A visual audit of diversity imagery across fortune 500 firms

Introduction

Companies draw from the largest pool of their most visible employees to showcase their organizational attractiveness. Organizational attractiveness is broadly defined as the extent to which potential job seekers perceive an organization as a desirable place to work or as a desirable entity with which to initiate a work-related relationship (Aiman-Smith et al., 2001; Ehrhart and Ziegert, 2005). As Gomes and Neves, 2011 explain, job applicants’ perceptions of organizational attributes such as pay, benefits, career development opportunities, and job security strongly shape attractiveness, which in turn fully mediates their intention to apply. In fact, the study found that organizational attributes and job feedback (“knowledge of results”) were the most influential predictors of attractiveness compared to other job characteristics, highlighting that applicants’ perceptions of what organizations communicate externally have a greater impact than technical details of the job itself (Gomes and Neves, 2011). Moreover, employer branding and organizational reputation act as additional layers of attractiveness by reinforcing favorable impressions of the company as a desirable workplace, meaning that visible employees and corporate communication efforts become critical in transmitting these signals and shaping these perceptions (Fiset et al., 2021; Hellmann et al., 2024; Turban and Greening, 1997).

Building on this, research in employer image theory distinguishes between two major types of attributes that shape perceived attractiveness: instrumental and symbolic (Lievens and Highhouse, 2003). Instrumental attributes describe the tangible, practical aspects of employment such as pay, bonuses, benefits, and location that allow job seekers to weigh concrete rewards and working conditions. In contrast, symbolic attributes capture the intangible, emotional, and personality-based qualities people associate with an organization, such as sincerity, prestige, or innovativeness. These symbolic cues enable potential employees to express their identity and align with the organization's perceived character (Aaker, 1997; Slaughter et al., 2004). Studies have shown that both types of attributes influence perceptions of organizational attractiveness, though symbolic attributes often play a crucial role in shaping emotional connection and perceived fit, especially in preliminary stages of job consideration (Lievens and Highhouse, 2003; van Hoye and Saks, 2011).

Consequently, attracting talent becomes both a strategic and essential priority, as organizations compete not only through job design but also through their image and reputation (Highhouse et al., 1999). Companies must understand which attributes draw potential employees and how they can differentiate themselves from competitors to secure top talent. For much of the past decade, diversity has been central to this appeal, with corporate DEI programs and public commitments proliferating in response to both social pressure and competitive advantage arguments (Dobbin and Kalev, 2016). This emphasis on diversity is also reflected in organizations’ broader communication strategies. For example, a survey of more than 2,500 marketers across five countries found that 88% of Generation × respondents and 90% of Millennial respondents believed that using more diverse representation enhances a brand's reputation (McCabe, 2018). These findings suggest that diverse visual representation has become an increasingly important mechanism through which organizations communicate organizational values and cultivate a favorable public image. However, the sociopolitical environment surrounding corporate DEI has shifted considerably. Following the 2023 U.S. Supreme Court decisions restricting race-conscious university admissions, corporate diversity initiatives faced growing legal uncertainty, political opposition, activist pressure, and public scrutiny (Follmer et al., 2024). These pressures intensified as several prominent companies reduced, renamed, or discontinued elements of their DEI programs in response to changing legal and political conditions. For example, Walmart, McDonald's, and Meta publicly scaled back or eliminated aspects of their diversity initiatives, whereas Apple, Microsoft, and Delta Air Lines reaffirmed their commitments despite heightened scrutiny (Gibson and Lyons, 2025; Meyersohn, 2025). At the same time, organizations continued to face expectations from employees, applicants, consumers, and other stakeholders to demonstrate a credible commitment to inclusion. Companies therefore operate within a contested institutional environment in which public DEI commitments may generate legitimacy among some audiences while producing legal, political, or reputational risks among others.

Yet the question remains: how do companies continue to look attractive to stakeholders in an era of heightened scrutiny? Some organizations invest in substantive DEI initiatives, while others make bold public statements with limited structural follow-through. Increasingly, many rely on imagery, such as photographs of leaders, employees, and events to carefully curate an image of inclusivity (Williamson et al., 2003). Corporate websites and recruitment pages have been shown to significantly shape perceptions of organizational attractiveness, with content influencing perceived person–organization fit. These visuals appear in multiple forms, including CEO portraits, annual reports, and “Our People” or “About Us” webpages, and they function less as neutral representations than as deliberate impression management strategies designed to signal legitimacy (Meyer and Rowan, 1977).

From an institutional perspective, these visual practices can also be understood as responses to broader societal expectations regarding diversity and inclusion. Organizations face pressures to demonstrate legitimacy by conforming to widely accepted norms and values (e.g., DiMaggio and Powell, 1983). Consequently, diversity imagery may serve as a recruitment signal directed toward prospective employees and as a means of demonstrating conformity with institutional expectations surrounding equity, representation, and social responsibility. This perspective complements employer image theory by explaining why organizations may adopt similar visual practices even when those images do not accurately reflect their workforce composition. In the next section, we draw on employer image theory (Cable and Turban, 2001; Lievens and Highhouse, 2003; Lievens and Slaughter, 2016) and institutional theory (Meyer and Rowan, 1977) to further explain why these visual signals play such a powerful role in shaping stakeholder perceptions.

Corporate images and organizational attractiveness

Employer-facing webpages have become one of the most strategically important, yet understudied, tools through which organizations construct and project their employer brand. As Theurer et al., 2022 note, these sites represent a central, company-controlled medium for communicating who the organization claims to be and what kinds of employees it seeks to attract. Recruitment professionals recognize this influence: in LinkedIn's global survey of recruitment trends, 61% of HR officers identified a firm's employment webpage as one of the most effective mechanisms for initiating and sustaining employer branding efforts (Schnidman et al., 2017). For job seekers, these pages function as interpretive spaces through which applicants infer an organization's culture, priorities, and values. Evidence suggests that employer websites are nearly universally consulted, with 98.2% of applicants in one study reporting that they used company webpages to discern what it might feel like to join and affiliate with the organization (Banks et al., 2019), and prior consistently shows that these digital materials shape applicant perceptions of organizational attractiveness (Dineen et al., 2007).

Despite the vital role of corporate websites in employer branding, scholars have only recently begun to examine a crucial component of these pages: the photographic images that firms choose to display. These visuals are not merely decorative; they communicate symbolic cues about who belongs, who is valued, and who is imagined as the “right” kind of employee. Employer image theory provides a useful framework for understanding why these visuals matter (Cable and Turban, 2001; Lievens and Highhouse, 2003; Lievens and Slaughter, 2016). Because applicants typically have limited direct knowledge of organizations, they rely on external signals to infer what an employer is like and whether it aligns with their identities and values. Lievens and Highhouse, 2003 differentiate between functional attributes (e.g., pay, security) and symbolic attributes (e.g., inclusiveness, prestige), noting that symbolic cues often have a particularly strong influence because they help job seekers evaluate the level of fit with their identity and values (Slaughter et al., 2004). Photographs are among the most powerful symbolic signals employers provide as they are processed rapidly, evoke immediate impressions, and strongly shape perceptions of belonging and culture (e.g., Burt et al., 2010).

Despite this growing body of research, important gaps remain. Existing studies have primarily examined whether diversity cues influence organizational attractiveness, applicant attention, perceived fit, or job-pursuit intentions. Far less is known about how organizations construct these visual signals in practice, including whether they rely on authentic employee photographs or stock imagery, which demographic groups and visible inclusion characteristics they portray, what contextual themes accompany these representations, and whether these practices vary across industries. Moreover, prior research has rarely integrated employer image theory with an institutional perspective to explain both how these images shape applicant perceptions and why organizations adopt similar diversity-oriented visual practices to establish legitimacy. This study addresses these gaps by examining the visual content of Fortune 500 diversity and inclusion webpages through both theoretical lenses.

Organizations frequently rely on photographs because appealing visual website design, including color, layout, and photographs, enhances perceptions of organizational attractiveness (Braddy et al., 2008). Research consistently demonstrates that photographs help applicants infer organizational values and assess fit (Walker et al., 2011), and that authentic photos of real employees foster greater trust than staged or stock images (Burt et al., 2010). Moreover, for historically underrepresented groups, authentic depictions of gender, racial, or generational diversity in photographs can signal belonging, fairness, and respect (van Hoye and Saks, 2011), and featuring women or racialized minorities can communicate an organization's commitment to diversity (Avery et al., 2013; Baum et al., 2016; Iseke and Pull, 2019).

Yet these visual cues also generate complex interpretations. Research suggests that diversity cues can shape distinct stages of applicant responses, from initial attention to downstream evaluations. For example, Mao et al., 2021 show that depicting women in recruitment photos increases women's attraction to the organization but can reduce job-pursuit intentions among men who hold traditional male breadwinner beliefs. Iseke and Pull, 2019 find that male applicants may infer from such imagery that organizations prioritize gender diversity and support women's advancement, which can shape their reactions. Walker et al., 2012 further demonstrate that diversity cues can increase visual attention and information recall across audiences, showing that both Black and white viewers spend more time on, and recall more information from, websites containing diversity cues. Taken together, this body of work indicates that diversity cues may broadly capture attention and enhance recall across audiences, while evaluative responses such as attraction and job-pursuit intentions vary depending on viewers’ beliefs and interpretations. Thus, these findings underscore that images play a significant role in shaping perceptions of organizational culture, expectations of belonging, and applicant decisions (e.g., Coulter and Punj, 1999), but they also reveal that diversity cues can carry unintended consequences.

Importantly, this imagery may operate more as symbolic presentation than as an accurate portrayal of organizational reality, particularly when firms use purchased diversity photos that misalign with their workforce demographics. Recent research further demonstrates that organizations strategically deploy visual representations of diversity as part of impression management processes, often signaling inclusivity while simultaneously reproducing dominant norms (Alves Ferreira et al., 2026; Ang et al., 2020). When diversity photos are not supported by genuine structural or cultural change, external audiences may interpret them as insincere, exposing organizations to reputational risk (Windscheid et al., 2016). Empirical evidence shows that organizations often overstate diversity visually. For example, Pippert, Essenburg and Matchett (2013) found that 82% of U.S. universities portrayed Black students in their admissions viewbooks at rates that more than doubled their actual enrollment. Similarly, Chichester, Jewell, LePrevost, and Lee, 2024 show that more diverse imagery is often paywalled in commercial stock libraries, creating both representational and structural barriers for all but the wealthiest clients.

The incentives to engage in such practices align with earlier findings that visual design elements significantly shape applicants’ perceptions of employer image (Cober et al., 2004; Coulter and Punj, 1999). For example, Cober and colleagues (2004) observed that 78% of Fortune's Best Companies included pictures of people, diverse groups, or awards on their employment homepages, a number that grew to 86% within just two years. These visuals help organizations present themselves as socially responsible and diversity-oriented (Williamson et al., 2003).

Within this logic, highlighting diversity photos to attract talent who share these values (Roberson, 2019) can serve as a form of symbolic window dressing (Dobbin and Kalev, 2016). Companies may use inexpensive or generic stock photography to construct a visually compelling image of inclusivity, leveraging their accessibility and low cost to develop an appealing employer brand in a resource-efficient manner. Although such imagery can help organizations project a diversity-oriented identity that appeals to applicants and enhances competitiveness (Tsai and Yang, 2010; Younis and Hammad, 2021), their impact depends on how these visual signals are interpreted by viewers.

In this paper, we examine how Fortune 500 companies use visuals on their diversity and inclusion webpages, a feature that has received far less scholarly attention than written content (Highhouse and Hoffman, 2001). Our analysis focuses on whether these webpages display photos and, if so, whether the images depict actual employees or rely on stock photography. We also examine demographic representation, visible inclusion characteristics, contextual themes, and industry-level differences in these visual practices. By integrating employer image theory with institutional theory, we consider how diversity imagery shapes perceptions of organizational identity and attractiveness while also helping firms respond to institutional expectations and signal legitimacy. While companies often use visuals to signal inclusiveness and belonging, prior research shows that such cues can exaggerate or symbolically construct diversity rather than reflect actual workforce composition (Chichester et al., 2024; Wilson and Meyer, 2009). Beyond identifying whether companies use stock photography, we examine the demographic representation, visible inclusion characteristic, and contextual themes presented in these images, as well as whether these visual practices vary across industries. This broader approach enables us to assess the source of the images that companies display and the ways in which visual content constructs and communicates an inclusive organizational identity. Accordingly, we ask the following research questions:

• RQ1: To what extent do Fortune 500 companies use stock photos on their diversity and inclusion webpages?
• RQ2: How are demographic groups, visible inclusion characteristics, and organizational contexts represented in the images displayed on Fortune 500 diversity and inclusion webpages?
• RQ3: To what extent do these visual practices differ across industries?

Methods and data analysis

Dataset construction and coding process

This study adopts a mixed-methods approach that combines qualitative coding of visual data with quantitative analysis using chi-square tests to evaluate contemporary corporate visual practices. We created our dataset by systematically collecting publicly available images from Fortune 500 company websites, examining all companies listed on the 2025 Fortune 500 to ensure a comprehensive and clearly defined sampling frame. We focused specifically on pages related to diversity, equity, and inclusion (DEI), as these sections most explicitly communicate companies’ commitments to diversity and representation. These pages were identified by systematically reviewing each company's main navigation menus (e.g., “Diversity, Equity, and Inclusion,” “Our People,” “Our Culture,” and “About Us”) and locating subsections explicitly referencing diversity, equity, inclusion, belonging, or employee diversity initiatives. When DEI-related content was not immediately accessible, we used the website's internal search function and applied consistent keyword searches (e.g., “diversity,” “inclusion,” “equity,” “belonging,” “culture,” “people”) to locate relevant pages. All data were collected within a defined two-month period, from March to April 2025, to ensure consistency across companies and minimize potential temporal variation.

From this sample, we selected only images that included people and excluded photos of products, buildings, or abstract graphics. Focusing on human images allowed us to analyze the most direct visual signals of DEI in corporate imagery. This process yielded 387 usable images representing 340 unique companies. Of the companies reviewed, 303 (61%) displayed one person-centered image and 37 (7%) displayed multiple images. In contrast, 160 companies (32%) did not display an accessible person-centered image on their DEI-related webpages. Although these firms were not included in the image-level analyses, their absence of person-centered imagery constitutes a distinct visual communication strategy that is considered in our interpretation of the findings. From there, each image was indexed using a photo ID, a company ID, and an industry classification. The primary unit of analysis in this study is the image, as it represents the level at which visual content and stock vs. original classifications are coded; however, we also incorporate individual-level coding (e.g., demographic characteristics within images) and firm-level aggregation (e.g., industry comparisons) to support complementary analyses.

The next stage involved coding each image across several dimensions of visibility and identity. Drawing on research in visual management studies and impression management, we conceptualize corporate images as symbolic artifacts that communicate organizational identity and values (Bell and Davison, 2013). Accordingly, our coding framework differentiates between three types of visual cues: demographic representation, which captures visible attributes such as perceived sex, age, and race proxies; and visible inclusion characteristic, defined as visually observable cues that may signal diversity or non-dominant identities (e.g., tattoos, religious symbols); and contextual signals, which reflect the broader activity or setting depicted in the image (e.g., teamwork, celebration, or industry-specific tasks). Together, these categories allow us to systematically examine how organizations visually construct and communicate inclusivity, belonging, and organizational values through corporate imagery. Table 1 provides a comprehensive overview of all coding categories, variables, and their operational definitions. In all cases, coding decisions were based strictly on visually observable cues and were only coded when characteristics were clearly and unambiguously apparent. Percentages for image-level variables, including visible inclusion characteristics, were calculated using the total number of coded images as the denominator (N = 387), whereas percentages for person-level variables, including perceived sex, age, and skin tone, were calculated using the total number of coded individuals as the denominator (N = 2,005). It is important to note that each demographic characteristic and visible inclusion characteristic was coded as a separate variable. Accordingly, the coding framework captures the presence of individual attributes but does not assess how multiple identities co-occur or interact within the same person or image.

Table 1. Coding categories, variables, and operational definitions.

We counted the number of people depicted and coded perceived sex as men or women based on visual presentation, recognizing that this approach could not capture non-binary, transgender, or other gender-diverse identities. We also recorded the presence of people of color (1 = yes, 0 = no). Age diversity was also coded by noting the presence of children (infant-16), young adults (16–25), adults (25–50), and older adults (50+), reflecting research showing that perceptions of organizational attractiveness are shaped by intergenerational cues (Lyons and Kuron, 2014). These categories capture broad, visually distinguishable groupings rather than precise age determinations, and ambiguous cases were not coded. To assess skin tone, we used a protocol adapted from the Massey-Martin New Immigrant Survey (NIS) Skin Color Scale (Massey and Martin, 2003) and further informed by emerging research using the Monk Skin Tone Scale (Monk, 2023). Coders compared the skin tone of each person shown in an image with a reference chart showing 10 identical hands ranging from very light (1) to very dark (10). They then classified each person's skin tone as Light (score: 1–2), Medium (score: 3–5), Dark (score: 6–10), or Not sure (score: Not applicable). Similar classifications have been used to assess photographs across a variety of fields (`e.g., Abu-Ghname et al., 2020; Chichester et al., 2024 Lamb et al., 2023);, to create a more streamlined classification system. The “Not Sure” category was used when lighting, image quality, or filters prevented a reliable assessment.

Beyond basic demographic variables, we also coded for visible inclusion characteristics in each image. These characteristics were coded as binary variables (1 = yes, 0 = no). Visible inclusion characteristics included identity-related markers, such as Pride, religious, and disability-related cues, as well as visible body-size representation. Specifically, photographs were coded for the presence of Pride symbols (e.g., rainbow flags, LGBTQIA + symbols, or Pride-themed apparel), evidence of disability (e.g., wheelchairs or prosthetic or bionic arms), cultural or religious identifiers (e.g., culturally specific garments such as hijabs or turbans), visible tattoos, markers of military or veteran status (e.g., fatigues or service medals), pregnancy (e.g., a visibly pregnant abdomen), body-size representation (e.g., visibly apparent larger body types), and Indigenous representation (e.g., Indigenous cultural dress or regalia). Obesity was coded as a visible body-size characteristic because body size is associated with social and workplace stigma and may broaden representations of inclusion (e.g., Lemmon et al., 2022). It was not treated as theoretically equivalent to religious, disability, or LGBTQ + identities. Visible symbols related to LGBTQIA + pride, disability, religion, and other forms of inclusion may function as cues of organizational inclusivity; however, such representations can also operate as symbolic gestures that do not necessarily reflect structural transformation (Ashley and Empson, 2017; Mavin and Grandy, 2016). To enhance consistency and reduce subjectivity, variables such as obesity and pregnancy were coded only when clearly and visibly apparent in the image and based on observable cues rather than clinical or inferred characteristics.

To capture the broader context in which individuals were depicted, photos were coded by the first and second author according to the primary activity or behavior displayed, with a focus on identifying the dominant contextual theme represented in each image. Because many photographs contained multiple individuals and activities, coders were instructed to assign a single primary theme based on the most salient visual message conveyed by the image rather than coding all possible activities present. Attention was paid to distinguishing between images that depicted work-related activities, social interactions, community engagement, and symbolic representations of inclusion.

These results were then compared across coders, with differences discussed and resolved through consensus, and subsequently refined and organized into clearly defined, mutually exclusive thematic categories. Several rounds of refinement were undertaken to reduce conceptual overlap among categories. For example, images depicting smiling individuals were not automatically classified as “Smiling/Positive Interactions” if they were clearly engaged in work-related collaboration, volunteering activities, celebrations, or role-specific tasks. Instead, the “Smiling/Positive Interactions” category was reserved for images in which positive affect, friendliness, or interpersonal warmth was the primary visual focus and no more specific contextual activity could be identified. Similarly, “Workplace Collaboration” was restricted to images depicting teamwork, meetings, mentoring, or task-oriented professional interactions, whereas “Industry- or Role-Specific Depictions” captured individuals performing specialized occupational duties or appearing in environments that clearly signaled a particular profession, industry, or organizational role.

This iterative process involved multiple rounds of discussion and refinement to ensure consistency and conceptual clarity in the final coding scheme. This process resulted in six overarching categories: (1) Smiling/Positive Interactions, defined as images emphasizing friendliness, approachability, or positive interpersonal engagement without a clearly identifiable task; (2) Workplace Collaboration, defined as task-oriented professional interactions such as meetings, teamwork, mentoring, or problem solving; (3) Community Participation, defined as volunteering, charitable activities, outreach initiatives, or engagement with community stakeholders; (4) Celebrations/Team Bonding, defined as group celebrations, recognition events, social gatherings, or activities intended to foster cohesion and unity; (5) Outdoor and Physical Activities, defined as recreational, wellness-oriented, or physically active pursuits occurring primarily outside traditional work settings; and (6) Industry- or Role-Specific Depictions, defined as images highlighting specialized occupations, technical expertise, or industry-specific work environments.

Descriptions of each category and inter-rater reliability are reported in Table 1, and illustrative examples of each code are provided in the Supplementary Materials to demonstrate how the categories were applied in practice. To assess coding reliability, both authors independently coded each photo in the dataset based on key variables, including demographic representation, visible inclusion characteristics, and contextual themes, and then met to compare their scores. Inter-rater agreement was 92% prior to discussion, after which all differences were reviewed and resolved to reach consensus.

As a last step, we identified whether each image was an original corporate photograph or a stock photo. Given the relative novelty of using reverse image search methods in organizational research, we provide additional detail here to enhance transparency and replicability. To do this, every collected image was uploaded into two reverse image search tools: TinEye (https://tineye.com), which is effective at tracing exact matches and locating licensing sources, and Google Image Search (https://images.google.com/), which identifies both exact and visually similar images. This process revealed when photos appeared on commercial stock platforms or across multiple corporate websites. Images were classified as stock photos if (a) they were directly traceable to a recognized stock image provider or (b) similar underlying images, including versions that had been cropped, resized, or otherwise minimally modified, appeared across multiple unrelated organizational websites, indicating non-exclusive use. When images appeared on multiple platforms, we prioritized the most authoritative sources (e.g., original stock provider) to confirm classification. Confirmed stock images were recorded along with their providers, including Shutterstock, iStock, Alamy, Adobe Stock, Getty Images, and Unsplash. In addition, we documented all instances in which the same stock image was used by more than one organization to capture patterns of image reuse across firms. This systematic procedure strengthens confidence in our classification and provides a replicable approach for future research using visual data.

A quantitative content analysis was conducted to systematically examine visual representations in corporate imagery. Frequencies and proportions were calculated to summarize patterns across all variables. To evaluate whether stock photo usage differed by industry, a chi-square test of independence was conducted across twenty-one industry categories. A second chi-square test of independence was performed to assess whether stock photo usage varied across six image context categories (smiling and positive interactions, workplace collaboration, community participation, celebrations and team bonding, outdoor activities, and industry- or role-specific depictions). For both analyses, standardized residuals were examined to identify which categories contributed most to significant associations.

Results

The following demographic percentages are based on the 2,005 coded individuals represented across the 387 coded images. In terms of perceived sex, women (n = 1,124; 56%) were depicted more frequently than men (n = 819; 41%), while 62 individuals (3%) could not reliably be classified. A similar pattern emerged for skin tone, with light-skinned individuals (n = 991; 49%) appearing more frequently than those with medium (n = 494; 25%) or dark skin tones (n = 427; 21%). An additional 93 individuals (5%) could not be reliably classified by skin tone. With respect to age, adults (n = 1,484; 74%) were the most frequently represented group, followed by older adults (n = 226; 11%), young adults (n = 112; 6%), and children (n = 13; 1%). However, 170 individuals (8%) could not be reliably classified into an age category.

The content analysis revealed a limited yet discernible presence of visible inclusion characteristics in the corporate imagery of the 387 Fortune 500 photographs. Body-size representation was the most frequently observed characteristic, with visible obesity appearing in 52 images (13.4%). Among identity-related inclusion markers, cultural or religious identifiers appeared in 31 images (8.0%), tattoos in 28 images (7.2%), disability-related markers in 22 images (5.7%), Pride representation in 21 images (5.4%), military or veteran status in 3 images (0.8%), pregnancy in 2 images (0.5%), and Indigenous representation in 0 images (0%).

Table 2 reports the distribution of Fortune 500 companies across industries by photo type, showing the proportion of firms that use original (non-stock) images, stock photos, or no photos relative to the total number of companies within each industry (i.e., the denominator is the number of firms in each industry category). Across industries, the Apparel sector had the largest number of companies displaying original, non-stock photos (n = 8, 100%), followed by Transportation (n = 7, 78%) and Household Products (n = 6, 75%). Among those that did use photos, the reliance on stock imagery was limited: Utilities showed the highest proportion of stock-photo use (n = 11, 52%), followed by Materials (n = 4, 50%) and Engineering & Construction (n = 11, 44%). In contrast, the Financials sector exhibited two related patterns. Of the 95 Financial firms, 25 (26%) used stock photography and 32 (34%) displayed no accessible person-centered imagery, while 38 (40%) used original photographs. Thus, the sector's reliance on stock imagery and its comparatively high proportion of firms without person-centered imagery are not contradictory. Together, these patterns indicate that Financials firms relied less heavily on original person-centered photography. Overall, these findings highlight substantial cross-industry variation in visual communication practices. However, percentages for industries containing relatively few companies, such as Media, Grocery and Drug Stores, Apparel, Materials, Household Products, and Telecommunications, should be interpreted cautiously because a small number of firms can produce substantial percentage differences. Accordingly, the industry percentages are presented primarily as descriptive patterns.

Table 2. Presence and type of diversity imagery on fortune 500 diversity and inclusion webpages by industry.

N = 500 companies. Residual numbers in bold indicate meaningful deviation from expected frequencies. Industries shown in bold indicate statistically significant differences (p < .05).

Across the 387 photographs, the most common contextual themes depicted in photographs were workplace collaboration and professional engagement (n = 158, 41%), followed by smiling and positive interactions (n = 92, 24%), industry- or role-specific depictions accounted for (n = 58, 15%), and celebrations, unity, and team bonding (n = 52, 13%), while community participation and volunteering appeared far less frequently (n = 19, 5%). Outdoor and physical activities were the least represented theme (n = 8, 2%). Collectively, these patterns suggest that companies most often use visual content to emphasize teamwork, professionalism, and positive interpersonal dynamics within workplace settings, rather than community-oriented engagement or physically active contexts (Table 3).

Table 3. Presence and type of diversity imagery on fortune 500 diversity and inclusion webpages by Photo depiction.

N = 387 photos. Residual numbers in bold indicate meaningful deviation from expected frequencies.

Stock photo usage

Of the 387 photos analyzed across 75 companies, 76 were identified as stock images, representing approximately 20% of all images examined. The most frequently used stock photo sources included Shutterstock (28 photos), Alamy (19 photos), and iStockphoto (20 photos). Additional sources such as Adobe Stock (7 photos), Getty Images (3 photos), and Unsplash (1 photo) also contributed to the dataset, though to a lesser extent. Notably, 22 of the 76 stock photos (29%) appeared on multiple platforms; therefore, each image was counted once in the overall stock-image total but attributed to every platform on which it was identified, making platform-specific source counts non-mutually exclusive.

Further analysis revealed that the photos from twelve companies (3%) used duplicate stock photographs across their DEI visual materials, representing five distinct images in total. One image appeared on the websites of four companies (DCP Midstream, PNC Financial, Edward Jones, and LPL Financial), while five additional pairs of companies used identical visuals. These images originated from major stock photo providers, including Shutterstock, Alamy, Getty Images, and iStockphoto.

We performed a chi-square test of independence to examine whether stock-photo usage differed across 21 industries. The results revealed a significant association, χ²(20) = 32.47, p = .039, indicating variation in the use of stock imagery across sectors.

The effect size was in the moderate range (Cramér's V = .25), based on Cohen's (1988) benchmarks for χ² tests (small = .07–.21; moderate = .21–.35; large >.35). We also examined expected cell counts and found that assumptions were satisfied; with 9% of cells had expected counts below 5, which falls within accepted guidelines for chi-square tests (i.e., no more than 20% of cells below 5 and none below 1). Although these values fall within commonly applied guidelines for chi-square analyses, findings involving industries with small cell sizes should be interpreted cautiously. Standardized residuals showed that the Financials (residual = 4.11) industry used significantly more stock photos than expected. In contrast, Energy (residual = −2.98), Retailing (residual = −2.84), Utilities (residual = −2.03) used significantly fewer stock images than expected. These findings suggest that while certain industries rely heavily on generic imagery to represent their workforce, others use original photographs, reflecting notable sector-level differences in visual branding practices and approaches to depicting organizational identity (See Table 2).

A second chi-square test of independence examined whether stock-photo usage differed across six image categories (smiling and positive interactions; workplace collaboration; community participation; celebrations and team bonding; outdoor activities; industry- or role-specific depictions) and revealed a significant association, χ²(5) = 31.08, p < .001. The effect size was also in the moderate range (Cramér's V = .28, Cohen, 1988). We also examined expected cell counts and found that assumptions were satisfied; only one cell had an expected count below 5, representing less than 20% of cells and therefore meeting recommended thresholds for chi-square tests. Standardized residuals indicated that Workplace collaboration images contained significantly more stock photos than expected (residual = 2.61), making this category the largest contributor to the chi-square value. In contrast, Celebrations and team bonding (residual = −3.03) and Industry- or role-specific depictions (residual = −2.68) contained significantly fewer stock images than expected, while Outdoor activities (residual = −1.27), Community participation (residual = .17), and Smiling and positive interactions (residual = –.46) were close to expected values. Taken together, these results suggest that organizations rely more frequently on stock photography when depicting professional collaboration, whereas images portraying celebrations or job-specific tasks are more likely to feature original photographs (See Table 3).

Discussion

Drawing on employer image theory (Cable and Turban, 2001; Lievens and Highhouse, 2003), and contemporary research on organizational attractiveness, our findings provide answers to each of our three research questions by documenting the prevalence of stock photography, describing the demographic and contextual content of corporate diversity imagery, and demonstrating that these visual practices vary systematically across industries. Our analysis shows that visual communication is not a neutral aesthetic decision but a strategic expression of organizational identity. Examining 387 corporate images from 340 Fortune 500 companies from 2025, we first found that the use of stock photography is widespread on corporate diversity websites. Using a novel reverse image search technique, we identified that one in five companies rely on pre-produced stock visuals rather than photographs featuring their own employees. These images typically portray staged scenes of multicultural teamwork and collaboration, which may project representations of inclusivity that may or may not reflect the organization's actual workforce or culture (Aiello et al., 2023). This pattern is consistent with a broader organizational tendency to convey diversity symbolically rather than substantively (Dobbin and Kalev, 2017; Meyer and Rowan, 1977) and suggests that corporate visuals can function as instrumental and symbolic signals that may shape perceptions of attractiveness, particularly through portrayals of inclusivity, collaboration, and belonging. At the same time, this raises important questions about what constitutes an appropriate or credible representation of diversity in corporate imagery, including which groups should be depicted and whether such representations should reflect organizational composition, broader societal demographics, or other benchmarks.

A notable feature of the findings is that, for most firms, DEI was represented through a single prominent person-centered image. This suggests that many Fortune 500 companies condense their public message of inclusion into one primary visual, making the people depicted especially important in signaling who belongs within the organization. Addressing our second research question, our findings characterize how demographic groups and visible inclusion characteristics are represented in corporate diversity imagery. The perceived sex, age, skin tone, body size, and visible inclusion characteristics shape the forms of diversity made visible. Body-size representation was more common than the identity-related markers examined, but it was not treated as conceptually equivalent to religious, disability, or LGBTQIA + identities. Rather, its inclusion reflects the relevance of weight stigma and the visibility of larger bodies in corporate imagery (Lemmon et al., 2022).

Our third research question examined whether these visual practices differed across industries, and the findings indicate that they do. The reliance on stock imagery also varies across industries. The Financials sector showed two related patterns: firms were more likely than expected to use stock imagery, and a comparatively high proportion displayed no accessible person-centered image. These findings are not contradictory. Rather, both indicate less reliance on original person-centered photography, with Financials firms tending either to use standardized stock imagery or to communicate without person-centered imagery. The office-based and comparatively less visually distinctive nature of financial work may make generic representations of professionalism, collaboration, and diversity more readily applicable. However, this explanation is interpretive and was not directly tested in the present study. In contrast, firms in the Energy, Retailing, and Utilities sectors are less likely to rely on stock photography, which may reflect the more visible, place-based, and operationally embedded nature of their activities. In these sectors, work is often tied to physical infrastructure, frontline employees, and customer interactions, making firm-specific imagery more meaningful and credible. Consequently, industries that rely more heavily on stock imagery may reproduce similar visual cues, potentially contributing to greater homogeneity in corporate branding (Hatch and Schultz, 2003).

This homogenization is especially visible when organizations use the same stock images. Among the 75 companies using stock photography, 12 shared at least one image with another company. In total, five images were reused across firms, including one that appeared on four company websites (DCP Midstream, PNC Financial, Edward Jones, & LPL Financials) and four additional instances in which pairs of companies used the same image: Vistra Corp and Jefferies Financial; Emerson Electric and Cigna; W.R. Berkley and M&T Bank; and Hartford Financial and Science Applications International (see Appendix). This level of repetition illustrates the degree to which many organizations depend on a common database of prefabricated imagery. Although such photos offer a convenient way to signal professionalism or inclusivity, their reuse across multiple organizations may serve as a visible cue that may shape how viewers interpret these images, particularly when identical visuals appear in different corporate contexts. Accordingly, rather than assuming reduced authenticity, our findings point to the potential for image repetition to shape audience interpretations, an issue that warrants further empirical investigation. More broadly, while our data do not allow us to determine what constitutes an “ideal” diversity image, it stands to reason that corporate imagery should align with the organization it represents, rather than risk appearing disconnected from its underlying workforce or practices. Taken together, these patterns are consistent with a reliance on symbolic signaling rather than authentic representation, reflecting the ceremonial approaches to diversity described above. These findings suggest that companies may visually signal inclusivity through imagery, although we do not directly assess whether such signals correspond to underlying workforce composition or organizational practices.

Theoretical and practical implications

These findings extend employer image theory (Cable and Turban, 2001; Lievens and Highhouse, 2003) by showing how authenticity, impression management, and legitimacy processes unfold in the visual domain. Visual diversity is not merely an aesthetic choice but a symbolic cue that shapes job seekers’ mental representations of an organization's identity, culture, and values. Staged or stock photography can be interpreted as an impression-management strategy (Burt et al., 2010), signaling inclusivity at the symbolic level while aligning with institutional accounts of ceremonial compliance that reflect image management rather than substantive DEI commitments. In contrast, authentic or employee-generated imagery may provide cues that are more likely to be interpreted as authentic, potentially strengthening perceptions of value congruence, belonging, and organizational credibility. Distinguishing between these forms of visual signaling offers a more nuanced understanding of whether corporate imagery conveys functional information about the actual workplace or merely constructs symbolic attributes of the employer brand.

By integrating visual analysis with employer image theory and institutional perspectives, this study demonstrates how corporate imagery operates at the intersection of branding, identity construction, and social signaling, raising questions about the authenticity and reliability of the DEI cues organizations project. This integration clarifies the distinct but complementary explanatory roles of the two frameworks. Employer image theory explains how prospective applicants interpret visual cues and incorporate them into judgments of organizational attractiveness and fit, whereas institutional theory (Meyer and Rowan, 1977) explains why organizations adopt and reproduce similar diversity-oriented visual practices in response to normative expectations and pressures for legitimacy. Together, these frameworks reveal that corporate diversity imagery simultaneously operates as an applicant-facing signal and an organization-facing legitimacy mechanism.

The sociopolitical context surrounding the study further clarifies the institutional significance of these visual practices. In the contemporary U.S. environment, organizations face competing pressures regarding whether and how they communicate their commitments to diversity and inclusion. Growing legal challenges, political opposition, state-level restrictions, and changes in federal DEI policy have encouraged some companies to reduce, rename, or deemphasize their formal DEI programs and public commitments (Follmer et al., 2024). At the same time, applicants, employees, consumers, and other stakeholders may continue to expect organizations to demonstrate inclusivity and provide evidence that members of historically underrepresented groups belong within the workplace. Corporate imagery may therefore offer organizations a flexible means of navigating these competing demands. Visual representations can preserve symbolic signals of inclusion even when companies reduce the visibility of explicit DEI terminology, goals, or programs.

From an institutional perspective, this tension may help explain why organizations retain diversity-oriented imagery while making fewer substantive or easily verifiable commitments. Stock photographs are particularly relevant in this context because they allow firms to construct an inclusive public identity without necessarily connecting that representation to identifiable employees, workforce demographics, or formal organizational practices. Such imagery may help organizations maintain legitimacy with audiences that value inclusion while limiting the specificity of the claims being communicated. However, this flexibility also heightens concerns about ceremonial compliance and decoupling, particularly when visual representations cannot be reconciled with employee experiences or organizational outcomes. Thus, the current sociopolitical environment makes the authenticity of visual DEI signals especially consequential, as the same image may be viewed as evidence of belonging by one audience, performative impression management by another, or a politically contested organizational statement by a third.

Corporate imagery plays a powerful role in recruitment by shaping job seekers’ perceptions of person–organization fit, belonging, and value alignment (Lievens and Highhouse, 2003; Lievens and Slaughter, 2016), and may be particularly valuable in linguistically diverse contexts where language differences can create misunderstanding and impede information exchange (Fiset et al., 2024). Because these visuals function as symbolic cues about organizational identity, authenticity in their presentation is essential. When diversity imagery appears superficial or merely decorative, it may be associated with lower perceived trust in the employer brand, particularly when job seekers can verify these signals against alternative sources of information. Glassdoor diversity ratings provide employee-generated evaluations of a company's EDI efforts, captured through a “Diversity & Inclusion” workplace factor rating in which current and former employees report their satisfaction with the organization's DEI efforts on a 5-point scale, with scores aggregated at the company level based on voluntary, anonymous feedback (Dube and Zhu, 2021). These ratings enable applicants to compare corporate representations with reported workplace experiences, which may make inconsistencies more visible and potentially affect perceptions of organizational credibility. Research by Mao et al., 2021 further shows that interpretations of diversity imagery vary across audiences; photos featuring women or racial minorities tend to attract applicants who value inclusivity but may deter individuals who hold traditional or gender-biased beliefs. Representation therefore may function as a signal that influences who feels welcomed or discouraged from pursuing employment.

These findings also highlight the dual function of visual representation in organizational communication. Companies can intentionally use DEI-oriented imagery to attract values-aligned applicants and to signal a commitment to inclusive norms (e.g., Dineen et al., 2007). The first function is consistent with employer image theory, because imagery influences how applicants evaluate organizational identity, fit, and attractiveness. The second reflects institutional theory, because the public display of diversity allows organizations to demonstrate conformity with prevailing expectations regarding inclusion and social responsibility. However, this strategy is effective only when the visuals appear credible. This credibility is especially important in the current sociopolitical environment, where corporate DEI claims are subject to intensified scrutiny from applicants, employees, political actors, advocacy groups, and consumers holding divergent expectations. Heavy reliance on generic stock imagery, especially idealized depictions of multicultural teams, may be perceived as less authentic and may create a disconnect between projected and lived organizational culture. Firms seeking to build trust, differentiate their brand, or authentically represent their workforce should invest in employee-based visuals that reflect actual demographic and cultural realities.

Limitations and future research directions

This study is limited by its reliance on visual data, which necessarily restricts analysis to characteristics that are readily observable in photographs. Many important dimensions of diversity are invisible and cannot be reliably inferred from images, including mental health conditions, neurodiversity, chronic illness, learning disabilities, sexual orientation (absent explicit cues), socioeconomic background, and most forms of disability (Buchanan and Settles, 2019; Clair et al., 2005). Consequently, our findings reflect how organizations elect to visually signal diversity rather than the full scope of diversity that may exist within their workforce. Additionally, the sample may limit representativeness, as it includes only publicly available corporate website images and excludes firms that did not display accessible person-centered imagery on their DEI-related webpages. Because the prevalence of such exclusions varies across industries, industry comparisons should be interpreted cautiously. This exclusion may be consequential, as firms without person-centered imagery may differ systematically from those that visually represent employees, applicants, or other stakeholders in their diversity communications. As a result, the findings should be interpreted as reflecting the visual practices of companies that chose to use person-centered DEI imagery, rather than the full population of Fortune 500 firms. Broader sociopolitical shifts, including changes in corporate diversity messaging following the election of Donald Trump and his administration's anti-DEI executive orders, may also shape how diversity is portrayed, which the current cross-sectional design cannot fully capture. Future research should integrate visual analysis with employee interviews, internal demographic data, or applicant surveys to examine how visible and invisible forms of diversity align or diverge in shaping perceptions of organizational inclusion.

A further limitation concerns the coding of demographic characteristics from visual cues. Perceived sex was coded based on perceived visual presentation of sex and therefore cannot capture identities that do not conform to conventional presentation norms, including non-binary, transgender, or gender-diverse identities. Similarly, although demographic characteristics such as gender and skin tone were coded, they were analyzed independently rather than intersectionally. As a result, the study cannot assess how multiple identities co-occur within the same images or how visual prominence, positioning, or organizational roles may differ across intersecting identities (e.g., whether women depicted are disproportionately light-skinned or whether racialized individuals are more frequently portrayed in background or subordinate positions). Future research should adopt an intersectional approach that examines how multiple dimensions of identity interact within corporate imagery to shape representations of inclusion and organizational power dynamics.

The result of this analysis is that disability representation is likely underestimated. Our coding captured only visible markers of disability (e.g., wheelchairs or prosthetic limbs), which represent a narrow subset of the disability spectrum (Prince, 2017; Santuzzi et al., 2014). Many disabilities are non-apparent or deliberately concealed due to stigma, professional norms, or organizational expectations of “fit,” resulting in selective visibility within corporate imagery. Moreover, although a structured coding scheme was employed, the interpretation of visual content introduces potential threats to coding reliability, as subjective judgments may influence how characteristics and categories are assigned. Although we report raw percentage agreement between coders, we did not calculate a chance-corrected reliability coefficient, such as Cohen's kappa or Krippendorff's alpha. Accordingly, the reported agreement should be interpreted as an indicator of coding consistency rather than a fully chance-adjusted estimate of inter-rater reliability. In addition, the two coders were also authors of the study, which may introduce potential bias in how coding categories were interpreted or applied. This is particularly relevant for variables such as body size and pregnancy, which were coded only when clearly and visibly apparent and based on observable cues rather than clinical or inferred characteristics. We also note the challenges associated with inferring demographic characteristics such as age from photographs and the broader constraints of visually coding identity. Future studies could examine how organizations decide whether to visually depict disability and how job seekers with visible vs. invisible disabilities interpret and respond to these cues, shedding light on how visual norms may reinforce or challenge ableist assumptions in employer branding. They could also incorporate external workforce demographic data and employ experimental or survey-based methods to evaluate how diverse types of imagery influence perceptions of authenticity.

Relatedly, future research should more directly examine the tension between visual appeal and perceived authenticity in corporate imagery. Research on digital aesthetics suggests that authenticity plays a critical role in shaping audience perceptions, with raw, unpolished images often appearing more genuine, whereas highly polished corporate photos may be perceived as more artificial (Colliander and Marder, 2018; Douglas, 2014; Pantti, 2013). At the same time, organizations frequently rely on polished imagery because it has been shown to produce favorable evaluations in controlled settings (Hao et al., 2015; Litvin and Mouri, 2009). This tension reflects longstanding insights from institutional theory, which suggest that many corporate practices, particularly those related to diversity, may be ceremonial, and designed to align with external expectations rather than substantively transform internal realities (Dobbin and Kalev, 2017; Edelman, 1990; Meyer and Rowan, 1977). Future work could explore how audiences navigate this tension and how different forms of visual presentation influence perceptions of authenticity, credibility, and organizational intent.

Finally, our analysis captures symbolic representation rather than lived experience or structural inclusion. Corporate photographs are curated artifacts shaped by branding strategies, institutional pressures, and reputational concerns, and the presence of visible diversity does not necessarily indicate meaningful inclusion, just as its absence does not confirm exclusion. Furthermore, the use of chi-square analyses does not permit causal inferences regarding why these patterns occur or how they influence audience perceptions. The chi-square analyses also assumed independence of observations; however, some companies contributed multiple images to the dataset. Because images from the same firm may reflect shared branding strategies, design templates, or communication guidelines, these observations may not be fully independent. This clustering may have affected the estimated chi-square statistics and associated significance levels. As such, the quantitative findings should be interpreted as descriptive associations among observed images rather than as causal or fully independent firm-level estimates. Future research should use firm-level analyses or statistical methods that account for the clustering of multiple images within companies. Future research should examine how job seekers assess the authenticity of corporate imagery and distinguish between real and stock photographs and their effects over time. This work could explore how diversity portrayed through corporate imagery aligns or fails to align with organizations’ stated diversity values and whether the use of stock photos is interpreted as a signal of espoused diversity commitments (Fiset et al., 2025). Experimental studies could also assess how variations in racial and gender composition, photo authenticity, and visual style shape perceptions of organizational trust, credibility, and application intentions across different demographic and ideological groups, helping organizations better understand how to design recruitment materials that both attract diverse talent and reflect genuine structural commitments to inclusivity.

In sum, this study contributes to understanding how visual diversity functions as both a recruitment signal and an impression management strategy. By integrating perspectives from institutional theory and recruitment psychology, we suggest that imagery may function as an important mechanism shaping perceptions of organizational attractiveness, while also raising questions about how visual representations align with broader organizational practices. Future research should examine the extent to which organizations’ visual narratives align with broader structural commitments to diversity and inclusion. Questions of authenticity in corporate imagery may depend not only on visual style but also on how audiences interpret the relationship between organizational imagery and employee experiences. As stakeholders and job seekers increasingly evaluate organizational communications across multiple information sources, perceptions of visual credibility may become an increasingly important component of employer branding and organizational reputation.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. ChatGPT was employed solely for grammar checks and was not involved in developing any theories or data analysis.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Raffaele Lombardi, Libera Università Maria SS. Assunta, Italy

Reviewed by: Marina Ciampi, Sapienza University of Rome, Italy

Cosimo Miraglia, Sapienza University of Rome, Italy