Ethnic Minorities in Advertising

Abstract

The literature on diversity suffers from several shortcomings, leading to conflicting views or unjustified beliefs about ethnic minorities in advertising. Evidence is needed to show whether advertising has indeed become more ethnically diverse over time, whether any cultural differences exist, and whether ethnic diversity in advertising in new media differs from that in traditional media advertisements. Building on extant research and social role theory, this study provides two empirical studies, a meta-analysis and content analysis, and questions commonly held beliefs in the diversity literature. The findings show that any increase in ethnic minority representation in advertising is fully captured by the fact that ethnic minorities in the general population have increased and that relevant changes have occurred only in Eastern societies where the overrepresentation of the white minority has decreased. Ethnic minorities are well represented in online and TV media but not in print media. Except for the product use of ethnic minority endorsers, stereotyping prevails and barely differs between old and new media. These findings have implications for both researchers and advertisers.

Numerous studies have investigated the representation and portrayal of ethnic and racial minorities in advertising. They have shown that ethnic minority endorsers, as compared to the dominant (often white) majority in a society, are underrepresented and portrayed stereotypically, adding to the long-standing criticism of the invisibility and stereotyping of minorities in advertising (e.g., Eisend, Muldrow, and Rosengren 2023; Grau and Zotos 2016). The underrepresentation and stereotyping of ethnic minorities jeopardize inclusivity and diversity in advertising, leading to misperceptions and negative views of ethnic minorities and cultivating prejudices and negative values in society, thereby increasing discrimination and eventually harming the well-being of ethnic minority consumers. Ethnicity also plays a major role in whether and how consumers respond to company communication (e.g., Bozkurt, Gligor, and Hollebeek 2021). The diversity literature suffers from several shortcomings, leading to conflicting views or unjustified beliefs about ethnic minorities in advertising, leaving researchers and practitioners with three unanswered but important questions. First, though advertising receives criticism about the lack of (ethnic) diversity, some authors have seen positive developments, and more recent content analyses have indicated that the overall representation of ethnic minority groups has been increasing over time (e.g., Taylor et al. 2019). Though the commonly held belief is that advertising becomes more diverse, an alternative interpretation is that any change simply reflects the increase in ethnic minority populations in many societies that are becoming more culturally diverse. Second, most research is Western centered, leading to a narrow view of ethnic diversity, and we do not know whether the findings apply to other contexts, in particular, non-Western societies where white people are a minority. Third, all prior content analysis studies have investigated minority endorsers in print or TV advertising, and we lack insights into portrayals of ethnic minorities in new media formats that differ in many aspects from their traditional counterparts, for instance, by addressing mostly younger groups of consumers (Roth-Cohen, Kanevska, and Eisend 2023). As a result, the literature is based on a Western-centric, traditional media-based commonly held, but empirically unsubstantiated, belief about the progress we have made toward more diversity and inclusion of ethnic minorities in advertising.

In questioning these beliefs, the present study follows an exploratory approach, builds on extant research on ethnic minorities in advertising and social role theory, and provides two empirical studies. First, 34 content analysis papers providing 116 estimates dealing with the representation of ethnic minorities in advertising were meta-analyzed to investigate and explain changes in ethnic minority representation over time and across cultures, thus addressing changes over time and across cultures. None of the prior content analysis studies have investigated new media channels. To address the differences between media, we conducted a content analysis based on a random sample of 1,130 advertisements with 2,861 human endorsers in three different types of media (television, print, and YouTube). The findings empirically substantiate our knowledge of and question commonly held beliefs about changes in ethnic minority depictions over time, across cultures, and between media types.

This study contributes to the literature in several ways. First, by considering prior research studies and using a measure of the percentage of minorities in advertising in relation to the ethnic makeup of the population, we quantitatively assess developments over time and provide evidence that no progress has been made regarding the representation of ethnic minorities in advertising but that developments simply reflect societal changes. Second, we measure the differences between Eastern and Western cultures and show that the development of ethnic minority representations in advertising differs between these cultures, thus rejecting the commonly held Western-centric beliefs about ethnic diversity in advertising. Third, we investigate the representation of minorities in different media types and extend prior knowledge about the portrayals of ethnic minorities based exclusively on portrayals in old media. These results have implications for advertisers and policymakers and are relevant for societies that strive for diversity and inclusion, particularly regarding different ethnic groups within a country.

Prior Research and Theoretical Background

A minority group is a sociological category that differs from the dominant social majority in a society based on personal characteristics, such as gender, race or ethnicity, religion, and age. Many studies on the representation and portrayal of ethnic and racial minorities in advertising (e.g., Bailey 2006; Bowen and Schmid 1997; Knobloch-Westerwick and Coates 2006) have revealed that ethnic minority endorsers are often underrepresented and portrayed stereotypically. In particular, the stereotypes of whites and Asians are mainly positive, whereas ethnic groups such as blacks and Hispanics are often depicted in inferior and unfavorable roles. Whites and Asians are often included in advertisements for high-value and sophisticated products, such as electronics and financial products, whereas blacks and Hispanics are often depicted in advertisements for food products and clothes (Davidson and Schumann 2005; Maher et al. 2008; Taylor and Bang 1997; Taylor, Lee, and Stern 1995; Taylor and Stern 1997). More recent literature suggests that progress has been made toward an increasing and more realistic representation of ethnic minority groups (Eisend, Muldrow, and Rosengren 2023; Taylor et al. 2019), but it is unclear whether any increase in representation is catching up with increasing ethnic diversity in many societies.

Even though the representation of ethnic minorities might have improved, subtle stereotypes persist (Taylor et al. 2019). For instance, though black endorsers are also depicted as celebrities, successful entertainers, or professional athletes, these depictions stereotype blacks as possessing genetically endowed skills (such as muscles and speed; Staples and Jones 1985), whereas the white majority is stereotyped as intellectually superior (Bristor, Lee, and Hunt 1995). The same holds true for Asian Americans, who are continuously depicted as well-educated, hardworking, and primarily in business settings. Prior research is Western-centered and neglects the situation in countries where white people are a minority group (e.g., in many Eastern countries). White people might be portrayed and represented similar to other minorities when they belong to a minority group in a country; that is, as underrepresented and stereotyped as inferior. However, the idea of white supremacy (i.e., whites being at the top of an ethnic hierarchy) suggests that the favorable white stereotype prevails globally, and white minorities are overrepresented and stereotyped as superior (Thomas, Johnson, and Grier 2023). Furthermore, many global brands are traditionally Western brands and might have employed the same advertisements in Eastern countries as in their home country, featuring mostly white people. Until now, only a few studies have investigated the portrayal of ethnic minorities, including white minorities in non-Western cultural contexts with a non-white majority, indicating that white models tend to be overrepresented (e.g., Prieler, Ivanov, and Hagiwara 2022).

Not only are beliefs about changes over time and across cultures unsubstantiated but differences in the media environment are also ignored. So far, no content analysis has investigated ethnic minorities outside print or TV media. Online formats are mainly used by young people, thus addressing a younger audience (Belanche, Flavián, and Pérez-Rueda 2017) with a large variety of content, feedback options, and technology to personalize advertising (De Keyzer, Dens, and De Pelsmacker 2015). A more progressive audience and the implementation of open-to-feedback formats in advertising in online media might favor more diversity, thus leading to improved representations and portrayals of ethnic minorities.

Social role theory provides a theoretical framework explaining the potential differences and changes in ethnic minority representations in advertising over time, between cultures, and in different media. Although developed for and applied in the context of gender roles and stereotyping (Eagly 1987; Schneider and Bos 2019), social role theory has been extended to a general theory of the stereotypes of different social groups (Koenig and Eagly 2014). Social role theory postulates that we observe that members of a group occupy particular social roles more often than members of other groups do and thus generalize to the entire group, leading to group stereotypes. The social roles of ethnic minorities and majority groups have historically originated from differences in occupation and status. For instance, ethnic minorities have immigrated to mostly Western countries and accepted low-wage jobs as unskilled or blue-collar workers, leading to typical social roles that differ from the roles of the white majority, such as business professionals (Koenig and Eagly 2014). Because advertising shows particular social roles (e.g., white-collar workers) more often than other roles (e.g., blue-collar workers; Paulson and O’Guinn 2012), ethnic minorities are underrepresented in advertising. Social role theory suggests that these roles are malleable, vary across cultures, and change over time (Diekman and Eagly 2000). For instance, immigration policies in many countries try to attract skilled workers from foreign countries to take over well-paid, high-status jobs (Hatton 2016). These policies, along with an increasing diversity, equity, and inclusion policy in many societies (Licsandru and Cui 2019), have led to changing social roles that ethnic minority members occupy, and differences in social roles held by ethnic minorities and the majority have eroded. Due to the similarity in social roles that ethnic minorities and the majority nowadays hold, we would expect more ethnic minority endorsers, and the appearance of ethnic minorities in advertising should resemble the population of ethnic minorities in society. Similar arguments extend to differences in cultures that experience different immigration rates (e.g., European countries have experienced strong increases in immigration due to war refugees during the last years) as well as media formats that differ in terms of the diversity awareness of their audience; for instance, online media target younger, more global, and more diversity-aware audiences (Roth-Cohen, Kanevska, and Eisend 2023). If the differences in social roles between ethnic minority and majority groups have become smaller, they should have led to less stereotyped roles and more realistic representations of ethnic minorities in advertising. However, if ethnic minority representation has only increased to the same extent as ethnic minority groups in the general population—that is, the representation gap continues to exist despite changing social roles—this would indicate strong evidence for ongoing stereotyping of ethnic minorities, because they continue to be underrepresented even though they occupy increasingly more social roles that are commonly portrayed in advertising. Based on this theory, we apply an exploratory research approach to analyze alternative explanations and propose the following research question:

Research Question: How did the representation of ethnic minorities in advertising develop over time, differ between cultures, and differ between traditional and online media formats?

Study 1, a meta-analysis, investigated changes in ethnic minority representation over time and across cultures, and Study 2, a content analysis, investigated differences in representation and stereotyping of ethnic minorities due to media formats.

Meta-Analysis

Paper Retrieval and Data

We collected data from content analysis papers that provided percentage figures on the representation of ethnic minorities in advertising in a country at a specific time. To identify relevant papers, we performed a keyword search of electronic databases (e.g., Google Scholar, Business Source Complete) using “advertis*” combined with “ethnic*,” “rac*,” “minorit*,” “divers*,” “African,” “Black,” “Asian,” “Hispanic,” “Latino,” “stereotyp*,” “presence,” “representation,” “portrayal,” and “content analy*” as keywords. Once a paper was identified, the reference list was searched for further publications. We selected content analysis studies only if they provided percentage figures of ethnic minority endorsers in advertising based on a random sample of advertisements from the media that could be considered representative of a particular country at a particular time. We excluded content analyses based on nonrandomly selected media, typically purposive samples of media outlets targeted at specific minority groups. We further excluded two papers that did not provide any information on the sampling year, which is a detail we needed for our analysis. We identified 34 papers that were appropriate for our analysis. Most papers provided results for more than one sample (e.g., for different years of data collection or for different countries) and for more than one minority group, which we treated as independent estimates. Overall, the 34 papers published between 1977 and 2022 provided 116 different estimates of the percentage of ethnic minorities in advertising (the supplemental online appendix provides an overview of all papers and the number of estimates in each paper).

Variables and Analytical Procedure

To capture the representation of ethnic minorities in advertising, which is our dependent variable, we use two different variables. The percentage of minority endorsers captures changes in the representation of ethnic minorities in advertising, but there is no benchmark that considers that the share of ethnic minorities in the general population also varies and changes over time. Therefore, we collected the percentage of minorities in the population of a particular country in the year of data collection from a country’s statistical office (e.g., Statistics Bureau Japan 2023; U.S. Census Bureau 2023). By dividing the percentage of minority endorsers in advertising by the percentage of the minority in the population, we computed a minority ratio variable that indicates whether ethnic minorities were either underrepresented in advertising (if the ratio was smaller than 1) or overrepresented (if the ratio was greater than 1). For nine estimates, we could not retrieve the population figures; therefore, they were not considered in the respective analyses. The minority ratio measure is a real-world standard that assesses whether advertising depicts diversity in a country as is (e.g., represents the minority group populations in a country) but is not an ideal standard that shows diversity as it should be (e.g., all ethnic groups should be represented regardless of population size; Eisend 2022). Real-world standards are easier to quantify and apply, particularly when developments over time are assessed.

As an independent variable that captures changes over time, we used the sampling year; that is, the year in which the data were collected. The variable region captures cultural differences and distinguishes between Western countries (including the United States, Canada, and European countries [the Netherlands and the UK]) and Eastern countries (including Hong Kong, Japan, and South Korea). To account for other influencing factors that show variation in our data and provide potential alternative explanations for the variation in the representation of ethnic minorities, we coded the following variables. The variable minority group distinguishes between studies that have coded either all and thus multiple ethnic minority groups or only a single ethnic minority group. The variable white identifies estimates from studies that have investigated white people as a minority group. The variable medium distinguishes between TV and print advertising.

We ran linear regression models in which we regressed either the percentage of minority endorsers or the minority ratio on sampling year, region, minority group, white, and medium. Furthermore, we added an interaction effect between the sampling year and region to identify country differences as potential boundary conditions for time-related changes, if any.

Results

The results in Table 1 show that the percentages of minority endorsers have increased over time, are smaller if the estimates refer only to specific minorities, are smaller in print than in TV advertising, and are larger in Eastern than in Western countries. The interaction term is marginally significant and shows that the positive effect over time applies to Western countries (t = 3.187, p = .002) but not to Eastern countries (t = −1.348, p = .181).

Table 1. Explaining changes in the representation of minority endorsers in advertising: Regression results.

	Percentage of minority endorsers (mean = 10.63, SD = 10.64)	Minority ratio (mean = 9.38, SD = 47.75)
Intercept	−673.50 (219.85)**	−133.62 (580.06)
Sampling year (continuous [mean = 2,000.15, SD = 10.04])	0.35 (0.11)**	0.07 (0.29)
Minority group (0 = all minorities [n = 24], 1 = specific minority [n = 99])	−9.92 (2.36)***	−3.09 (6.33)
White (0 = other minorities [n = 119], 1 = white as minority [n = 4])	6.60 (5.20)	224.56 (14.53)***
Medium (0 = TV [n = 57], 1 = print [n = 67])	−3.71 (2.12)^+	−0.35 (5.55)
Region (0 = Western country [n = 103], 1 = Eastern country [n = 21])	2,358.60 (1,260.52)^+	13,073.84 (3,789.58)***
Sampling year × Region	−1.18 (0.63)^+	−6.51 (1.89)***
Model fit
F	5.07***	56.46***
R^2	0.22	0.77
Highest variance inflation factor	1.69	1.75
N	116	107

⁺Significant at p < .10. *Significant at p < .05. **Significant at p < .01. ***Significant at p < .001.

The minority ratio relates these findings to societal changes and delivers a more nuanced picture for our interpretation. We do not find any effects of time; the increase in the percentage of ethnic minority endorsers in advertising seems to be fully captured by the fact that the proportion of ethnic minorities in the general population has increased. The effect of region was positive; ethnic minorities in Eastern countries were less likely to be underrepresented than those in Western countries. The interaction between time and region is significant and shows that the minority ratio over time has remained unchanged in Western countries (t = 0.238, p = .812) but has decreased in Eastern countries (t = −3.471, p < .001). Eastern countries tend to overrepresent minorities due to many white endorsers, but this overrepresentation of white endorsers has decreased over time. Supporting this finding, we also found that the minority ratio is higher for white people as a minority, indicating that white minorities are less likely to be underrepresented. This finding is in line with the white supremacy argument. In summary, we did not find any progress in the representation of ethnic minorities over time in Western cultures but did find changes in Eastern cultures.

Content Analysis

Sampling and Data

Content analysis studies are typically restricted to a country context and are therefore difficult to generalize globally in a way that meta-analytic results can do. We selected Germany for our content analysis for several reasons. It is a large advertising market, the second-largest market in Europe after the UK, with a distribution of spending between online and offline advertising that reflects global figures (Statista 2022). YouTube, the online advertising media we choose in our content analysis, has high reach rates for adult audiences (Hootsuite 2021). Although Germany is a Western country, reunification more than 30 years ago has made it a melting pot of Western and Eastern influences. Germany has a long history of immigration of different ethnic groups over time, currently having the highest absolute number of immigrants after the United States (International Organization for Migration 2022). Though this does not allow generalization across country contexts, Germany seems to be a good candidate for content analysis of ethnic minorities, especially regarding the comparison of online and offline media.

The sampling procedure for advertisements was based on previous studies (Bailey 2006; Bowen and Schmid 1997; Taylor et al. 2019). The current content analysis considers both traditional (i.e., print magazines and television) and new advertising media (i.e., YouTube). YouTube is the second most popular social media platform worldwide. It is the only platform that allows access to advertisements without a private account and provides options to control for personalized advertising.

Television Advertisements

We selected two of the most popular private television channels in Germany: RTL and ProSieben (market shares in March 2023 among viewers between 14 and 49: 10.3% [RTL] and 8.6% [ProSieben]; AGF Videoforschung 2023). During a period of 30 days in June 2021, 15 recording sessions were completed every alternate day. The 2- to 3-hour recording sessions were rotated between the morning (9–11 am), afternoon (2–5 pm), and evening (7–10 pm). This approach is in line with prior content analysis (Eisend, Plagemann, and Sollwedel 2014). A total of 1,412 advertisements were recorded. We discarded 926 duplicates, which resulted in 486 advertisements.

Print Advertisements

To cover a wide range of readership, we chose the following genres in line with prior content analyses (e.g., Bowen and Schmid 1997): news, sports, business, technology, lifestyle, beauty (one aimed at women and one aimed at men), automobiles, houses, gardening, and nutrition. For each category, we selected the two magazines with the highest readership. For magazines that appeared more often than once a month, we randomly selected one issue per month. Because some issues were no longer available, we extended the 12-month period and included issues printed during the period from June 2020 to September 2021 (a 14-month period). We included 12 issues per magazine if available. We included all ads that were one-third or more of a page or larger (Bailey 2006; Knobloch-Westerwick and Coates 2006). In total, 2,529 advertisements were obtained. After discarding 419 duplicates, 2,110 unique advertisements were included in the sample.

YouTube Advertisements

During the period from June 1 to June 30 and June 7 to July 9, 2021, we recorded pre-roll advertisements that appeared before the content viewers planned to watch. These ads require a minimum of 5 seconds’ exposure time before one can skip to the content. Because most advertisements are targeted, it is difficult to draw a fully representative pool of pre-roll advertisements. We applied the procedures for YouTube ad sampling provided in literature to draw a semi-representative set of ads for predefined YouTube users (Roth-Cohen, Kanevska, and Eisend 2023). To minimize personalization and mimic the average YouTube consumer, we created a playlist of trending videos in the “Best of YouTube” categories “music,” gaming, “news,” and “movies.” Because they received the highest number of views, this content is likely to reach the majority of YouTube viewers. To avoid categorization of our activities (e.g., if we start watching the trending videos in the category gaming first, we will be categorized as a “gamer”), each playlist was composed in a predefined order, rotating between different categories. At the beginning of every session, YouTube was accessed from an Incognito browser, and we cleaned the watching history, cookies, and caches to avoid personalization. Similar to the television recording, the 2- to 3-hour sessions were rotated between the morning (9–11 am), afternoon (2–5 pm), and evening (7–10 pm). Because geo-targeting cannot be avoided on YouTube, the recordings were conducted in different cities and villages in Germany. In total, we recorded 2,456 ads. After excluding 1,978 duplicates, 478 unique ads were retained.

After recording and data collection, advertisements containing non-human characters (i.e., animated characters, cartoon characters), music video advertisements, movies, TV trailers, and self-promotion ads were excluded (Bristor, Lee, and Hunt 1995; Knobloch-Westerwick and Coates 2006; Taylor et al. 2019). Up to ten characters portrayed in each advertisement were coded. After applying all specified exclusion criteria, 1,130 ads with 2,861 human characters (1,224 characters in TV ads, 672 in print ads, and 965 in YouTube ads) were included.

Coding Procedure, Measures, and Reliability

Table 2 provides an overview of the variables and their coding values. Two graduate students (one ethnic minority, one ethnic majority, one female, and one male) coded all of the advertisements independently. One of the authors thoroughly trained the coders. Intercoder reliability was assessed using Cohen’s kappa. The reliability coefficients in Table 2 indicate a satisfactory intercoder agreement. To resolve disagreements between coders, an independent third coder was recruited to reconcile inconsistencies.

Table 2. Content analysis coding scheme.

Variable	Variable values	Prior studies that applied the coding	Agreement rate %	Cohen’s kappa
Ethnic minority	1 = No 2 = Yes 3 = Cannot tell	Adapted from Kunz and Fleras (1998), MacGregor (1989)	85.58	0.67
Ethnicity	1 = Caucasian or White 2 = Black 3 = Asian 4 = Mediterranean European/Arabic or Middle Eastern 5 = South or Latin American 6 = Other (please specify)	Adapted from Daalmans and Odink (2019), Taylor et al. (2019)	83.60	0.67
Gender	1 = Female (cisgender) 2 = Male (cisgender) 3 = Transgender (LGBTQIA+) characters 4 = Child or infant 5 = Cannot tell	Adapted from Miller (2017), Taylor et al. (2019)	97.31	0.95
Setting	1 = Business setting 2 = Home setting (indoor or outdoor) 3 = Outdoor or natural scenery 4 = Social setting outside the home 5 = Multiple settings 6 = Other	Adapted from Taylor et al. (2019)	70.99	0.63
Occupation	1 = White collar or professional 2 = Middle-class working 3 = Blue collar or laborer 4 = Family or kids 5 = Celebrity 6 = Other 7 = Unknown	Bowen and Schmid (1997)	77.47	0.62
Interaction	1 = No, just one ethnic group 2 = Yes, members from two ethnic groups 3 = Yes, members from three ethnic groups 4 = Yes, members from four or more ethnic groups 5 = I cannot tell	Adapted from Bang and Reece (2003)	73.22	0.60
Product value	1 = Inexpensive or low-end 2 = Expensive or high-end 3 = Unknown or service	Tanner and Lee (2005)	77.39	0.62
Product type	0 = Nontechnical 1 = Technical	Taylor et al. (2019)	91.17	0.82

Results

The descriptive results in Table 3 are presented in cross-tabs, distinguishing between ethnic minority and majority characters for the whole sample, as well as separately for three different types of media. Chi-square tests were applied to test whether the categories of the variables differed between the minority and majority groups. A significant test result indicated that the depiction of the ethnic minority differed from that of the ethnic majority regarding the respective variables.

Table 3. Characteristics of ethnic minorities and majorities in advertisements across the three media types.

	TV ads (1,224 characters, 378 ads)			Print ads (672 characters, 447 ads)			YouTube ads (965 characters, 305 ads)			All ads (2,861 characters, 1,130 ads)
Variables	Minority % (n = 435^a)	Majority % (n = 785)	χ^2	Minority % (n = 125)	Majority % (n = 547)	χ^2	Minority % (n = 335)	Majority % (n = 630)	χ^2	Minority % (n = 894)	Majority % (n = 1,962)	χ^2
Gender
Female	52.1	42.9	21.31**	49.6	47.3	0.60	50.7	38.3	16.07**	51.2	42.7	26.07***
Male	39.9	44.2		42.4	45.9		40.6	53.8		40.5	47.8
Non-binary	0.7	0.0		0.0	0.0		0.3	0.2		0.4	0.1
Child or infant	7.4	12.9		8.0	6.8		8.4	7.8		7.8	9.5
Setting
Business	8.5	13.1	29.81***	3.2	7.3	14.19*	22.7	22.9	10.78	13.1	14.6	39.70***
Home	36.2	39.7		24.0	20.1		25.7	32.2		30.5	31.9
Outdoor	24.2	24.5		36.8	27.1		21.2	21.0		24.8	24.1
Social	7.1	4.5		5.6	2.6		11.3	8.7		8.5	5.3
Sports	7.1	3.6		0.8	1.6		4.8	2.4		5.4	2.7
Other^b	16.8	14.6		29.6	41.3		14.3	12.8		17.7	21.5
Occupation
White collar or professional	2.1	3.6	26.62**	3.2	9.9	11.15	4.2	7.8	16.36*	3.0	6.7	39.26***
Middle-class working	4.8	6.5		6.4	7.9		11.6	15.1		7.6	9.7
Blue collar or laborer	1.8	2.6		1.6	0.5		4.2	2.4		2.7	1.9
Family or kids	15.2	22.9		13.6	23.3		9.0	11.3		12.6	16.5
Celebrity	6.9	5		8.0	11.7		4.5	4.9		6.2	6.8
Other or unknown	69.2	59.4		67.2	56.7		66.5	58.5		67.9	58.4
Interaction
No interaction	15.0	48.4	171.79***	60.8	92.0	82.92***	20.3	52.7	100.47****	23.4	61.9	403.92***
Interaction with one other group	36.6	28.7		32.8	6.6		25.7	19.7		32.0	19.6
Interaction with two other groups	31.1	17.1		6.4	1.5		30.7	15.9		27.5	12.3
Interaction with four or more other groups	17.3	5.9		0.0	0.0		23.3	11.7		17.1	6.1
Product value^c
Inexpensive or low-end	57.4	59.4		43.7	58.9		52.2	47.8		52.8	56.0
Expensive or high-end	32.7	30.0		45.6	26.1		38.4	33.7		37.3	29.6
Unknown or service	9.9	10.5		10.7	14.9		9.4	18.4		9.9	14.4
Product type^c
Nontechnical	60.5	60.4		53.4	58.1		39.9	36.4		52.6	53.2
Technical	39.5	39.6		46.6	41.9		60.1	63.6		47.4	46.8

^a The figures refer to the number of characters.

^b This category includes multiple settings, no background, and other.

^c The analysis of product value and product type is based on the number of ads, not the number of characters.

*p < .05. **p < .01. ***p < .001.

Ethnic minorities were depicted less frequently in advertising than ethnic majorities (χ² = 71.54, p < .001). Of the 2,861 characters across the three media platforms, 31.3% belonged to an ethnic minority (35.5% of the characters on TV, 34.7% of the characters on YouTube, and 18.6% of the characters in print media). The 2021 micro-census revealed that 27.2% of the German population had a migration background (Federal Statistical Office 2021). Note that these figures are proxies for ethnic minorities, because they include white immigrants. For historical reasons, the German government has not collected population statistics based on ethnicity or race (Federal Ministry of the Interior and Community 2018). Based on these figures, the representation exceeds the actual number of people who belong to an ethnic minority across all ads and on TV and YouTube ads but not in print media. One reason for the differences between media formats seems to be a younger and thus more diversity-aware audience in online media and most TV stations compared to print media (Deloitte 2019).

Regarding the variables describing differences in portrayals, thus showing the potential stereotyping of ethnic minorities, we found several systematic differences between ethnic minorities and majority characters that also differed across media platforms.

Gender

Ethnic minorities are more likely to be female than male. This pattern is similar in TV and YouTube ads but not in print ads.

Setting

Overall, ethnic minorities are more likely to be depicted in social and sports or gym settings (1.6 and 2 times more often, respectively) than ethnic majority characters, and the figures for business and home settings are very similar. Though there are no significant differences in settings in YouTube advertising, print, and TV, most characters are more likely to be in business settings than ethnic minorities, showing that stereotyping is more prevalent in traditional media.

Occupation

The majority, compared to minority, characters are more likely to be depicted as white-collar workers or professionals. This difference did not appear in print ads but did appear in TV and YouTube ads. Compared to minority characters, majority endorsers are more likely to be shown with children or families.

Interaction

The variable shows the largest differences overall and in all media types, indicating that majority endorsers are more likely to be depicted with no interaction, whereas minority endorsers are more likely to be depicted in interactions with at least one other group.

Product Value and Product Type

Majority and minority characters are equally likely to advertise high- and low-end products as well as technical and nontechnical products.

Overall, the way ethnic minorities are portrayed still reveals stereotypes: Ethnic minorities are more likely to be female, less likely to be depicted in business settings, less likely to be white collar workers, and more likely to be depicted in interaction with other groups. There were no differences regarding products. The differences across media platforms are minor and do not show any clear trend that distinguishes stereotyping in traditional and new media.

Discussion

The aim of this study was to question commonly held beliefs about the progress advertising makes toward more diversity and inclusion regarding the representation and portrayal of ethnic minorities. Though the literature suggests that progress has occurred (e.g., Eisend, Muldrow, and Rosengren 2023), our quantitative analysis refutes this belief and shows a more nuanced picture. Though the representation of ethnic minority endorsers in advertising has increased over time, this reflects the fact that ethnic minorities in the general population have also increased. Diversity improvements in representation occurred only in Eastern societies, where the overrepresentation of white endorsers has decreased, but not in Western societies. We found that in online media as well as on TV, the underrepresentation of ethnic minorities no longer seems to be an issue but prevails in more traditional print media. Regarding portrayals, we found that stereotyping of ethnic minorities regarding product use no longer exists, indicating that advertisers have achieved diversity and inclusion of ethnic minorities in their specific roles as consumers. However, stereotypes regarding settings, occupations, and interactions still prevailed.

From a social role theory perspective, eroding differences in social roles held by ethnic minorities and majority groups would predict an increasing representation of ethnic minorities in advertising, but our findings provide strong evidence of ongoing and even increasing stereotyping because ethnic minorities continue to be underrepresented despite occupying social roles that are commonly portrayed in advertising. Thus, the social role perspective supports the ongoing stereotyping of ethnic minorities who are excluded because of their ethnicity. However, though specific role expectations regarding occupation, setting, and interaction indicate ongoing stereotyping, because they seem more strongly rooted in historically developed ethnic stereotypes, social role adaptions occur for product categories. Such adaptions provide advantages for marketers, such as a higher readiness of minorities to act or accept changes and innovations (Aaker and Akutsu 2009), which explains why advertisers have been more open and willing to reduce stereotypical portrayals of ethnic minorities regarding product categories.

Cultural differences that indicate a reduction in the overrepresentation of white minorities in Eastern cultures reduce white supremacy in advertising (Thomas, Johnson, and Grier 2023). However, the reduction occurs in favor of an increased representation of the majority group in Eastern societies and not in favor of the representation of other disadvantaged minorities, which jeopardizes its contribution to increased ethnic diversity and inclusion in advertising globally.

The findings regarding the differences between media formats are encouraging regarding the representation of ethnic minorities in new media formats that reach a broader audience than traditional media formats. However, the findings regarding stereotyping paint a mixed picture, similar to prior studies (Roth-Cohen, Kanevska, and Eisend 2023), indicating that stereotyping is reduced only when it benefits advertisers.

The study has methodological and substantial implications for researchers. Our approach to meta-analyzing content analysis studies has only one predecessor (Eisend 2010) but is the first to compute a ratio value that describes the difference in representation of an ethnic minority in advertising compared to reality (i.e., the population value). The real-world benchmark captures whether advertising describes diversity as is in the real world (Eisend 2022) and can be applied in further studies that attempt to describe the temporal development of the representation of minorities or disadvantaged groups (e.g., older people, females, homosexuals) to quantitatively assess whether any changes have occurred. A substantial implication is to continue with research on ethnic minorities in different countries and across different media platforms, because stereotyping prevails but their patterns change, and these changes need to be monitored for a better understanding of the topic.

Though advertisers, especially in many Western societies, seem open to more inclusive ads and representations, our findings jeopardize the conclusion that advertising has, in general, become more diverse and inclusive, and “woke advertising” with respect to social and political issues regarding ethnic minorities is not supported by our findings. Consumers and public policy are advised to use stricter criteria and quantitative benchmarks to evaluate the progress of advertising toward diversity and inclusion. The way in which ethnic minorities are depicted remains stereotypical. A more critical evaluation of woke advertising is needed to avoid marketers benefiting from seemingly prosocial activities that do not seem to have relevant social effects.

Disclosure Statement

The authors have no conflicts of interest to declare.

Additional information

Funding

This research received financial support from a research grant by the German Research Foundation (DFG), Germany (Grant No. EI 508/21-1).

Notes on contributors

Anna Rößner

Anna Rößner (Ph.D., European University Viadrina) is a postdoctoral researcher, European University Viadrina.

Martin Eisend

Martin Eisend (Ph.D., Free University Berlin) is a professor of marketing, European University Viadrina, and an adjunct professor of marketing, Copenhagen Business School.