‘Idealised’ female models, who are most commonly used in advertisements, often have extremely thin figures (U.K. body-size 4–6) with features not usually possessed by ordinary people, making them appear desirable (Dillavou, 2009). These models tend to have a Body Mass Index (BMI) of 16 (Firger, 2016) which is considerably lower than the ‘healthy’ range of 18.5 to 24.9 (NHS, 2019). Consequently, it has been argued and demonstrated, exposure to idealised models can negatively impact female consumers’ body image (Groesz et al., 2002), and, in extreme cases, can lead to eating disorders (Stice et al., 2001). Inevitably, there is a range of shapes that may be described as idealised, which can be between ‘ultra-thin’ to ‘healthily thin’ as defined by BMI. In this study, we use models that may be described as the latter.

Some brands have introduced more diversity by including average and plus-size models (Malmelin and Hakala, 2009; Zlotnick, 2020). These average and plus-size models are defined as ‘non-idealised’ because they are larger in size and more realistic in comparison to idealised models (Aagerup, 2011). While the main priority of advertisers is to create an effective advertisement, they are also morally responsible for the adverse effects of their advertisements on consumers (Lee and Jin, 2019).

This study aims to investigate whether the use of non-idealised models in advertisements can positively affect female consumer’s body image, which has societal importance (BPS, 2019), and potentially a positive influence on advertisement effectiveness, which has financial benefits for brands (Tellis, 2004). It extends the current literature in three ways: comparing thin and plus-size models, using social media vs. traditional (print, TV, ecommerce) advertisements, and using memory (as a key measure of advertising effectiveness) in addition to brand attitude as outcome variables. Many studies that have attempted to measure the effectiveness of advertising have looked at memory, often thought of as a more robust measure of advertisement effectiveness than attitudes to the brand, or intention to buy (Furnham, 2019).

There is an extensive literature on model body size effects in advertising (Aagerup, 2021; Czerniawski, 2021; Janssen and Paas, 2014; Joo and Wu, 2021; Kordrostami and Laczniak, 2022; Paek et al., 2023; Selensky and Carels, 2021; Vazquez et al., 2024; Viglia et al., 2023; Yim et al., 2024). They tend to show that advertisers choose thinner models with very few using non-idealised, particularly plus-sized models. Researchers have been particularly interested in individual difference factors (such as personality, body size and body appreciation/satisfaction) and how they influence the evaluation of the model and the desire to purchase the product. Various explanations have been proposed such as that of Borau and Bonnefon (2017) who argued that previous inconsistent findings on the effects of non-idealised body shapes might have resulted from a complex causal framework. They suggested that “natural” models have an impact on performance through two affective mediators (body anxiety, and “repulsion toward the model”), while allowing moderation by the viewer’s own body mass index (BMI). In another study Lou and Tse (2020) showed that, women assimilate to an average-sized model over a thin or plus-size model in advertising, which did not affect their subsequent body satisfaction. Moreover, women’s body mass index (BMI) moderated the mediating relationship between model size and purchase intentions via perceived similarity to the model.

Marketers tend to over-represent idealised models in their advertisements (Yu et al., 2011). Using idealised models in traditional advertisements (e.g., TV, magazines) causes body image concerns in 21% of women (Mental Health Foundation, 2021), possibly because individuals think they are undesirable if they do not possess this thin body (Mills et al., 2017). Research has shown that exposure to idealised models in advertising results in women reporting lower self-esteem (Richins, 1991), greater body dissatisfaction (Grabe et al., 2008), and a more negative body image (Groesz et al., 2002). However, the effect sizes in these studies were small or moderate.

These findings can partly be explained by social comparison theory (Festinger, 1954); individuals evaluate their characteristics against others if some objective measures are unavailable. When individuals make upward comparisons to those who are better than they are, (e.g., consumers evaluating their body size against idealised models), their self-esteem is reduced. This is consistent with a recent meta-analysis of 48 articles, which found significant negative effects of upward comparison on social media users’ body image, well-being, mental health and self-esteem (McComb et al., 2023).

However, not all women engage in social comparisons, and some women instead demonstrate a self-enhancement effect when exposed to idealised models (Myers and Biocca, 1992; Smeesters and Mandel, 2006). Self-enhancement is the motivation to maximise positive views of oneself (Giacomin and Jordan, 2017), therefore, by perceiving idealised models as aspirational, this motivates these women to achieve their ideal physique which positively influences their body image (Mussweiler and Strack, 2000).

Other studies have found that idealised models do not negatively influence consumers’ body image, but rather non-idealised models shift consumers’ body image perceptions in a positive direction. Both Halliwell et al. (2005) and Holmstrom (2004) found that viewing magazine advertisements containing idealised models did not significantly increase females’ body-focused anxiety or body dissatisfaction, but non-idealised models elicited a relief effect, decreasing females’ body-focused anxiety and body dissatisfaction. Females make downward social comparisons to non-idealised models who are larger than they are, leading them to feel better about their body shape. Diedrichs and Lee (2011) also found that females who internalise ideal beauty standards displayed a more positive body image after exposure to average-sized models compared to thin or no models. Yet it may also be that lateral comparisons are also very likely when comparing oneself to a non-idealised model, especially when the model is average-sized rather than plus-sized.

These findings suggest that using non-idealised models in place of idealised models within advertisements would benefit society by creating a more realistic beauty standard. However, research thus far has focused mainly on traditional advertising (e.g., magazines), and not more recent forms of advertising such as that found in social media, which this study aims to do.

In this study, we look at body appreciation, which is closely related to the concept of body satisfaction. As Linardon et al. (2022) note, body appreciation, defined as accepting, holding favourable attitudes towards, and respecting the body, is the most widely studied facet of positive body image.

Idealised models are traditionally used in advertisements because they possess a body type with good physical and social health which is perceived as desirable in Western cultures (Levy-Navarro, 2009), and “the function of advertising is to create desire” (Yakhlef, 1999, p. 137). Marketers assume that these models are more effective because female consumers tend to prefer advertisements containing idealised models (Hesse-Biber et al., 2006), and develop more favourable attitudes towards the brand (Henderson-King et al., 2001), which elicits higher purchase intentions for the product (Till and Busler, 2000). One question that has been answered (Borau and Bonnefon, 2017; Janssen and Paas, 2014; Lou and Tse, 2020) but worth replicating, with different stimuli, is whether non-idealised models can be more effective than idealised models.

Although the early 2000s promoted extreme thinness as the beauty standard, evidence shows that this is no longer deemed to be aspirational (Barry, 2014). An increasing number of female consumers have criticised the use of unrealistically thin models in advertisements (Halliwell and Dittmar, 2004), and have demanded more diversity of body shapes and sizes via their social media platforms (Pounders and Mabry-Flynn, 2019). Therefore, brands have started to include more average and plus-size models, such as Rihanna’s SavageXFenty collection, which changes the societal norms by broadening the beauty standard for women. Research has shown that advertisements do not need to contain extremely thin idealised models to be effective, and incorporating non-idealised models is more socially responsible (D’Alessandro and Chitty, 2011).

Many consumers, particularly millennials, are interested in how brands are meeting their social responsibilities (Kotler, 2011), and they tend to make purchases based on ethical grounds rather than aesthetic ones (Aagerup, 2011). For example, consumers may exhibit positive attitudes towards a brand that uses non-idealised models as the brand holds values, such as body inclusivity, which are congruent with their values (Ridgway, 2016). Barry (2014) also found that non-idealised models are perceived as honest, aspirational, socially responsible, and therefore more appealing, in contrast to idealised models who are perceived as dishonest and unrealistic, leading to a rejection of the ‘thin-ideal’.

There is a literature on non-idealised models and advertising effectiveness going back over a decade (Åkestam et al., 2017; Diedrichs and Lee, 2010, 2011). There is also some recent evidence to suggest that women have responded positively to the use of more realistic models after being solely exposed to thin models for so long (Pounders and Mabry-Flynn, 2019), and therefore, may develop more positive attitudes towards brands that use non-idealised models. It is hypothesised that:

Furthermore, there may be an interaction between the consumer’s body image and the model’s size on brand attitudes (Potter, 1986). Yu et al. (2011) found that brand attitudes were indirectly influenced by participants’ body image via their perceived similarity to either the idealised or the non-idealised models. Specifically, women with a more negative body image perceive non-idealised models as aspirational (Barry, 2014), and perceive idealised models as less aspirational, which can be attributed to the substantial gap between their actual and ideal-self (Malär et al., 2011). Also, women who are less body-conscious perceive idealised models as aspirational and appealing (Malär et al., 2011). Thus, we test the idea that the viewing of different models has an effect on the relationship between body appreciation and attitude to brands. It is hypothesized that:

For all analyses reported in this section, the significance level was set at the conventional p = 0.05.

The first hypothesis was partially supported, as an interaction was found between the type of model and the change in the BAS scores pre-exposure and post-exposure to the advertisements; non-idealised models elicited a greater change in BAS scores compared to idealised models. Participants exposed to non-idealised models showed an increase in their BAS scores, however, participants exposed to idealised models showed no change in their BAS scores, rather than the decrease that was predicted. These findings are congruent with Holmstrom (2004) and Halliwell et al. (2005) who found that idealised models have no effect on consumers’ body image, but non-idealised models improve women’s body image. The positive effect of non-idealised models can be explained in terms of social comparison theory (Festinger, 1954). When making downward social comparisons to others with inferior attributes, individuals feel better about their attributes. However, we should acknowledge that we did not explicitly measure any comparison process—which could be seen as a limitation.

The current findings contradict studies in which it was found that idealised models elicit greater body dissatisfaction (Grabe et al., 2008) and a more negative body image (Groesz et al., 2002). This was possibly because some participants were inspired by the idealised models and demonstrated a self-enhancement effect, as in the study by Mussweiler and Strack (2000). Alternatively, the idealised models in this study may not have represented extremely thin real-life models as they were digitally created by slimming down the non-idealised versions. In the free recall task, some participants responded with “thin pear-shaped lady,” “normal women,” “average-sized woman,” indicating that the idealised models were not perceived as extremely thin. Possibly participants believed there was not much discrepancy between their body-size and the idealised models, therefore, appreciation of their own bodies was unaffected.

The second hypothesis was supported; the non-idealised models elicited more positive attitudes towards the brands compared to the idealised models. This may be because brands using non-idealised models are seen as being more honest, socially responsible, and appealing than those using idealised models (Barry, 2014), and consumers have recently demanded more diverse body-sizes in advertisements (Pounders and Mabry-Flynn, 2019). It is possible that participants believed in the importance of body inclusivity and therefore demonstrated more positive brand attitudes in the non-idealised condition, as research shows that consumers exhibit positive attitudes toward ethical brands which hold values that are congruent with their own (Ridgway, 2016). However, there are likely to be individual differences in body inclusivity beliefs, therefore, future research should measure these beliefs as it may moderate consumers’ brand attitudes towards advertisements using different sized models. Currently, these ideas however remain as essentially little more than untested speculations.

The finding above suggests that brands may benefit from using non-idealised models in their advertisements, in contrast to earlier findings which suggest that consumers demonstrate more positive attitudes towards brands using idealised models (for example, Henderson-King et al., 2001; Hesse-Biber et al., 2006; Till and Busler, 2000). The current findings also contradict those obtained by Halliwell and Dittmar (2004) who suggested that model size does not influence brand attitudes. This inconsistency may be due to their use of deodorant as the product category in their advertisements which is unrelated to body-size, in contrast to the use of clothing in the current investigation. When consumers purchase a practical product (e.g., a deodorant) to fulfil their needs, the model in the advertisement might be irrelevant (Aagerup, 2011). However, when consumers purchase clothing for aesthetic purposes the model’s size contributes to the appearance of the item (Kozar and Damhorsti, 2008). Therefore, future research could explore the use of different product categories, and investigate whether this modulates the influence of different sized models on brand attitudes.

One unanticipated finding was an association between body appreciation and brand attitude. While no other study has directly measured this relationship, a possible explanation could be that body appreciation is significantly associated with life satisfaction (Swami and Ng, 2015), and life satisfaction is positively correlated with brand behaviour, including brand awareness (Abbas et al., 2016).

The third hypothesis was that there would be an interaction between consumers’ pre-exposure BAS score and the type of model viewed, on brand attitude, and this was only partially supported. Participants viewing non-idealised models did not show a negative relationship between body appreciation and brand attitude. These findings are consistent with those of Yu et al. (2011), who found that women with a more negative body image did not demonstrate significantly higher brand attitudes after viewing non-idealised models compared to women with a more positive body image. However, as predicted, this investigation found that participants viewing idealised models showed a significant positive relationship between body appreciation and brand attitude. This finding is congruent with that of Malär et al. (2011) who found that less self-conscious individuals are more likely to perceive idealised models as aspirational as they are similar to them. Therefore, the current finding could be explained in terms of perceived similarity; participants with higher BAS scores are probably more likely to relate to the idealised models resulting in higher brand attitudes, as Kelly et al. (2014) found that individuals give positive evaluations to others that are similar to them.

The fourth hypothesis was that there would be an effect of model type on free recall, and this was supported; brands and products were recalled better by participants exposed to advertisements containing non-idealised models compared to idealised models. The most plausible explanation is that non-idealised models are unexpected, as they violate social conventions in fashion (Dahl et al., 2003), which grabs the consumers’ attention (Wu and Huberman, 2008), resulting in better memory, as there is a clear link between attention and memory (Chun and Turk-Browne, 2007). Furthermore, the pre-exposure BAS scores were not correlated with free recall suggesting consumers’ body image does not influence their memorability for advertisements containing an idealised or non-idealised model.

While a positive correlation was found between free recall and recognition scores in both the idealised and non-idealised conditions, the fifth hypothesis that there would be an effect of model type on recognition memory performance was not supported. No previous study has directly measured this for brand recognition specifically, however, this finding somewhat contradicts those obtained by Clayton et al. (2017) and Ridgway (2016) who found that plus-size models were recognised more than average-size models, followed by thin models, using a yes/no recognition task. It is not entirely clear why there is a difference in the findings, but it could be that model size affects model recognition, but not brand recognition. Alternatively, the discrepancy could be attributed to the type of recognition memory task used (yes/no vs. multiple-choice task) as they rely on slightly different mechanisms (see Smith and Duncan, 2004).

An explanation for the effect of model type on free recall but not recognition could be due to individual differences in focus of attention (i.e., on the models vs. product or brand); the recognition task was solely for brands, while free recall responses were mostly of the products. Individuals pay more attention to plus-size models than thin models (Clayton et al., 2017). Therefore, most participants probably focused heavily on the non-idealised models which increased their free recall of the products (clothing), but this distracted them from the brand names, resulting in poorer brand recognition. However, other participants may have been less distracted by the non-idealised models so may have also focused on the brand names resulting in better free recall and brand recognition scores. The high and low recognition scores of participants in the non-idealised condition, due to individual differences in their focus of attention, could have counteracted each other leading to a non-significant effect of model type on recognition. Future investigations could use eye-tracking to measure participants’ focus of attention as this may influence the distinction between brand and product memorability.

Furthermore, the negative correlation found between brand attitude and recognition scores in the non-idealised condition could also be explained by the focus of attention. The unconventional nature of non-idealised models may have drawn participants’ attention to this aspect of the advertisement resulting in higher brand attitude scores, at the expense of other information—such as the brand names—resulting in lower recognition scores. However, for the advertisements with idealised models, this did not occur, possibly because these models are conventional and draw less attention.