How is gender-discriminatory content measured on social media? A systematic review revealing a critical gap in measurement tools

1 Introduction

Gender discrimination represents one of the most persistent forms of social inequality, manifesting through practices, attitudes, and behaviors that systematically disadvantage women as a social group (Klonoff and Landrine, 1995). These patterns lead to structural inequalities in the workplace, education, and personal relationships, and they raise the risk for women of harassment and violence throughout their lives.

From a socio-psychological perspective, gender discrimination goes beyond isolated prejudiced attitudes or occasional acts of hostility, but can be seen as a systemic phenomenon maintained by culturally shared beliefs, normative expectations, and institutionalized power imbalances that justify the subordinate status of women as a social group (Glick and Fiske, 1996; Nye et al., 2009; Bareket and Fiske, 2023).

Despite significant progress, empirical evidence consistently shows that gender discrimination remains widespread in modern societies, including among younger generations. For example, recent European data indicate ongoing support for traditional gender roles, different moral expectations for women and men, and acceptance of behaviors that limit women's autonomy and social value (European Institute for Gender Equality, 2023; OECD, 2023).

Within this broader framework, gender discrimination can be expressed through multiple, conceptually distinct manifestations. These include sexism, gender stereotyping, sexual objectification, and sexualization. Although these constructs differ in their theoretical definitions and operationalizations, they share the common feature of reinforcing unequal gender relations and contributing to the social devaluation of women. In contemporary digital environments, these manifestations frequently coexist and overlap within media content. Accordingly, the present review does not treat sexism, gender stereotyping, sexual objectification, and sexualization as equivalent constructs. Rather, they are conceptualized as related manifestations of a broader domain of gender-discriminatory content toward women. Thus, the common denominator across included studies is not the construct assessed (e.g., sexism, stereotyping, objectification, or sexualization), but the attempt to operationalize exposure to gender-discriminatory social media content.

Specifically, sexism represents one of its most extensively studied expressions, referring specifically to prejudiced attitudes and discriminatory behaviors grounded in gender-based stereotypes (Glick and Fiske, 1996; Nye et al., 2009).

Ambivalent sexism theory provides a useful framework for understanding how these inequalities are maintained through two complementary components: hostile sexism, which openly penalizes women who challenge traditional gender hierarchies, and benevolent sexism, which idealizes women who conform to conventional roles while reinforcing their dependence and subordination (Glick and Fiske, 2001, 2011; Bareket and Fiske, 2023; Ward and Grower, 2020). These beliefs are embedded in broader cultural narratives about femininity, sexuality, and relational roles and contribute to double standards that reward male agency while constraining female autonomy (Endendijk et al., 2020; Leaper and Gutierrez, 2024).

Sexual objectification represents another central expression of gender discrimination, involving the reduction of women to their bodies or sexual attributes and the prioritization of appearance over competence, agency, or voice (Fredrickson and Roberts, 1997). Sexualization is a related but broader process in which a person's value is primarily derived from sexual appeal or behavior, physical attractiveness is equated with sexual desirability, or the person is treated as an object for others' sexual use American Psychological Association Task Force on the Sexualization of Girls, 2008. In digital environments, sexual objectification and sexualization frequently intersect, as women's visibility and social approval may become tied to conformity with sexualized aesthetic standards.

These processes are not confined to traditional media. Social media may intensify them by encouraging self-presentation practices in which sexualized content is rewarded through metrics of visibility and approval, such as likes and follower counts. Consequently, objectification may be not only externally imposed but also internalized and reproduced through participation in platform norms, particularly among adolescents and young adults (Choukas-Bradley et al., 2022; Tiggemann, 2022).

Gender stereotyping refers to the attribution of specific traits, characteristics, and social roles to individuals based on their gender (Eagly and Wood, 2012; Ellemers, 2018). These stereotypes include both descriptive beliefs about how women and men are perceived to be and prescriptive beliefs about how they are expected to behave (Heilman, 2012). By reinforcing traditional gender roles and expectations, gender stereotypes contribute to the maintenance of gender inequalities across social contexts. In online environments, stereotypical representations of women are frequently reproduced through social media content, advertising, influencer communication, and user-generated content, potentially shaping users' perceptions of gender norms and social roles (Lee and Yuan, 2023).

Although gender discrimination predates digital media, social media do not merely reproduce offline inequalities. Their affordances reshape how discriminatory representations are produced, circulated, validated, and interpreted, altering their visibility and normative status in everyday interaction (Megarry, 2014; Semenzin and Bainotti, 2020; Dutta et al., 2025; He, 2025; Villacampa and Méndez-Hernández, 2026).

Several affordances distinguish the online circulation of gender-discriminatory content from offline encounters. First, scalability and persistence allow such content to reach large audiences and remain accessible over time (Dutta et al., 2025). Second, recommender systems may amplify content that conforms to sexualized or stereotypical norms, increasing its repeated visibility among young users (Regehr et al., 2025). Third, discriminatory messages may be conveyed through humor, irony, trends, or aestheticized forms of self-presentation, making them less readily recognizable as discriminatory (Megarry, 2014; Sills et al., 2016). Finally, interactive metrics such as likes, comments, and shares provide immediate social validation and may reinforce gendered norms through peer approval.

Together, these affordances may increase the visibility and routine circulation of gender-discriminatory content, while also making it more difficult to identify and measure consistently (Sills et al., 2016; Dutta et al., 2025).

A central feature of online gender discrimination is the routine circulation of content that represents women in stereotypical, sexualized, or degrading ways. Because such content is frequently embedded in humor, irony, trends, aesthetic norms, or apparently individual forms of self-expression, it may be experienced as ordinary rather than explicitly discriminatory (Megarry, 2014; Sills et al., 2016). This normalization can obscure its structural implications and reduce the likelihood of critical recognition and engagement (Megarry, 2014; Villacampa and Méndez-Hernández, 2026).

Repeated exposure to stereotypical, sexualized, or objectifying representations has been associated with self-objectification, body surveillance, internalization of gender norms, endorsement of traditional gender roles, and reduced sensitivity to gender-based violence (Fredrickson and Roberts, 1997; Karsay et al., 2018; Rollero et al., 2021; Rollero and Tartaglia, 2019; Karsay et al., 2023; Nassen et al., 2024). These findings indicate that the consequences of exposure extend beyond immediate reactions to individual posts and may contribute to the broader normalization of unequal gender relations.

Adolescents and young adults warrant particular attention because these developmental stages involve heightened sensitivity to peer evaluation, social norms, identity formation, and body image (Leaper and Brown, 2018; Rogers et al., 2022). Social media are central contexts in which meanings related to gender, sexuality, and social values are negotiated during these years. Longitudinal evidence also indicates that girls report higher levels of emotional difficulties than boys from early adolescence, with this gap widening over time (Yoon et al., 2023).

Despite the developmental relevance of these processes, the available measurement landscape remains unclear. Previous reviews have documented fragmentation in the quantification of online sexism and in automatic sexism-detection approaches (Dutta et al., 2025; Lei et al., 2024), but they do not provide a focused mapping of the instruments used to assess gender-discriminatory content among adolescents and young adults. Without such a mapping, it remains difficult to compare findings across studies, monitor changes over time, and develop evidence-based preventive or educational interventions.

Moreover a critical issue emerging from the broader literature on digital media measurement concerns the distinction between different levels at which exposure can be assessed. Exposure to online content may refer to: (a) perceived exposure, namely individuals' subjective reports of the frequency or intensity with which they encounter specific content; (b) actual exposure, referring to objectively recorded behavioral traces such as logged platform activity, browsing histories, or digital trace data; and (c) content properties, namely the objective characteristics of the content itself, assessed through content analysis procedures or coding frameworks. These levels capture related but conceptually distinct phenomena and should not be considered interchangeable (De Vreese and Neijens, 2016; Parry et al., 2021; Valkenburg, 2022). In particular, evidence suggests that self-reported media exposure often diverges substantially from objectively logged behavior (Scharkow, 2016; Ellis et al., 2019; Parry et al., 2021). Consequently, understanding how gender-discriminatory content is measured requires distinguishing whether studies assess users' perceptions of exposure, their actual exposure, or the characteristics of the content circulating within digital environments (Araujo et al., 2017; Guess et al., 2023; Riffe et al., 2019).

1.1 Aim

The primary objective of this systematic review is to provide a comprehensive and theoretically informed examination of the instruments developed to assess exposure to social media content that stereotypes and/or discriminates against women, including expressions such as sexism, gender stereotyping, sexual objectification, and sexualization. Rather than treating these phenomena as isolated constructs, the review conceptualizes them as multiple expressions of a broader system of gender discrimination and examines how this system is operationalized at the level of exposure in digital environments. In doing so, the review considers the full range of measurement approaches documented in the literature, including both self-report instruments assessing perceived exposure and tools designed for direct content analysis. The review is further guided by a conceptual distinction between perceived exposure, actual exposure, and content properties. Accordingly, particular attention is devoted to identifying which of these dimensions are addressed by existing measurement approaches and which remain underrepresented in the literature. The relative distribution of these approaches across the included studies is itself treated as an empirical finding, reflecting the current state of methodological development in this area.

2 Methods

This systematic review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA; Liberati et al., 2009; Page et al., 2021).

2.1 Inclusion criteria

To be included in this systematic review, studies had to meet the following criteria:

• a) Studies involving social media users, specifically adolescents and/or young adults, were included. Eligible studies focused on users of major social networking platforms, including, but not limited to, Facebook, Instagram, Twitter/X, YouTube, Twitch, Discord, Reddit and similar digital environments, with no restrictions placed on platform type. Moreover, studies examining exposure to media content, including music videos, reality television, or other sexualized formats, distributed and consumed through the named platforms were also considered eligible.
• b) Studies were included if they described, developed, validated, or used instruments designed to assess the presence, circulation, or exposure to online content that stereotypes, discriminates against, sexualizes, or objectifies women. Eligible instruments included, but were not limited to, self-report questionnaires and psychometric scales, standardized or ad hoc measurement tools, quantitative metrics, and content analysis tools and coding schemes, when relevant to assessing sexist or gender-stereotypical online content. The decision to include both self-report measures and content analysis tools was intentional, reflecting the aim of mapping the full landscape of available measurement approaches rather than restricting the review to a single methodological tradition.
• c) Only empirical studies were considered, including quantitative, qualitative, and mixed methods designs. Studies had to be published in peer-reviewed journals and written in English.
• d) Studies published from March 11, 2020, onward were included, corresponding to a period marked by a substantial increase in social media use following the COVID-19 pandemic. This temporal boundary was selected because the COVID-19 pandemic marked a significant turning point in social media use patterns among adolescents and young adults, substantially altering both the volume and the nature of online gender-discriminatory content. Moreover, this temporal boundary was applied to the study context and publication date, not to the instruments themselves; tools developed prior to 2020 are acknowledged where they appear in the included studies, but the primary validation literature for those instruments falls outside the scope of this review.

No restrictions were placed on study design (quantitative, qualitative, or mixed methods) or on the geographical area in which the studies were conducted.

2.2 Exclusion criteria

Studies were excluded if they met any of the following criteria:

• a) Studies did not involve social media users (e.g., studies focusing exclusively on journalism, advertising, television, or other non-interactive media) or focused exclusively on clinical populations, children, or adults over 30 years of age.
• b) Studies were excluded if they were theoretical papers, editorials, commentaries, or opinion pieces without the development or use of empirical measurement tools, and did not involve the development, validation, or application of any measurement instrument or address sexism or gender discrimination without explicit reference to social media or online social platforms.
• c) Studies were excluded if they examined hate speech or other forms of discrimination (e.g., racism, xenophobia, or discrimination against LGBTQIA+ communities) without an explicit and substantive focus on sexism or discrimination against women. This restriction was theoretically motivated by the need for conceptual coherence: instruments developed to assess discrimination against other groups operate within different epistemological and definitional frameworks, and their inclusion would have required integrating constructs that are not directly comparable with the phenomena under review.
• d) The following were excluded: narrative reviews and systematic reviews; gray literature, including theses, dissertations, and non–peer-reviewed reports; conference papers, conference abstracts, letters to the editor, editorials, and books or book chapters.
• f) Studies conducted in disciplines outside Psychology, such as Management, Sport Sciences, Economics, or Virology, were excluded when the focus was not aligned with psychological, psychosocial, or behavioral processes relevant to the study aims.

2.3 Search procedures

Starting from these criteria, we formulated the following search string:

• (adolescent^* OR teen^* OR young^* OR youth^*) AND (social media OR social network^* OR online platform^* OR Facebook OR YouTube OR Twitter OR Instagram OR TikTok OR WhatsApp OR Telegram) AND (tool^* OR scale^* OR questionnaire^* OR instrument^* OR metric^* OR assessment OR inventory OR test^*) AND (sexis^* OR gender OR misogyn^* OR discrimination^*).

The asterisk was used as a truncation operator to retrieve different lexical variants of the search terms, including “sexism” and “sexist,” as well as “misogyny,” “misogynist,” and “misogynistic.” The search was conducted on September 12, 2025, in the electronic databases Scopus, Web of Science, and PsycINFO.

Specifically, in Web of Science, the search was conducted using the Topic field (including title, abstract, and keywords). The initial search returned 4,930 records. Filters were subsequently applied to retain English-language journal articles, resulting in 3,997 records. Further refinement based on relevant research areas (including Psychology, Behavioral Sciences, Sociology, Education Research, Communication, Social Issues, Social Sciences, Women's Studies, Anthropology, Cultural Studies, Development Studies, and related fields) yielded 3,510 records.

In Scopus, searches were conducted using the Title, Abstract, and Keywords fields. Applying the search string and restricting results to publications from 2020 onward produced 462 records. After limiting the results to articles, published and indexed records, and English-language publications, and restricting subject areas to Social Sciences, Psychology (multidisciplinary), and Arts and Humanities, a total of 168 records were retained.

In PsycINFO, the search was conducted using subject terms and applied filters for English-language, published, peer-reviewed journal articles. This search yielded 21 records.

All identified records were imported into Rayyan for screening. The combined database search returned 3,699 records. After duplicate detection, 125 duplicate records were identified and removed. The final dataset consisted of 3,574 records.

2.4 Study selection

After title and abstract screening, 3,505 records were excluded for failing to meet the inclusion criteria. Specifically, studies were excluded because they were classified as “wrong outcome,” “wrong population,” “wrong publication type,” or “wrong date.”

Articles presenting research not aligned with the objective of this systematic review were considered “off topic” (wrong outcome). Specifically, studies that did not measure the spread of online content that stereotypes, sexualizes, or discriminates against women on social media, or that addressed sexism exclusively in offline contexts or without explicit reference to social media platforms, were excluded. Studies focusing on other forms of discrimination (e.g., racism or generic hate speech) without a clear and specific connection to sexism against women were also excluded. Regarding the “wrong population” criterion, articles reporting research conducted exclusively on children, adults over 30, older adults, or clinical populations, as well as studies that did not involve social media users, were excluded.

Studies focusing on occupational groups, such as healthcare workers, educators, or media professionals, whose social media use was examined in a work-related context, rather than as personal users, were also excluded as they did not correspond to the target population defined in the inclusion criteria

Articles classified as “wrong publication type” included narrative or systematic reviews, editorials, commentaries, conference abstracts, conference papers, books or book chapters, and other non-peer-reviewed gray literature. Finally, studies published before March 11, 2020, were excluded, consistent with the timeframe defined in the inclusion criteria.

After this phase, 69 articles were deemed eligible and underwent full-text screening. These studies were then analyzed for content and assessed for methodological quality to determine their suitability for inclusion in the systematic review.

2.5 Extraction of data and quality assessment of studies

The content-based record selection was conducted by two team members with extensive backgrounds in gender studies. The two experts carefully read the 69 articles independently.

To assess inter-rater reliability, Cohen's κ was calculated on the initial independent screening decisions, prior to any discussion or consensus procedures. The resulting k value was 0.95, indicating near-perfect agreement. Disagreements were subsequently resolved through discussion among the research team.

At the end of this process, 61 articles were excluded. As shown in the flow diagram, of these: 54 were excluded because they were deemed “off-topic,” meaning they did not focus on tools that assess exposure to sexist online content but instead addressed related topics; 3 were excluded because the articles focused on populations other than our own; and 4 were excluded because, despite having English titles and abstracts, they were written in a different language.

Once the eight eligible articles were selected based on their content, they were assessed for quality.

Quality assessment was conducted independently by two researchers using the Critical Appraisal Skills Programme (CASP) appraisal tools. Given the methodological heterogeneity of the included studies, the specific CASP checklist corresponding to each study design was applied (e.g., CASP Randomized Controlled Trial Checklist, CASP Cohort Study Checklist, and CASP Qualitative Study Checklist). This approach allowed methodological quality to be assessed according to the characteristics of each individual study design [Critical Appraisal Skills Programme (CASP), 2023].

At the end of the assessment, the two researchers reported a low risk of bias, and Cohen's k was calculated again on the reviewers' initial independent quality ratings before consensus discussions and yielded a value of 1.00, indicating complete agreement.

Therefore, at the end of the risk of bias assessment process, 8 articles were definitively included in this systematic review. The study selection process is illustrated in Figure 1.

2.6 Narrative synthesis

Given the methodological heterogeneity of the included studies in terms of design, populations, outcomes, and measurement instruments, a meta-analysis was not considered appropriate. The findings were therefore synthesized using a structured narrative synthesis approach. Following data extraction, studies were grouped according to their primary measurement strategy: (a) self-report measures of perceived exposure to gender-discriminatory content, (b) structured content-analysis instruments designed to directly assess online content, and (c) qualitative content-analysis approaches. Within each category, studies were compared with respect to the type of construct assessed, measurement procedures, methodological strengths and limitations, and their capacity to directly evaluate gender-discriminatory content. Cross-study patterns were then identified and used to generate the overarching conclusions regarding current measurement practices and methodological gaps in the literature.

3 Results

The systematic review included eight empirical studies published between 2021 and 2024, characterized by marked methodological heterogeneity but united by their interest in gender dynamics and sexualization in digital contexts frequented by adolescents and young adults. The presentation of results follows the narrative synthesis framework described in the Methods section, with studies organized according to their measurement approach to facilitate comparison across methodological traditions.

Details of each study are shown in Table 1.

Table 1. Characteristics of the included studies.

Overall, quantitative cross-sectional survey-based designs predominately, accompanied by a longitudinal study, a randomized controlled trial, a qualitative content analysis, and an observational study to develop and validate a content analysis tool.

The samples mainly included participants aged 13 to 30, with highly variable sizes, ranging from 150 subjects to over 6,000 adolescents, while the platforms considered include Instagram, TikTok, Twitch, Discord, Facebook, Twitter/X, Reddit and, more generally, digital environments with a high visual or interactive component. Despite this variety, a recurring element emerges from the overall analysis: in most cases, social media constitutes the context of exposure or reference for use, rather than the direct object of systematic content measurement.

To facilitate comparison across studies, the included contributions were also examined according to their measurement approach and associated methodological strengths and limitations (see Table 1). Three broad approaches emerged. First, most studies relied on self-report measures of perceived exposure or engagement with gender-related content on social media (Alleva et al., 2024; Chadwick-Brown and Endendijk, 2024; Gil Bermejo et al., 2021; Hernández-Ruiz et al., 2024; Aubrey et al., 2024; Vranken et al., 2023). These studies benefited from the use of validated psychometric instruments and were particularly useful for examining psychological and behavioral correlates of exposure. However, they provided only indirect information about the characteristics of the content itself.

Second, Anciones-Anguita and Checa-Romero (2024) adopted a structured content-analysis approach and developed a dedicated coding instrument to objectively assess sexualization in Twitch livestreams. This represented the most advanced attempt to directly measure online content, although its scope was limited to sexualization and a single platform.

Third, Herbenick et al. (2023) directly analyzed social media content through qualitative coding procedures. While this approach offered rich contextual insights into gendered representations, the absence of a standardized and validated coding framework limited its comparability and replicability.

Overall, the comparison highlights a clear methodological imbalance within the literature: research is largely concentrated on perceived exposure and its outcomes, whereas validated approaches for directly measuring gender-discriminatory content remain scarce.

A closer examination of the included studies shows that the study by Alleva et al. (2024) is the only experimental contribution included in the review that focuses on the effects of exposure to idealized images on Instagram, testing whether cognitive strategies of “protective filtering,” in particular an orientation focused on body functionality, can mitigate the negative effects on body image. In a sample of 150 women aged 18-30, the authors use validated psychometric instruments to assess body appreciation, functional appreciation, body satisfaction, and self-objectification in state form. The results show that, although the emphasis on functionality has protective effects on some indicators of body wellbeing, exposure to content that conforms to dominant aesthetic ideals still increases self-objectification. Methodological rigor, experimental control, and the use of established scales are clear strengths; however, with respect to the objective of this review, the limitation is substantial, as Instagram content is not systematically analyzed or coded, and no tool is developed to measure its dissemination or structural characteristics.

A markedly different approach is adopted by Anciones-Anguita and Checa-Romero (2024), whose contribution is most directly aligned with the objective of this review. The authors analyze 1,920 livestream clips on Twitch across different thematic categories and developed a Sexualization Coding Scale organized into dimensions related to context and behavior, with scores ranging from 0 to 14 and good indices of inter-coder reliability and internal consistency. The results show significantly higher levels of sexualization in content produced by female streamers and in specific categories such as ASMR or Pools/Hot Tubs. The main strength lies in having constructed a replicable, quantitative grid applied to real user-generated content; however, the tool remains limited to the dimension of sexualization, does not include other forms of sexism or objectification, and is validated on a single platform, without a specific focus on adolescents as users. Despite these limitations, this study is the only structured attempt to objectively measure content among those included.

Chadwick-Brown and Endendijk (2024) explore the association between consumption of sexualized content on social media and sexual coercion in a sample of 255 Dutch adolescents aged 16 to 20. Exposure to “sexy posts” on Instagram and Facebook was assessed via self-report measures, along with variables such as sexual double standards and perpetration and victimization behaviors. The results indicate that greater exposure is associated with a higher probability of coercion, with effects modulated by sexual double standards. This study is relevant for understanding the link between digital content and concrete behaviors, but it has a central methodological limitation with respect to the objective of the review: exposure is not verified through direct content analysis but is entirely based on the subjective perception of the participants, and the category of “sexy posts” is not operationalized using systematic coding criteria.

A similar approach is taken by Gil Bermejo et al. (2021), who use a mixed-methods design involving 400 Spanish adolescents aged between 14 and 16 to investigate the link between ambivalent sexism and social media use. Using validated tools such as the Ambivalent Sexism Inventory and Detected Sexism in Adolescents, supplemented by focus groups, the study highlights how levels of sexism are associated with specific digital practices, particularly on visual platforms such as Instagram and Snapchat. Qualitative data highlighted pressures toward stereotypical self-presentation and the internalization of normative gender models. The strength lies in the methodological triangulation and the adolescent focus; however, even in this case, there is no direct analysis of the content, and no tool has been developed to quantify its spread or characteristics.

Herbenick et al. (2023) focus on 316 memes related to choking collected on Twitter/X, Instagram, Facebook, and Reddit, with the aim of exploring how such content contributes to the normalization of potentially dangerous sexual practices, often embedded in gender power dynamics. Through inductive qualitative coding, categories such as “romanticization of violence,” “trivialization of risk,” and “ironic representations of coercion” are identified. The contribution is valuable for the direct analysis of materials circulating online and for understanding the underlying cultural meanings; however, the absence of psychometric validation, quantitative prevalence indicators, and a standardized grid limits its usefulness as a systematic tool for assessing the circulation of such content.

Hernández-Ruiz et al. (2024) developed and validated a 38-item scale to measure perceptions of gender performance and identification with influencers on different platforms, in a sample of 622 adolescents and young adults. Psychometric validation showed good reliability and adaptation indices, and the results highlight a significant differentiation between platforms in the perception of gender models, with Twitch and Discord perceived as more masculinized and Instagram as more open to non-binary expressions. The strength lies in the methodological soundness and multi-platform focus; however, the scale assesses subjective perceptions rather than the actual presence or circulation of sexist content and does not include any direct analysis of online materials.

Aubrey et al. (2024) analyze the association between exposure to sexualized media and sexting behaviors in a large sample of 6,093 US adolescents. Exposure, measured through self-reports, includes online pornography, reality shows, and music videos. The results show strong associations, mediated by attitudes toward sexting. The sample size and statistical robustness are significant strengths; however, the focus was not specifically on social media platforms in the strict sense, and no tool is developed to analyze or quantify digital content. This study was nonetheless retained because the media formats examined are extensively distributed through platforms explicitly named in the inclusion criteria, most notably YouTube, and the measurement approach was considered relevant to the review's objective of mapping tools for assessing exposure to sexualized content in digital environments frequented by adolescents and young adults.

Finally, Vranken et al. (2023) adopt a three-wave longitudinal design on 1,286 young Belgian adults to examine the relationship between exposure to counter-stereotypical content and professional sexism, introducing a Professional Sexism Scale with good reliability indices. The results show that at the between-person level, greater exposure is associated with lower levels of professional sexism, while within-person effects are not stable over time. Longitudinal design is a significant methodological strength; however, even in this case, exposure is self-reported and no systematic analysis of platform content was conducted.

Overall, the integrated analysis of the eight contributions reveals a consistent pattern: the literature focuses mainly on psychological effects, attitudes, or individual perceptions, whereas the development of standardized, validated tools to objectively measure the prevalence and characteristics of sexist, objectifying, or sexualizing content in social media is largely absent. The only partial exception is the study by Anciones-Anguita and Checa-Romero (2024), which provides a promising methodological basis but is still limited in scope and platform. This reveals a clear methodological gap in contemporary literature, which is the main finding of this review.

4 Discussion

This systematic review examined tools and instruments used to assess exposure to social media content that stereotypes and/or discriminates against women, including expressions such as sexism, gender stereotyping, sexual objectification, and sexualization.

Despite growing interest in online gender discrimination, the results highlight a fragmented methodological framework characterized by significant heterogeneity in research approaches.

Across the included studies, quantitative approaches based on surveys clearly predominate. Exposure to gendered content is most often captured through self-reported indicators, and research tends to concentrate on the associations between perceived exposure and a range of psychological or behavioral outcomes, including self-objectification, endorsement of sexist beliefs, internalization of body ideals, and engagement in online sexual behaviors. This orientation is consistent with a well-established trend in the literature on media and sexualization, which has traditionally prioritized the study of individual-level effects over the systematic analysis of media content itself (Karsay et al., 2018; Ward et al., 2023).

A further recurring pattern concerns how exposure to online content is operationalized in empirical research. In most studies, social media are approached primarily as contexts of use rather than as objects of direct content analysis. As a consequence, exposure to sexualized or stereotypical representations is typically inferred through self-reported measures, such as frequency of platform use, time spent online, or participants' subjective perceptions of encountering specific types of content (Parry et al., 2021; Verbeij et al., 2021; Tiggemann, 2022).

This approach is widely adopted in research on digital media effects among adolescents and young adults, where the focus is often placed on the relationship between social media use, perceived exposure, and psychological outcomes such as body image concerns, mental health indicators, or online sexual behaviors (Choukas-Bradley et al., 2022; Valkenburg, 2022; Verbeij et al., 2021). For example, participants are frequently asked to estimate how often they encounter sexualized content or how intensively they engage with particular platforms. While such measures are useful for examining psychological correlates of media experience, they do not allow for a systematic description of the actual characteristics of the content circulating within digital environments.

At the same time, a growing body of research highlights that social media have become one of the primary contexts of socialization during adolescence, shaping perceived norms, relational expectations, and standards of attractiveness (Choukas-Bradley et al., 2020; Nesi et al., 2020; Odgers et al., 2020; Tiggemann and Slater, 2015). Higher levels of use or exposure to specific online practices have been associated with a range of outcomes, including poorer mental health, body dissatisfaction, and engagement in risky sexual behaviors (Karsay et al., 2023). However, even in these studies, the presence of sexualized, stereotypical, or discriminatory content is typically inferred indirectly from users' reported experiences, rather than examined through structured content analysis.

Overall, this pattern reflects a broader tendency in contemporary digital media research: while there is increasing attention to the effects of exposure, methodological tools for systematically measuring the nature, frequency, and distribution of content within social media ecosystems remain relatively underdeveloped. As a result, literature continues to focus primarily on individual perceptions and attitudes, leaving the structural dimension of content production and circulation comparatively underexplored.

Beyond the methodological imbalance identified in the reviewed studies, the findings also highlight the importance of distinguishing between different conceptual levels of exposure assessment. The studies included in this review predominantly measured perceived exposure, relying on participants‘ subjective estimates of how often they encountered sexualized, stereotypical, or discriminatory content on social media. These approaches provide valuable information about users' experiences and interpretations of digital environments, but they do not necessarily correspond to either the content actually encountered or the objective characteristics of that content (Scharkow, 2016; Parry et al., 2021).

A second level, actual exposure, refers to exposure assessed through behavioral or logged data, including digital trace information, platform analytics, browsing histories, passive sensing methodologies, or other forms of objectively recorded online behavior (Araujo et al., 2017; Guess et al., 2023). Notably, none of the studies included in this review adopted such approaches. This absence is particularly relevant because previous research has consistently documented discrepancies between self-reported and objectively measured digital media use, suggesting that perceived exposure and actual exposure may capture different aspects of individuals' online experiences (Scharkow, 2016; Ellis et al., 2019; Parry et al., 2021).

A third level concerns the properties of the content itself. This perspective focuses on the objective assessment of online materials through structured coding schemes, content-analysis procedures, or computational methods designed to identify characteristics of media messages independently of users‘ perceptions (Riffe et al., 2019; De Vreese and Neijens, 2016). Only two studies included in the review directly examined content, and only one proposed a standardized measurement framework. Consequently, current evidence is considerably stronger regarding users' perceptions of exposure than regarding either actual exposure or the prevalence and characteristics of gender-discriminatory content within social media ecosystems.

Framing the literature through these three levels helps clarify the nature of the methodological gap identified in the present review. The limitation is not simply the scarcity of content-analysis tools, but the broader absence of research capable of integrating subjective experiences, objective exposure indicators, and direct assessments of content characteristics within a common measurement framework. Such integration has increasingly been advocated in contemporary digital media research as a necessary step toward understanding how online environments shape attitudes, beliefs, and social behavior (Valkenburg, 2022; Guess et al., 2023).

One of the most salient findings of the present review concerns this methodological gap. As identified through the narrative synthesis, the included studies clustered into three main measurement approaches, allowing the comparison of methodological characteristics across otherwise heterogeneous designs. The structured synthesis of the included studies revealed a clear imbalance across measurement approaches. Most studies relied on self-report measures of perceived exposure, which were primarily designed to investigate psychological and behavioral correlates of social media use. By contrast, direct approaches to content assessment were rare and consisted of either a single structured content-analysis instrument or qualitative coding procedures. While self-report measures offer valuable insights into users' experiences and perceptions, they provide only indirect evidence regarding the characteristics of the content itself. Conversely, direct content-analysis approaches allow the objective assessment of online material but remain underdeveloped and limited in scope. This imbalance helps explain why current evidence is stronger with respect to the consequences of exposure than to the systematic characterization of gender-discriminatory content circulating on social media.

Standardized and validated tools for directly analyzing gender-discriminatory content on social media are largely absent. Among the studies included, only the work by Anciones-Anguita and Checa-Romero (2024) proposes a structured instrument designed to code sexualization in digital content through a content analysis grid applied to livestreams on Twitch. This contribution demonstrates the feasibility of developing replicable tools for the quantitative analysis of user-generated content and represents an important methodological step forward.

However, the scope of this instrument remains limited. It focuses exclusively on sexualization and is validated within a single platform, which restricts its ability to capture the broader range of gender discrimination phenomena present in online environments. These include not only sexualization, but also stereotyping, objectification, and more subtle or normalized forms of sexism that are often embedded in everyday digital practices (Megarry, 2014; Sills et al., 2016).

The predominance of self-reported measures in the remaining studies likely reflects both their relative methodological accessibility and the challenges associated with analyzing digital content, which is highly dynamic, heterogeneous, and shaped by platform algorithms (Dutta et al., 2025; Semenzin and Bainotti, 2020). Nevertheless, reliance on subjective indicators entails important limitations. Perceived exposure may be influenced by cognitive biases, selective recall, or individual interpretations of what constitutes discriminatory content, making it difficult to disentangle actual exposure from subjective experience.

In this sense, the findings suggest the need for more integrated methodological approaches that combine perceived exposure with systematic content analysis. Such integration would allow for a more accurate representation of the informational environments in which users are immersed and would help bridge the gap between subjective experience and the structural properties of digital content. This measurement gap also has important implications from an intervention perspective. Recent review evidence on digital tools and serious games aimed at promoting wellbeing from a gender perspective suggests that this field is expanding among adolescents and young adults, but remains methodologically heterogeneous and often pays limited attention to contextual-level variables and gender-specific processes (Esposito et al., 2024). In this sense, more robust tools for assessing gender-discriminatory content on social media may help bridge intervention research and a more precise understanding of the digital environments in which young people are embedded.

From a theoretical perspective, the results highlight a tension between two levels of analysis that are often treated separately in the literature: individual exposure to media content and the broader cultural structure of the content circulating in digital environments. Theories of sexualization and objectification (Fredrickson and Roberts, 1997; Ward, 2016) have consistently shown that repeated exposure to stereotypical representations contributes to the internalization of gender norms and the reproduction of symbolic inequalities. However, applying these frameworks to social media requires a more precise understanding of the content itself, including how it is produced, amplified, and normalized within platform-specific dynamics.

Unlike traditional media, social media are characterized by decentralized content production, active user participation, and algorithmic curation, all of which shape visibility and engagement. In this context, gender discrimination may not only take the form of explicit hostility but also emerge through subtle, normalized practices of sexualization or stereotyping embedded in everyday interactions and self-presentation (Megarry, 2014; Sills et al., 2016). The lack of adequate tools to systematically analyze such content limits the development of more comprehensive theoretical models capable of accounting for these processes.

Finally, these issues are particularly relevant when considering adolescent and young adult populations. These developmental stages are critical for identity formation, the development of intimate relationships, and the construction of body image (Leaper and Brown, 2018). Social media play a central role in these processes, functioning as key environments for social comparison and normative learning. Exposure to content that sexualizes or stereotypes women may therefore have significant implications for psychological wellbeing, self-perception, and attitudes toward gender relations. However, in the absence of precise measurement tools, it remains difficult to determine which specific characteristics of digital content are most influential in shaping these outcomes.

Overall, advancing research in this field requires moving beyond an exclusive focus on perceived exposure and toward a more comprehensive approach that integrates multiple levels of analysis. Developing standardized, theory-driven instruments for content analysis represents a crucial step for strengthening both empirical research and theoretical understanding.

5 Limitations and future proposals

Several limitations of the present review should be considered when interpreting the findings.

First, the relatively small number of included studies reflects both the specificity of the inclusion criteria and the limited availability of research that develops or applies tools for directly measuring gender-discriminatory content on social media. While this highlights an important gap in the literature, it also restricts the generalizability of the conclusions.

Second, the review focused exclusively on studies published in peer-reviewed journals and written in English, which may have introduced publication and language bias. Relevant contributions addressing measurement of gender-discriminatory content on social media exist in other languages, including Spanish, Portuguese, and German, and their exclusion may have further reduced the pool of eligible studies.

Third, the temporal restriction to studies published from March 2020 onward, although motivated by the documented acceleration in social media use during and after the COVID-19 pandemic, has less justification when the object of review is measurement instruments rather than exposure effects. Validated tools for content analysis are not necessarily tied to a specific technological moment, and relevant measurement work conducted prior to 2020 may have been excluded by this criterion. In fact, it should also be noted that several instruments used in the included studies were originally developed and validated before 2020; the foundational psychometric literature for these tools therefore falls outside the search window.

Future reviews may benefit from extending the temporal range to capture earlier methodological contributions.

Fourth, an additional limitation concerns the absence of prospective protocol registration. Although the review followed PRISMA 2020 recommendations, the review process was not preregistered in PROSPERO and no publicly available protocol was established before study selection. Consequently, methodological decisions were not externally documented in advance.

Finally, the marked methodological heterogeneity of the included studies, ranging from experimental designs to qualitative analyses, limits the possibility of systematic comparison across instruments and approaches.

Considered together, these criteria may have cumulatively contributed to the small number of included studies. While the findings are consistent with a genuine scarcity of tools specifically designed to measure gender-discriminatory content on social media, we cannot exclude that the selection strategy introduced a degree of systematic underrepresentation.

Despite these limitations, the review identifies several promising directions for future research. A key priority concerns the development of standardized and theoretically grounded tools for the direct analysis of digital content. Such instruments should be capable of capturing the multidimensional nature of gender discrimination, including not only sexualization but also stereotyping, objectification, and implicit or normalized forms of sexism.

In addition, future studies would benefit from integrating multiple data sources, combining content analysis with self-reported exposure and behavioral indicators of user interaction. Interdisciplinary approaches that incorporate computational methods may offer valuable opportunities to scale content analysis and address the complexity of contemporary digital ecosystems (Dutta et al., 2025).

Given the rapid evolution of social media platforms, it is essential that measurement tools remain adaptable to emerging formats, including short-form video, livestreaming, and algorithmically curated feeds, which increasingly shape the experiences of adolescents and young adults.

Moreover, future methodological development may benefit from explicitly integrating the three levels identified in the present review: perceived exposure, actual exposure, and content properties. Combining self-report measures with behavioral trace data and direct content analysis could provide a more comprehensive representation of young people's digital environments. Such multimethod approaches would allow researchers to examine not only what users believe they encounter online, but also what they actually encounter and the objective characteristics of the content to which they are exposed (Scharkow, 2016; Parry et al., 2021).

This integration is increasingly recommended within digital media research, where scholars have argued that understanding media effects requires moving beyond single-source measurement strategies and incorporating both user-centered and content-centered indicators of exposure (Valkenburg, 2022; Guess et al., 2023). In the context of gender-discriminatory content, such approaches could help clarify whether observed outcomes are driven by the actual prevalence of discriminatory material, by individuals' interpretations of that material, or by the interaction between the two. Future studies may therefore benefit from developing integrated measurement frameworks capable of linking objectively assessed content characteristics with behavioral exposure indicators and subjective user experiences.

Beyond documenting the current methodological limitations of the literature, the findings of this review also suggest a possible direction for future measurement development. One recurring pattern across the included studies is that exposure to gender-discriminatory content is generally treated as a unidimensional construct, typically assessed through self-reported frequency of exposure. However, the relationship between individuals and online discriminatory content may be more complex.

Drawing on the attitude literature (Rosenberg and Hovland, 1960; Eagly and Chaiken, 1993), future research may benefit from considering at least three complementary dimensions. A first dimension concerns the cognitive recognition of discriminatory content. In contemporary social media environments, sexism is often communicated through humor, irony, memes, or highly normalized representations (Megarry, 2014; Sills et al., 2016; He, 2025), making it difficult to determine whether users actually identify such content as discriminatory. A second dimension concerns individuals' affective responses. Exposure does not necessarily imply endorsement, and users may differ substantially in the extent to which they perceive sexist content as acceptable, harmless, entertaining, or problematic. Finally, a third dimension concerns behavioral tendencies, including the willingness to engage with, endorse, or share such content within digital platforms.

Considering these dimensions together may provide a more comprehensive understanding of how gender-discriminatory content operates within social media ecosystems. In particular, it may help explain why exposure to similar content can be associated with different outcomes across individuals. Some users may readily recognize discriminatory messages and reject them, whereas others may perceive them as acceptable forms of humor or social interaction and contribute to their circulation (Semenzin and Bainotti, 2020; Dutta et al., 2025).

From this perspective, advancing measurement efforts may require moving beyond the exclusive assessment of exposure and toward a broader evaluation of how users interpret, experience, and respond to gender-discriminatory content. The methodological gap identified in the present review highlights the need for instruments capable of capturing these processes, particularly among adolescents and young adults, for whom social media represent a central context of socialization and identity development (Leaper and Brown, 2018; Choukas-Bradley et al., 2022).

6 Conclusion

This systematic review examined the instruments currently used to assess exposure to gender-discriminatory content on social media among adolescents and young adults. The findings indicate that, despite growing scholarly attention to gender inequality in digital environments, the literature remains largely focused on individual perceptions, attitudes, and outcomes, with comparatively limited attention to the systematic measurement of online content.

Across studies, exposure is predominantly operationalized using self-report measures, whereas direct analyses of digital content are rare and methodologically underdeveloped. More specifically, the available literature is largely restricted to measures of perceived exposure, while studies assessing actual exposure through behavioral or logged data or directly measuring content properties remain extremely rare. This imbalance limits our ability to determine the extent to which subjective experiences of online environments correspond to the content that users actually encounter and to the objective characteristics of that content (Parry et al., 2021; Valkenburg, 2022). As a result, existing research provides important insights into the potential effects of gendered representations but offers a more limited understanding of how such representations are structured, distributed, and normalized within social media ecosystems.

Addressing this gap requires the development of integrated measurement approaches capable of simultaneously capturing users' perceptions, objectively recorded exposure patterns, and the structural features of gender-discriminatory content circulating within social media ecosystems. Such advances would strengthen both the empirical precision and the theoretical explanatory power of future research on online gender discrimination.

Advancing research in this direction is essential not only for improving empirical accuracy but also for deepening our understanding of how social media contribute to contemporary processes of gender socialization and inequality.

The findings of this review suggest that future measurement efforts may benefit from moving beyond the exclusive assessment of exposure to gender-discriminatory content. Considering how individuals recognize such content, how they emotionally respond to it, and how they engage with it within social media environments may provide a more comprehensive understanding of its role in contemporary processes of gender socialization. Such a perspective could support the development of more robust and theoretically grounded instruments for both research and intervention purposes.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

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: Chiara Rollero, University of Turin, Italy

Reviewed by: Cecilia Ines Calero, Torcuato di Tella University, Argentina

M. Paz Toldos, Monterrey Institute of Technology and Higher Education (ITESM), Mexico