Veiled on Instagram? Representation of Veiled versus Nonveiled Women in Western and Egyptian Instagram Posts

Abstract

This study uses visual, quantitative content analysis to examine whether there is a significant difference in the representation of veiled versus nonveiled women in branded Instagram posts. We also tested the representation of women in Western and Egyptian branded Instagram posts. We used a sample of 200 branded Instagram posts created to promote products and services. The study tested five categories based on Goffman’s and then Kang’s gender analysis models: feminine touch, ritualization of subordination, licensed withdrawal, body display, and independence. We report a significant difference in body display between veiled women in Western versus Egyptian brands’ Instagram posts, as well as between veiled and nonveiled models in Western brands’ posts. We also found that both veiled and nonveiled women in Western brands’ posts show significantly more feminine touch compared to Egyptian brands.

Keywords:

• Context effects
• cross-cultural
• gender
• Instagram
• visual quantitative content analysis

The representation of women on Instagram and the perceptions of visual social media marketing content (Fox, Nakhata, and Deitz 2019) raise many questions about inferred stereotypes (Bo Li et al. 2021; Yang et al. 2021). This study investigated how Egyptian and Western brands portray veiled versus nonveiled women in posts on Instagram. The Egyptian brands typify Islamic and Arab cultures, while the Western brands represent U.S. and European cultures. We studied how advertising practitioners and content creators frame and portray veiled and nonveiled women on Instagram in a unique paired representation.

The impact of hijab on how women are perceived in the minds of brand managers and creative directors was important to understand because “advertising diffuses its meanings into society’s belief systems” (Kang 1997, p. 980). Advertisements naturalize symbols, making them appear to be inherent rather than constructed social devices promoting the ideologies of society (Umiker-Sebeok 1981, p. 209). Judith Williamson (1978) describes an ideological process or metastructure in advertising that gives meaning to products beyond their utility structures to become symbolic structures cocreated with the audiences (Jhally 1987, p. 130), drawing on their social knowledge.

For the past 50 years, the male-dominant advertising industry has been blamed for misrepresenting women, confining them to certain roles, and limiting their career potentials as convenient from the male perspective (Courtney and Lockeretz 1971; Sullivan and O’Connor 1988). Alarmingly, most recent studies still show similar findings (Marques 2021; Middleton et al. 2020). Even as female creative directors rise and practice image management, “women are ‘faking’ behaviors to both fit in and stand out within the strongly gendered environment of advertising” (Thompson-Whiteside, Turnbull, and Howe-Walsh 2021, p. 294).

Goffman’s (1979, p. 23) landmark research revealed how advertising staged gestures, rituals, and postures that mirrored and molded offensive social stereotypes of women. Goffman’s model for decoding representations deliberates on facial expressions, head postures, eyes, head–eye aversion, knees, hands, finger biting and touching, relative size, position, and placement. Goffman developed scales for measuring the portrayal of women in advertising in five categories: relative size (women’s size relative to men), feminine touch (women touching themselves or caressing objects), function ranking (occupation), the ritualization of subordination (knee bending, lying down), and licensed withdrawal (gazing away and being an accessory to the scene).

Kang (1997) analyzed the gender behavior portrayed in advertisements utilizing Goffman’s model of decoding behavior to observe changes in women’s images since 1979. She added two new categories: body display and independence or self-assertiveness. She found that “overall, the extent of sexism in magazine ads remained approximately the same from 1979 to 1991” (p. 988). Stereotyping, in fact, significantly increased in both body display and licensed withdrawal categories (p. 993). Kang recommended cross-cultural replications, and this research attempts that in two dimensions.

The Representation of Veiled versus Nonveiled Women in Advertising

The global Islamic clothing market is expected to exceed $88 billion in 2025. Its growth has attracted attention and significant investments in the sector from mainstream fashion industry players, including DKNY, Mango, Uniqlo, and Tommy Hilfiger (Grand View Research 2018). Westerners have always seen the veil as the ultimate symbol of gender oppression in Islamic cultures, yet many Islamic women are veiled voluntarily and claim it as a mark of resistance, agency, and cultural membership (Hirschmann 1997). Hijab is not only a clothing style but an ideology to be modest, both in character and appearance—not just physical modesty, but also modesty in one’s thoughts, speech, and actions (Abid 2020).

The idea is to remove focus from the physical aspects of a woman so that the focus may be on her mind, character, and spirituality rather than on her body and appearance. It sends a message that she is a Muslim, has respect for herself, and expects to be treated respectfully, especially by the opposite sex (Abdo 2008). Hijab is considered a liberating experience from societal expectations and judgments over a woman’s body and the bondage of the swinging pendulum of the fashion industry and other institutions that exploit females (Abdo 2008).

Veiled models and role models have been depicted playing sports while wearing the Nike Pro hijab (Nike Pro Hijab 2018). American Eagle advocated its new “denim hijab” through an Instagram post: “Thank you @americaneagle for encouraging young America to follow their passions, express their individuality, and pursue their unique paths” (Moneyish 2017). Adidas, Fenty Beauty, Gap, H&M, Uniqlo, Lane Bryant, and many other brands followed with hijab ads to draw the Muslim community’s attention toward their brands.

Women in the West are still not portrayed fairly in advertising (Poole, Page, and Sidlova 2021), but does that also apply to veiled models? We posit that nonveiled models will display significantly more of the five categories (feminine touch, the ritualization of subordination, licensed withdrawal, body display, and independence) than veiled women in Instagram posts (see Table 1, hypotheses 1 to 5).

Table 1. Results of analysis of variance for 2 × 2 design showing the differences between veiled and nonveiled women and between Egyptian and Western brand posts on Instagram.

Results of Two-Tailed t Test Comparing the Display of Each of the Five Variables for Veiled versus Nonveiled Models	Results of Two-Tailed t Test Comparing the Display of Each of the Five Variables for Western versus Egyptian Brands
H1: Nonveiled women display significantly more of feminine touch than veiled women in branded Instagram posts	H6: Women in Western brands’ Instagram posts display significantly more of feminine touch than in Egyptian brands’ Instagram posts
t (185) = 0.610, p = .543, Cohen’s d = 0.089	t (185) = 3.631, p = .000**, Cohen’s d = 0.531
H6(a): There is a significant difference in the feminine touch by veiled women on Instagram of Western brands than Egyptian brands.	H6(b): There is a significant difference in the feminine touch by nonveiled women on Instagram of Western brands than Egyptian brands.
t (92) = 2.406, p = .018*, Cohen’s d = 0.496	t (91) = 2.705, p = .008**, Cohen’s d = 0.560
H2: Nonveiled women display significantly more ritualization of subordination than veiled women in branded Instagram posts.	H7: Women in Western brands’ Instagram posts display significantly more ritualization of subordination than in Egyptian brands’ Instagram posts.
t (188) = 1.296, p = .197, Cohen’s d = 0.188	t (188) = 0.398, p = .691, Cohen’s d = 0.058
H3: Nonveiled women display significantly more licensed withdrawal than veiled women in branded Instagram posts.	H8: Women in Western brands’ Instagram posts display significantly more licensed withdrawal than in Egyptian brands’ Instagram posts.
t (198) = 0.515, p = .607, Cohen’s d = 0.073	t (198) = 1.360, p = .176, Cohen’s d = 0.192
H4: Nonveiled women display significantly more of body display than veiled women in branded Instagram posts.	H9: Women in Western brands’ Instagram posts display significantly more of body display than in Egyptian brands’ Instagram posts.
t (198) = 2.904, p = .004**, Cohen’s d = 0.411	t (198) = 2.336, p = .020*, Cohen’s d = 0.330
H4(a): There is a significant difference in the body display by veiled versus nonveiled women on Instagram of Western brands.	H9(a): There is a significant difference in the body display by veiled women on Instagram in Western brands than Egyptian brands.
t (98) = 2.769, p = .007**, Cohen’s d = 0.411	t (98) = 3.089, p = .003**, Cohen’s d = 0.618
H4(b): There is a significant difference in the body display by veiled versus nonveiled women on Instagram of Egyptian brands.	H9(b): There is a significant difference in the body display by nonveiled women in Western brands than Egyptian brands.
t (98) = 1.274, p = .206, Cohen’s d = 0.255	t (98) = 0.669, p = .505, Cohen’s d = 0.134
H5: Nonveiled women display significantly more independence than veiled women in branded Instagram posts.	H10: Women in Western brands’ Instagram posts display significantly more independence than in Egyptian brands’ Instagram posts.
t (198) = 0.163, p = .871, Cohen’s d = 0.023	t (198) = 1.254, p = .212, Cohen’s d = 0.177

*Significant p < .05; **significant p < .001.

The Representation of Women in Western versus Egyptian Ads

For four decades, the study of female stereotypes in academia has been ongoing (e.g., Goffman 1979; Kang 1997; Furnham and Mak 1999; Wolin 2003; Eisend 2010; Wirtz, Sparks, and Zimbres 2017). Middleton, Turnbull, and de Oliveira (2020) have provided an anthology of studies of female role stereotypes in Western advertising since 2000. They highlighted four key role stereotypes for women as housewives, sexual objects, decorative objects of beauty, and career-oriented professionals. Women are usually depicted as nurturing and loving but rather dependent. They are featured in ads promoting products that are supposed to decrease stress and make managing life easier. A study of gender representation in 1,300 prime-time TV ads found that women were underrepresented as a primary character in commercials. Yet women are usually the ones who purchase most of the products advertised (Chu, Lee, and Kim 2016).

As a response to the massive backlash to stereotypical ads and the rise of women in leadership positions in the advertising and creative industries (Hsu 2018), different businesses are taking this as an opportunity to rebrand their advertising messages to include words of empowerment to women and to promote gender equality (Varghese and Kumar 2020). This viral practice, known as “femvertising,” basically increases advertising scripts that empower women and decreases gender stereotypes and gender inequality with a celebratory tone focusing on women and their achievements (Champlin et al. 2019).

Despite being highly celebrated by many people, femvertising has its critics. Critics claim femvertising exploits female empowerment and gender equality by leveraging the language of liberation to sell products (Windels et al. 2020). Sterbenk et al. (2021) found that 81% of award-winning companies’ female empowerment/femvertising commercials were often not in line with their broader corporate social responsibility (CSR) activities. Also, The Guardian has negatively criticized femvertising, referring to it as a type of sexism (Iqbal 2015).

In Egypt, Arafa (2013) concluded that women in Egyptian advertising are generally “demoralized” and represented as sex objects, feeding into negative gender stereotypes discussed by Goffman and other scholars. However, she noted that a different image of Egyptian women is portrayed in food and home products advertisements targeting females. Both local and multinational brands represented women as conservative, caring housewives. Older, famous actresses were chosen to promote images of dedicated veiled Egyptian mothers (Arafa 2013).

In 2017, the FP7 advertising agency, which is part of the McCann World Group, created a campaign focused on the misconception around the veil and veiled women, correcting faulty stereotypical perceptions around the idea of hijab. The campaign, titled “Ala Rasi” or “On My Head,” gained attention on Egyptian social media, with more than 3 million views on Facebook. Recently, the Supreme Council for Media Regulation in Egypt banned several ads deemed offensive to women and general morals (Eyada 2018). There is evidence that there seems to be some response by brand managers, creative directors, and ad agencies, particularly when the ad is targeting women of higher socioeconomic class (El-Mohandes 2020)

We hypothesize that women in Western branded Instagram posts display significantly more of the five categories (feminine touch, the ritualization of subordination, licensed withdrawal, body display, and independence) than women in Egyptian branded Instagram posts (see Table 1, hypotheses 6 to 10).

Interactions of Brand Origin and Model Veil

To complete that analysis of variance, if any of our 10 hypotheses shows a significant difference, it would be necessary also to investigate the interactions contributing to this difference. Two possible interactions should be studied, and for any hypothesis HN we will call the possible interaction hypotheses N(a) and N(b). In all cases, we will assume that any of the five variables in Goffman–Kang categories will be significantly higher in Western brands compared to Egyptian brands and for nonveiled models compared to veiled models. For parsimony, only the potentially significant interactions are reported in Table 1.

Research Design and Methods

This study is based on Goffman’s (1979) gender analysis model that included six categories explaining how gender is represented in advertising. However, we followed Kang’s (1997) work, which developed Goffman’s categories for print ads. Kang excluded three categories of Goffman’s gender analysis (relative size, function ranking, and family) and added two categories, body display and independence, which matches the current research context. Table 2 describes the three categories explored by the researcher from Goffman’s model and the two categories adopted from Kang’s model.

Table 2. Description and coding schemes of variables used as guides for the coders.

Categories	Variables	Description
Feminine touch	1. Caressing objects 2. Touching self	Women depicted using their fingers and hands to trace the outlines of an object, cradle it, caress its surface, or effect “just barely touching”."
Ritualization of subordination	1. Female lowering herself 2. Lying/sitting on bed/sofa 3. Shy knee bend	Behaviors illustrating actions and acts of subordination, e.g., lowering oneself physically in some form or other of prostration
Licensed withdrawal	1. Covering mouth or face with hand 2. Head down or gaze aversion (looking down) 3. Expansive exaggerated smiling 4. Finger to mouth	Behaviors removing depicted women psychologically and physically from the social situation at large, leaving them unoriented in it and to it
Body display	1. Showing parts of the body 2. Face makeup 3. Nail polish 4. Seductive posture	Body-revealing clothes, excessive face makeup, nail polish, and exhibiting sexual behavior
Independence	1. Relying on/ contacting others in the picture 2. Shielding behind an object or a person	This category is based on Mee-Eun Kang to evaluate the depiction of women in terms of independence and self-assurance

Note. Based on Goffman (1979) and Kang (1997).

Unlike Goffman’s gender analysis, which compares the depiction of males and females in advertising, our study compares veiled to nonveiled women in the same campaign for the same brand in the same medium (Instagram) and by the same creative team. We apply the all-else-equal principle to isolate the effect using a veiled model versus a nonveiled model for shooting the post. We also tested the difference between Egyptian and Western branded posts.

Quantitative Visual Content Analysis

We used visual quantitative content analysis following Bock, Isermann, and Knieper (2011) to investigate whether replicating the shot with a veiled model caused the creative directors to portray veiled women differently, avoiding negative stereotypical representations of women in advertising. Fixing the procedures and interpretation rules led to objectivity, reliability, and validity of the quantitative visual content analysis (Rössler 2017). This method (in contrast to qualitative content analysis) can deal with large samples of visual data and allows for generalizations (Bock, Isermann, and Knieper (2011).

Sampling Procedures

Following Bock, Isermann, and Knieper’s (2011) steps, the main unit of analysis was Instagram images used by Egyptian and Western brands. Uribe and Manzur (2012) suggested that the most effective sample sizes for stratified samples are under 100 per group. Our sampling frame was the results of searches for two hashtags, #modest fashion and #hijab fashion. We inspected the branded posts resulting from the searches in order of appearance. We expanded the posts to look at the brand campaign. We considered 210 Egyptian and Western brand campaigns on Instagram until 100 pairs of posts met our criteria. Eligible branded posts were selected so that, in each case, the same brand post appeared twice for the same product, once with a veiled woman and the other time with a nonveiled woman. If the same brand had posts for more than one product category—for example, shirts, jeans, bags, or shoes—only the first eligible post was considered for each category. One hundred posts (50 pairs) represented Western brands, and 100 (50 pairs) represented Egyptian brands. From another perspective, 100 posts were of veiled women, and 100 were of nonveiled women. Table 3 explains the research design with sample posts.

Table 3. Matrix 2 × 2 design showing the distribution of Instagram images of veiled versus nonveiled women in Egyptian and Western brands, with an example of veiled versus nonveiled women photos from GAP as a Western brand and Ganubi as an Egyptian brand on Instagram.

Brand	Veiled (100)	Nonveiled (100)
Western (100 paired Instagram posts) Brand name GAP	n = 50 Instagram posts	n = 50 Instagram posts
Egyptian (100 paired Instagram posts) Brand name Ganubi	n = 50 Instagram posts	n = 50 Instagram posts

Coding Procedure

Each Instagram post was coded using the structure represented in Table 2. We recruited five coders with three to 10 years of marketing communication experience to analyze the brands’ posts. We developed a coding booklet, including the five categories and the scales to evaluate them. After training the coders, the sample was evenly divided, where each coder was asked to code 40 images (20 pairs); the first 10 pairs came from Egyptian brands, while the second 10 pairs came from Western brands on Instagram. The coders had to assign a score from 1 to 5 for each variable for each post. Each Instagram post was evaluated across the five categories of our model as presented in Table 2 to evaluate 15 different variables using 5-point scales (coding scheme). The score for any category was the average score of the variables it included.

It is important to mention that the fourth category, body display, was added but not developed beyond body-revealing clothes in Kang’s (1997, p. 987) study. The researchers developed all four aspects under this category (showing parts of the body, full face makeup, nail polish, seductive posture) to suit this study’s purpose. This research investigates veiled women, who usually do not show any parts of their body except for their hands and faces; therefore, we included face makeup, nail polish, and seductive posture as observable attributes.

Coders’ Training

All five coders received training by the researchers describing the scope of the research and its relation to gender analysis studies by Goffman, and all were given examples from previous research. The main research question was presented to the trainees using pictures and explanations from Hoffman’s original book, Gender Advertisements. We developed and shared an online coding booklet, including a description of all the variables and coding schemes, with the coders to ensure they had a holistic view of the research and coding requirements. A pilot of the coding sheet was conducted for the same 10 dummy pairs to test their understanding of the variables and detect any changes needed to the online form. Cohen’s kappa interrater reliability in the pilot was 63%, which we did not accept. We held another session to deliberate and explain the differences and why they may have happened. We concluded the session by evaluating another 10 dummy pairs, and Cohen’s kappa rose to an acceptable 86%. We then discussed the few differences among the five coders, and this concluded their training. The dummy pairs were not part of the study sample.

After coding the full sample, SPSS 24 was used to conduct an analysis of variance (ANOVA) of the 200 posts (veiled versus nonveiled models; Western versus Egyptian brands), followed by further analyses for interactions of significant differences.

Results

The only significant difference between the veiled and nonveiled women among the five categories was body display, hypothesis 4: t (198) = 2.904, p = .004. This category constitutes four attributes: showing parts of the body, facial makeup, nail polish, and seductive posture. The four other categories—feminine touch, ritualization of subordination, licensed withdrawal, and independence— did not differ significantly between veiled and nonveiled models in general. We further investigated the two possible components of body display variance. The first was hypothesis 4(a), the difference in body display by veiled versus nonveiled women on Instagram of Western brands, which turned out to be significant, t (98) = 2.769, p = .007. The other was hypothesis 4(b), the difference in body display by veiled versus nonveiled women on Instagram of Egyptian brands, which was not significant, t (98) = 1.274, p = .206.

Two of the five categories showed significant differences between Western and Egyptian brands: hypothesis 6’s feminine touch and hypothesis 9’s body display. Further analysis showed a significant difference in feminine touch by both veiled, t (92) = 2.406, p = .018*, and nonveiled, t (91) = 2.705, p = .008, models of Western brands compared to Egyptian brands. Women in Western brands’ Instagram posts showed significantly more body display compared to Egyptian brands, hypothesis 9, t (198) = 2.336, p = .020. Further analysis of the two components showed that this significant difference was attributed to hypothesis 9(a), veiled models, t (98) = 3.089, p = .003, rather than hypothesis 9(b), nonveiled models, t (98) = 0.669, p = .505.

Discussion

We shared the findings with two creative directors and a brand manager and integrated their insights into the discussion.

The Differences between Veiled and Nonveiled Women

The ANOVA showed that four categories of Goffman’s gender analysis model (feminine touch, ritualization of subordination, licensed withdrawal, and independence) were not significantly different between veiled or nonveiled models. This result contradicts the literature, where the general belief is that there are significant differences in the representation of veiled versus nonveiled women in advertising and media.

Our results showed significantly more (p > .01) body display of nonveiled versus veiled models in Instagram posts as predicted. But the surprising finding was that the significant difference in body display between veiled and nonveiled was attributed to Western brands only. Egyptian brands showed no difference in body display mean scores between veiled and nonveiled models, and both were high and close to that of Western nonveiled models.

But why were all other hypotheses rejected? “Our global marketing department sends us the shots with European models, and we have to reshoot them as close as possible with Egyptian models, whether veiled or nonveiled or both,” explained a female multinational footwear brand manager. Two phenomena are in action here. The feminist forces of femvertising push nonveiled women to resist the sex object and decorative display stereotypes (Windels et al. 2020). On the other hand, the veiled creative director we interviewed explained that “there is the determination by veiled women to show that they are equally active and proud to display their feelings and success. Both phenomena bring veiled and nonveiled women closer together toward a new norm of women refusing stereotypes either way.” This, in a way, explains the insignificance of the differences in the four other categories—feminine touch, ritualization of subordination, licensed withdrawal, and independence—between veiled and nonveiled models.

The Differences between Egyptian and Western Brands

Comparing Western and Egyptian branded posts showed significant differences in feminine touch and body display. Middleton, Turnbull, and de Oliveira (2020) explain the differences in body display. In Western cultures, people tend to evaluate and judge others’ identities based on personality traits. Western countries consider the body to be part of identity, where body esteem is nonseparate from self-esteem. In Western cultures with individualistic philosophies, women’s physical attractiveness is assessed based on very strict concepts promoting self and body dissatisfaction (Chu, Lee, and Kim 2016), unlike collectivistic cultures, which many Middle Eastern countries are (De Mooij 2018).

A young male creative director tried to explain why the differences were pronounced only in body display. He stated: “Western creative directors are under pressure from their societies and clients to show more respect to women and stay away from stereotypes. In practice, all creative directors develop a vision for the shot and then apply that vision to both the veiled and nonveiled models. Sometimes the model wears the veil only for the shot.”

Interestingly, there was no significant difference in body display between the veiled and nonveiled women in the Egyptian posts on Instagram. The research studied 50 Egyptian brands with veiled and nonveiled women in paired Instagram posts. The significant difference in body display seen in the general sample of veiled versus nonveiled and in the sample of Western brands Instagram posts were not found in the Egyptian brands.

The second creative director explained that “[in Egyptian culture] exaggerated body display is distasteful and undesirable, regardless of the veil, particularly if the post targets a female audience. In general, sexy ads are frowned upon in this society.” His insight explains the significant difference between Egyptian and Western branded posts in feminine touch, whether the model is veiled or unveiled.

One limitation of this study is that the posts are not necessarily ads but what Fox, Nakhata, and Deitz (2019) call visual social media marketing content. Another limitation is the failure to support seven out of 10 hypotheses, which calls for scholars to revise many of their implicit assumptions. Our hypotheses were based on findings of previous research that studied advertising content in different media. Qualitative research is recommended to develop a profound understanding of the idiosyncrasies of visual social media marketing content on Instagram and similar social media networks.

Western brands included U.S. and European brands, while local brands were all Egyptian. Including regional brands from the Middle East and North Africa can enrich and expand the insights. A final limitation is the snapshot effect of collecting all the data in 12 days; a more longitudinal study could provide additional insights.

The research explored the difference between women’s veiled and nonveiled representation along Goffman’s and Kang’s five categories in Western and Egyptian brands on Instagram. Future research could expand the research beyond Egypt to the region and compare gender display across Islamic or Arab countries, and beyond Instagram to other networked media contexts (Butkowski 2021). Researchers could explore the spontaneous regional self-representation of veiled versus nonveiled women selfies on Instagram or other social media (Butkowski et al. 2020). All models, veiled or not, are trained to realize the visions of creative directors. Therefore, more qualitative research is needed with male and female creative directors to enrich the debate about the lack of progress in representation of women in the 40 years since Goffman’s alert.

Behaviors are driven by perceptions rather than reality. Ads are conceived to manage perceptions to drive consumers toward the objectives of the agency’s client. Evolution will happen if and when clients see an economic or social gain in representing women more decently, fairly, and progressively.