social connectedness with the people in the advertisements was measured. At the end of the questionnaire, the same manipulation control questions used in the previous studies were added for model age. Attention check questions were also used to ensure active participation (See Appendix B for all scale items and reliability measures).
Analysis and Results
A t test performed to address H2 showed more positive attitudes toward the advertisement featuring an older female decorative model (Molder = 4.96, SD = 1.35; Myounger = 4.17, SD = 1.45), t(155) = 3.55, p < .001). H2 was thus supported. To explore whether this effect might be influenced by consumer age, a twoway ANOVA was performed to examine the effects of model age (older model versus younger model) and consumer age (older consumer versus younger consumer) on attitudes toward the advertisements. The overall test was significant, F(3, 153) = 13.57, p < .001, η2 = .210; and so were the main effects of model age, F(1, 153) = 11.96, p < .001, η2 = .073; and consumer age, F(1, 153) = 25.91, p < .001, η2 = .145. The interaction effect between model age and consumer age was not significant (p = .606), indicating that consumer age did not moderate the effect. H2 was thus supported for both the older and the younger female consumers (See Figure 2).
Results for Attitudes toward the Advertisements, Study 3
In line with H1, a t test showed higher social connectedness for the advertisement featuring an older female decorative model (Molder = 3.94, SD = 1.62; Myounger = 3.00, SD = 1.19), t(143) = 4.17, p < .001; H1 was thus supported. To explore whether this effect might be contingent on consumer age, a two-way ANOVA was performed. The overall test was significant, F(3, 153) = 23.73, p < .001, η2 = .318; and so were the main effects of model age, F(1, 153) = 18.15, p < .001, η2 = .106; and consumer age, F(1, 153) = 33.75, p < .001, η2 = .181. The interaction effect between model age and consumer age was also significant, F(1, 153) = 14.70, p < .001, η2 = .088. A subsequent analysis of contrasts indicated that the older female consumers reported a stronger social connectedness with the older model, Molder = 4.82, SE = .19; Myounger = 3.20, SE = .21), F(1, 153) = 33.37, p < .001. For the younger female consumers, however, there was no significant difference in social connectedness between the younger and the older models (Molder = 2.90, SE = .21; Myounger = 2.81, SE = .20), F(1, 153) = .09, p = .765. Consumer age thus moderated the effects of the advertisement featuring an older female decorative model on social connectedness, so that only the older female consumers experienced higher social connectedness with the older model.
A mediation analysis was performed to address H3 and to examine the underlying theoretical explanation for the effects predicted for attitudes toward the advertisements in H2 (See Figure 3). More specifically, a moderated mediation analysis was performed using bootstrap analysis (Preacher and Hayes, 2008) in PROCESS, version 3.5, model 7, to test the mediation illustrated in Figure 3, where the effect of model age on social connectedness was moderated by consumer age. The analysis showed that the overall social connectedness outcome model was significant, F(3, 153) = 23.73, p < .001; as was the attitudes toward the advertisements outcome model, F(2, 154) = 59.57, p < .001. In the mediated
Consumers often experience social connectedness with advertising portrayals of people who are dissimilar to themselves in terms of, for example, gender and sexuality, indicating that social connectedness should be able to extend beyond externally visible sources for identification such as age.
moderation model, the direct effect of model age on advertisement attitudes was not significant (5,000 bootstrap samples, 95 percent confidence interval [95 percent CI]: −.57 to .15). The index of moderated mediation for this model was, however, significant (5,000 bootstrap samples, 95 percent CI: .48 to 1.42). The moderated mediation results indicated that the effects of model age were mediated by social connectedness in the older consumer group (mean effect: −.99; 5,000 bootstrap samples, 95 percent CI: −1.43 to −.59), but not in the younger consumer group (5,000 bootstrap samples, 95 percent CI: −.33 to .20). H3 was thus supported, but only for the older female consumer group.
Mediation Analysis Model, Study 3
Additional analyses showed that, as in the pretest, the older model was perceived to be older both in years (Molder = 56.59, SD = 7.71; Myounger = 34.19, SD = 6.95), t(154) = 19.06, p < .001; and on the age scale (Molder = 4.49, SD = .85; Myounger = 2.94, SD = .98), t(155) = 10.60, p < .001). A two-way ANOVA examining the effects of model age and consumer age on perceived model age in years showed a main effect of model age, F(3, 152) = 357.47, p < .001, η2 = .702; but neither the main effect of consumer age (p = .864) nor the interaction effect between model age and consumer age (p = .971) was significant. A similar two-way ANOVA examining the effects of model age and consumer age on model age measured on a scale (young–old) showed significant main effects of model age, F(3, 152) = 129.74, p < .001, η2 = .459; and consumer age, F(3, 152) = 16.79, p < .001, η2 = .099; but no significant interaction effect between model age and consumer age (p = .146).
Discussion
Study 3 replicated the findings from the previous studies regarding the more positive attitudes toward the advertisement featuring an older female decorative model for both younger and older female consumers. This was encouraging, as Study 3 used a new set of advertisements including new model photos, an anonymous brand, and a new product category (skin care). Furthermore, the results of Study 3 supported the theoretical reasoning in this article, because social connectedness was stronger for the advertisement featuring an older decorative model than for that featuring a younger decorative model. Study 3 also showed that social connectedness with the decorative model mediated the positive effects of the advertisement featuring an older female decorative model on attitudes toward the advertisement. The authors also explored another potential moderator for the positive effects on attitudes toward the advertisements: consumer age. The results from Study 3 indicated that consumer age did not moderate the effects of advertisements featuring older female decorative models on attitudes for the advertisements. Attitudes were more positive for the advertisements featuring an older female decorative model among both younger and older female consumers. Consumer age, however, did affect the degree to which the consumers experienced increased social connectedness with the decorative models in the advertisements.
The younger female consumers did not experience stronger social connectedness with the older decorative models. This was an unexpected result, as previous research indicates that social connectedness is not directly related to identification, in that consumers can experience higher social connectedness with decorative models in advertising with whom they do not necessarily identify. Consumers often experience social connectedness with advertising portrayals of people who are dissimilar to themselves in terms of, for example, gender and sexuality (Åkestam et al., 2017; Liljedal et al., 2020), indicating that social connectedness should be able to extend beyond externally visible sources for identification such as age. The results of the current study do not indicate any other explanation for the positive effects on attitudes toward the advertisements among the younger female consumers. Previous studies, however, have demonstrated negative responses among younger women toward young and attractive female decorative models in terms of negative emotions and social comparison (Bower, 2001; Dittmar and Howard, 2004).
STUDY 4
Study 4 was designed as a between-subjects experiment to test H1, H2, and H3 using an online sample of female consumers from the United Kingdom. In Study 4, the authors also measured brand attitudes and purchase intentions to enable an analysis of the downstream effects of advertisements featuring older female decorative models on these variables.
Methodology
Stimuli
Study 4 used four different advertisements. Two fictive advertisements for a new product, a nutritional supplement (omega-3 capsules), were created for this study to be used together with the two skin-care advertisements used in Study 3. The authors created the two omega-3 advertisements to be identical to the two skin-care advertisements, except that the product photo was changed from skin care to omega-3 supplements. The product photo was purchased from Shutterstock (See Figure 1 and Figure 4 for the advertisements used in Study 4).
Stimuli Advertisements for Omega-3 Supplement Used in Study 4 Showing an Advertisement Featuring an Older Female Decorative Model (Left) and an Advertisement Featuring a Younger Female Decorative Model (Right).
(Mock advertisements with photos from Shutterstock.)
Participants and Procedure
Study 4 used a U.K. online convenience sample from Prolific (201 women, aged 18–74 years, mean age = 37.24). Participants were each compensated with £1.25 GBP and randomly assigned to one of the four experimental groups so that each participant viewed and answered an online questionnaire about one of the advertisements only.
Measurements
The same measurement as in the previous studies was used to measure attitudes toward the advertisements and then was used to measure brand attitudes and purchase intentions. This was followed by the same measurement for social connectedness as that used in Study 3. At the end of the questionnaire, the same manipulation control questions as in the previous studies measuring model age were used, and attention check questions were included to ensure active participation (See Appendix B for all scale items and reliability measures).
Analysis and Results
A t test performed to address H2 revealed more positive attitudes for the advertisements featuring older female decorative models (Molder = 5.08, SD = 1.46; Myounger = 4.29, SD = 1.53), t(199) = 3.74, p < .001. H2 was thus supported. To explore whether this effect differed between product categories, a two-way ANOVA was performed to examine the effects of model age (older model versus younger model) and product category (skin-care product versus omega-3 supplement) on attitudes toward the advertisements. The overall ANOVA was significant, F(3, 197) = 5.96, p < .001, η2 = .083; and so was the main effect of model age, F(1, 197) = 13.64, p < .001, η2 = .065. The main effect of product category was, however, not significant (p = .098), and neither was the interaction effect between model age and product category (p = .304), indicating that product category did not moderate the effects of the advertisements featuring older female decorative models on attitudes for the advertisements.
As predicted in H1, a t test also showed higher social connectedness with the advertisements featuring older female decorative models (Molder = 3.74, SD = 1.39; Myounger = 3.14, SD = 1.27), t(199) = 3.18, p < .01. H1 was thus supported. To explore whether this effect might differ between product categories, a two-way ANOVA was performed to examine the effects of model age (older model versus younger model) and product category (skin-care product versus omega-3 supplement) on social connectedness. The overall ANOVA was significant, F(3, 197) = 4.51, p < .01, η2 = .064; and so was the main effect of model age, F(1, 197) = 9.91, p < .01, η2 = .048. The main effect of product category was not significant (p = .076); neither was the interaction effect between model age and product category (p = .679), indicating that the effect was not contingent on product category.
Additional analyses were also performed to examine the effects on brand attitudes and purchase intentions. A t test revealed more positive brand attitudes for the advertisements featuring older female decorative models (Molder = 4.87, SD = 1.33; Myounger = 4.37, SD = 1.28), t(199) = 2.75, p < .01. A two-way ANOVA was also performed to examine the effects of model age and product category on brand attitudes. The main test was significant, F(3, 197) = 4.81, p < .01, η2 = .068; and so were the main effects of model age, F(1, 197) = 7.31, p < .01, η2 = .036; and product category, F(1, 197) = 5.50, p < .05, η2 = .027; but the interaction effect between model age and product category was not significant (p = .259). Another t test showed stronger purchase intentions for the advertisements featuring older female decorative models (Molder = 4.04, SD = 1.73; Myounger = 3.55, SD = 1.73), t(199) = 2.01, p < .05. A two-way ANOVA was performed to examine the effects of model age and product category on purchase intentions. The main test was significant, F(3, 197) = 2.57, p < .05, η2 = .038; and so was the main effect of model age, F(1, 197) = 4.04, p < .05, η2 = .020. The main effect for product category (p = .070) and the interaction effect between model age and product category (p = .612) were not significant (See Figure 5).
Results for Dependent Variables, Study 4
To address H3 and to further examine the mediating effects of social connectedness on the positive effects for advertisements featuring older female decorative models on advertisement attitudes (H2), brand attitudes, and purchase intentions, a serial mediation analysis was performed using bootstrap analysis in PROCESS, version 3.5, model 6, to test the mediation model illustrated in Figure 6.
Mediation Analysis Model, Study 4
The analysis showed that the overall social connectedness out-come model was significant, F(1, 199) = 10.13, p < .01; as was the advertisement attitudes outcome model, F(2, 198) = 78.75, p < .001; the brand attitudes outcome model, F(3, 197) = 114.30, p < .001; and the purchase intentions outcome model, F(4, 196) = 89.00, p < .001. In the mediation model, the direct effect of model age on purchase intentions was not significant (5,000 bootstrap samples, 95 percent CI: −.09 to .52). The mediation analysis produced a significant mean indirect effect of −.118 (5,000 bootstrap samples, 95 percent CI: −.21 to −.04), indicating that the positive effects of the advertisements featuring older female decorative models on attitudes for the advertisements, mediated by social connectedness, in turn, mediated the positive effects on brand attitudes and purchase intentions. H3 was thus supported.
Additional analyses also showed that the older model was perceived to be older both in years (Molder = 56.70, SD = 6.15; Myounger = 34.55, SD = 6.59), t(199) = 24.62, p < .001; and on the age scale (Molder = 4.44, SD = .86; Myounger = 3.06, SD = .86), t(199) = 11.36, p < .001. A two-way ANOVA examining the effects of model age and product category on perceived model age in years showed a main effect of model age, F(3, 197) = 616.63, p < .001, η2 = .758; and product category, F(3, 197) = 4.88, p < .05, η2 = .024. The interaction effect between model age and product category (p = .700) was, however, not significant. A two-way ANOVA examining the effects of model age and product category on model age measured on a scale (young versus old) showed a significant main effect of model age, F(3, 197) = 127.64, p < .001, η2 = .393. The main effect of consumer age (p = .584) and the interaction effect between model age and product category (p = .499) were not significant.
Discussion
Study 4 replicated the positive effects of advertisements featuring older female decorative models on attitudes toward the advertisements in women from the three previous studies for yet another product category. The study also replicated the positive effects on social connectedness found in the third study. Social connectedness again mediated the positive effects of advertisements featuring older female decorative models on advertisement attitudes. Study 4 also showed that the positive effects were not limited to attitudes for the advertisements but that brand attitudes were more positive and purchase intentions were higher for the advertisements featuring older female decorative models.
GENERAL DISCUSSION
The results of the four empirical studies presented in this article indicated support for all the research hypotheses. The advertisements featuring older female decorative models had positive effects on attitudes toward the advertisements in female consumers across four experiments, using several different products and advertisements, as well as samples from both the United States and the United Kingdom. The mediation analyses indicated some support for the proposed theoretical explanation that social connectedness mediates the positive effects on attitudes toward the advertisements featuring older female decorative models. These findings are in line with previous research that has documented positive advertisement effects and social effects of challenging the underrepresentation of other groups in advertising (Åkestam et al., 2017, 2021; Bond and Farrell, 2020). The studies also showed that consumer gender moderated the positive effects of using advertisements featuring older female decorative models, as there were no positive effects from the advertisements featuring older female decorative models for male consumers, only for female consumers.
It should be noted that the authors did not find any negative effects from the advertisements featuring older female decorative models among male consumers. For the male consumers, there were no differences in attitudes for the advertisements featuring older and younger female decorative models. These findings could have been a result of gender differences in how advertising messages are processed. Male consumers tend to focus their processing of advertisements on the main message, for example, whereas female consumers tend to process contextual cues more extensively (Meyers-Levy and Loken, 2015). The gender differences in Study 2 may also have been a result of studying a gendered product and a female model, as both may have been perceived as less relevant by the male participants. The absence of negative effects, together with the positive effects found in female consumers (across age groups, products, and cultural contexts), makes a strong case for recommending advertisers to use advertisements featuring older female decorative models and is in line with the previous studies showing positive effects of challenging the underrepresentation of other groups (Bond and Farrell, 2020; Northey et al., 2020).
Theoretical Implications
This article will help advertisers and researchers better understand the effects of featuring older people in advertising. The current research adds an empirical examination of the earlier research recommendations to avoid featuring older women in advertising (Greco, 1988; Saad, 2004). Previous empirical research in the area has predominantly focused on documenting the representation of older people in advertising, mainly by conducting content analyses of advertising (Eisend, 2022; Phillips, 2022) and not by empirically examining consumer responses to older decorative models in advertising. The results of the studies presented in this article indicate that the older research recommendations for advertisers to avoid featuring older female decorative models in advertising (Greco, 1988; Saad, 2004) may simply be outdated. Rather than the negative effects of advertisements featuring older female decorative models proposed by this older research, the findings in this article indicate positive commercial and social advertisement effects from featuring older female models in advertisements. As mentioned earlier, these findings are in line with the recent research that demonstrates positive effects of advertisements featuring people from underrepresented groups (Åkestam et al., 2017; Bond and Farrell, 2020; Northey et al., 2020).
As a theoretical explanation of the effects of featuring older female decorative models in advertisements, the authors proposed studying social connectedness. The social effects of advertising are still understudied, but there is a growing interest in this research area (Eisend, 2019), and more research on the social effects of featuring older people in advertising is needed (Eisend, 2022). One such social effect in particular, social connectedness (Hutcherson et al., 2008), is receiving increased research attention, probably because of its ability to explain not only positive advertising and brand effects (Åkestam et al., 2017; Jeong and Kim, 2021; Liljedal et al., 2020) but also how audience connectedness can help create successful diversity initiatives (Burgess, Wilkie, and Dolan, 2020).
The findings in this article indicate positive commercial and social advertisement effects from featuring older female models in advertisements. These findings are in line with the recent research that demonstrates positive effects of advertisements featuring people from underrepresented groups.
Managerial Implications
The findings presented in this article offer support for the brands with recent advertising that features older female decorative models (Jefferson, 2018). The authors of the current article hope that the findings can, in some way, help counteract the persistent hesitancy in the advertising industry about the effectiveness of advertisements featuring older female decorative models (Windels, 2016) and thereby help lessen the underrepresentation of older women in advertising. On the topic of older women in advertising, some practitioners still question how effective advertisements featuring older decorative models will be for younger consumers (Szmigin and Carrigan, 2000). This is understandable, considering the old research recommendations for advertisers to avoid the use of older female decorative models (Greco, 1988; Saad, 2004) and the prevailing dominance of younger female decorative models in advertising (Gurrieri, 2021; Lewis et al., 2011). The results presented in this article should go some way in reducing any remaining hesitation among advertisers to feature older decorative models in their advertising. As the findings show, practitioners who feature older female decorative models in their advertising might, in fact, benefit considerably from doing so. Not only do advertisements featuring older female decorative models attract female consumers across ages, but they also manage to do so without alienating male consumers. Consumers of genders other than that targeted by an advertisement are always important to consider, as they may purchase the advertised products for themselves, purchase them as gifts, or act as advisors in shopping decisions (Åkestam et al., 2021). The lack of negative effects across consumer groups in the current research is particularly important, because it increases the practical relevance of the findings.
As many countries are currently seeing rapidly aging populations (Sudbury-Riley, Kohlbacher, and Hofmeister, 2015), the authors expect advertising professionals to become much more aware of the underrepresentation of older people in advertising moving forward. Aging populations mean that the elderly consumers will make up a larger share of the market, and advertising will have to reflect this development. Underrepresentation in media and advertising can be harmful for elderly consumers (Baumann and de Laat, 2012). Who gets represented in advertising matters because advertising is also very much a part of the cultural expressions of society, and representations in advertising not only reflect the current values of society but also affect how consumers view themselves and others (Amatulli et al., 2018; Bradley and Longino, 2001).
Limitations and Further Research
In the empirical studies presented in this article, the older women featured in the advertisements were all anonymous decorative models. Many older people featured in advertisements, however, are not decorative models but endorsers. The perceived age of endorsers featured in advertisements can affect advertisement responses and how the brand is perceived in terms of, for example, brand age (Huber, Meyer, Vogel, et al., 2013). Future research may want to examine whether the findings in this article can be applied to endorsers. Additionally, future research may want to examine potential mediators for the positive effects on attitudes for the advertisements that the current research demonstrated in younger women in Study 3. Although the younger women displayed more positive attitudes for the advertisements featuring older female decorative models, social connectedness did not mediate this effect in this consumer group (up to 25 years old). An examination of other potential mediators to this effect would be valuable.
An additional aspect of this research is that it exclusively focused on advertisements featuring older female decorative models. Consumer gender was found to moderate the positive effects of the advertisements featuring older female decorative models, because men did not respond more positively (or negatively) to the advertisements featuring older female decorative models. This naturally leads to the question of how consumers would respond to advertisements featuring older male decorative models. Overall, however, older men are more frequently represented in advertising than are older women, making the underrepresentation of this group less pronounced (Phillips, 2022). Future research may still want to explore how consumers of all genders respond to advertisements featuring older men. The current research also mostly focused on advertisements for female-gendered products (handbags), and future research may want to examine responses to more gender-neutral products advertised with older female decorative models.
Another question to consider for future research is the downstream consequences of the underrepresentation of older women. As the results in this article do not point to any negative effects of including older women in advertising, it seems likely that excluding them from advertising (Baumann and de Laat, 2012; Zhang et al., 2006) has no clear benefit. Instead, the consequences may include potentially serious effects in terms of, for example, older women being perceived as a less relevant market segment that will, therefore, have unmet needs. Future research may want to examine other social effects of challenging the underrepresentation of older women in advertising than social connectedness. One such social effect of interest would be empathy, as previous research on advertisements featuring underrepresented groups has demonstrated positive effects on empathy (Åkestam et al., 2017). Finally, although the current research replicated the positive effects of advertisements featuring older female decorative models across four studies with both U.S. and U.K. samples, there is still a need to explore the cultural aspects of the effects of these advertisements. The effects found in the four studies in this article could thus be further investigated in radically different cultures; for example, in countries with very small percentages of elderly consumers or cultures where ageism is less pronounced.
ABOUT THE AUTHORS
Hanna Berg is an associate professor at the Center for Consumer Marketing at Stockholm School of Economics, Sweden. Her main research interests are consumer behavior, visual marketing, advertising, and online retailing. Berg’s work has been published in the Journal of Advertising Research, European Journal of Marketing, Journal of Business Research, International Journal of Consumer Studies, and Psychology & Marketing, among other journals.
Karina T. Liljedal is an assistant professor at the Center for Consumer Marketing and a research fellow and lecturer at the Center for Retailing at Stockholm School of Economics, Sweden. Her research interests are consumer behavior, advertising, and retail. Liljedal’s work can be found in the Journal of Advertising Research, European Journal of Marketing, International Journal of Consumer Studies, and International Review of Retail, Distribution and Consumer Research, among others.
Appendix A
Overview of Previous Research about the Effects of Featuring Older People in Advertising on Attitudes toward the Advertisements
Appendix B Measures, Reliability, and Items
- Received September 3, 2021.
- Received (in revised form) October 14, 2022.
- Accepted October 18, 2022.
- Copyright © 2023 ARF. All rights reserved.
REFERENCES
Vol 64 Issue 3
