The Power of Social Media: Stigmatizing Content Affects Perceptions of Mental Health Care

Abstract

A great discrepancy exists between the number of people who experience mental health concerns and the number of people who seek help for those problems. Negative attitudes toward mental health treatment contribute to this discrepancy. In the current study, we tested the extent to which exposure to social media posts that derogate mental health treatment increases such negative attitudes. Participants were exposed to 10 ostensibly real Twitter posts. Whereas some participants viewed 10 posts that did not relate to mental health, others viewed five such innocuous posts interspersed with five posts that derogated mental health treatment. Participants then reported their levels of mental health self-stigma and attitudes toward help-seeking. These dependent variables were combined into a single measure of stigma toward mental health treatment. Results indicated that among men and among women high in feminine gender role espousal, there was no effect of the manipulation on stigma against mental health treatment. However, women low in feminine gender role espousal were impacted by the manipulation: Among these women, the stigmatizing tweets increased the extent to which they stigmatized mental health services. Our research thus shows that even brief exposure to messages derogating mental health treatment may have a profound impact on associated attitudes but that such effects are moderated by gender identification and the extent to which people espouse traditional gender roles.

According to the National Institute of Mental Health (NIMH, 2020), about 51.5 million adults in the United States experienced a mental health condition in 2019. Of those adults, only about 23 million (44.8%) received professional psychological help during 2019 (NIMH, 2020). Similar results have been found across the world, including in the United Kingdom (McManus et al., 2016), Argentina (Cia et al., 2019), and Singapore (Chong et al., 2012). Several factors might explain why so few people who experience mental health disorders seek and receive treatment. One potential explanation could be logistical challenges, such as lack of transportation, health insurance, or financial resources (Andrade et al., 2014). People’s attitudes toward mental illness, self-reliance, and help-seeking also impact whether they seek help for mental health concerns (Andrade et al., 2014). In particular, stigma surrounding mental health may prevent individuals from seeking help for mental health concerns (Andrade et al., 2014). In addition, gender and gender role beliefs have been shown to predict help-seeking attitudes toward mental illness (e.g., Bradbury, 2020; Chandra & Minkovitz, 2006; Cole & Ingram, 2020; Juvrud & Rennels, 2017).

In the current research, we explore the extent to which social media content may also play such a role. A wealth of scholarship has provided evidence for the powerful effects of social media in a variety of contexts (see, e.g., Boulianne, 2015; Huang, 2022; Ibrahim, 2021 for meta-analyses). Mental health is one such context that has attracted a great deal of scholarly attention. For example, McComb et al. (2023) found evidence that social media use is associated with heightened upward social comparison, which in turn predicts worse self-perceptions of body image, subjective well-being, self-esteem, and overall mental health. Braghieri et al. (2022) found that the staggered introduction of Facebook to a college campus in the mid-2000s was associated with increases in self-reported mental health challenges. Primack et al. (2017) reported that respondents who used more social media platforms reported higher rates of anxiety and depression than respondents who used fewer social media platforms. Scholars have also used the Uses and Gratifications Theory (e.g., Ruggerio, 2000) to understand the interplay between social media and mental health. For example, Ferris et al. (2021) found that people’s motives for using social media predicted the severity of the addictive consequences of social media, while Kircaburun et al. (2020) found that certain social media use motives, including meeting new people and expressing a more “popular” self, were associated with problematic usage of social media.

Here, in an experimental study, we explore the impact of exposure to social media content that stigmatizes mental health treatment on attitudes toward mental health. In doing so, we assess the moderating roles of gender and gender role espousal.

Mental Health Stigma and Help-Seeking

Whereas Sheehan et al. (2017) described eight forms of stigma related to mental health, Corrigan and Rao (2012) described two main forms: public stigma and self-stigma. Public stigma is defined as prejudice and discrimination toward people with mental health disorders. When people internalize these stereotypes and begin to believe that they are true, self-stigma is said to occur (Corrigan & Rao, 2012, p. 464). Many forms of health stigma are interconnected, and these stigmas may manifest in a variety of ways, including experienced discrimination, experienced or public stigma, internalized or self-stigma, perceived stigma, anticipated stigma, and secondary or associative stigma experienced by those close to the individual (Stangl et al., 2019). Stangl et al.’s (2019) Health Stigma and Discrimination Framework articulate ways in which all members of society may experience, reinforce, perceive, anticipate, internalize, and perpetuate health-related stigma.

Mental Health Self-Stigma and Attitudes toward Help-Seeking

Overall, perceived public stigma about mental health appears to serve as a barrier to attitudes toward help-seeking, although cultural factors may moderate this relationship. One consequence of public stigma toward mental health treatment is that individuals internalize mental health stigma (Corrigan & Rao, 2012), and public stigma is a notable driver of self-stigma, making differences between those with and without mental illness more salient (McCullock & Scrivano, 2023). Similar to public mental health stigma and perceived public mental health stigma (Chandra & Minkovitz, 2006; Hom et al., 2018; Jennings et al., 2015; Nadeem et al., 2007), self-stigma predicts negative attitudes toward help-seeking across a range of populations (Cole & Ingram, 2020; Jennings et al., 2015; Lannin et al., 2016; Vogel et al., 2006, 2011). Importantly, sociocultural means of communicating mental health stigma (e.g., comments from parents, peers, media, and people in positions of power; Ward & Grower, 2020) may be particularly relevant in understanding barriers to help-seeking.

Reduced self-stigma has been shown to be positively associated with secure attachment styles (Zhao et al., 2015), strong peer relationships (Zhao et al., 2015), greater mental health literacy and motivation to seek help (DeBate et al., 2018), weaker self-reliance beliefs (e.g., “Strong people can get over psychological problems by themselves,” Jennings et al., 2015), more-positive attitudes toward counseling, and choosing to seek help or information about mental health (Lannin et al., 2016). Those experiencing severe distress with higher self-stigma were half as likely to seek out information or counseling as those with lower self-stigma (Lannin et al., 2016). Finally, in addition to predicting attitudes toward help-seeking, self-stigma differentiated those who did and did not seek help over a 2-month follow-up period in a large mixed-gender sample of U.S. college students (Vogel et al., 2006). Thus, research shows that self-stigma predicts both attitudes toward help-seeking and actual help-seeking.

Gender and Gender Role Espousal

In addition to the body of research demonstrating a relationship between mental health stigma and attitudes toward seeking treatment, research also suggests that this relationship is moderated by gender and espousal of gender roles (Cole & Ingram, 2020; Juvrud & Rennels, 2017), among other factors. Gender role was originally defined as “behaviors, expectations, and role sets defined by society as masculine or feminine, which are embodied in the behavior of the individual . . . and culturally regarded as appropriate to males or females” (O’Neil, 1981, p. 203). These gender roles are socially constructed (Sharpe & Heppner, 1991) and often reinforce the spurious idea of a gender binary. Espousal of gender roles refers to the incorporation of these socially constructed beliefs about gender and “appropriate” gender roles into one’s own attitudes and behaviors, which then predicts multiple broad facets of our functioning, including perception, memory, attention, social behaviors, interests, and self-perceived competencies (Ward & Grower, 2020).

Within and across cultures, gender and gender role espousal predict mental health stigma and attitudes toward help-seeking (Vogel et al., 2011). In particular, women and girls hold more favorable attitudes toward help-seeking and report lower self-stigma (Bradbury, 2020; Chandra & Minkovitz, 2006; Cole & Ingram, 2020; Juvrud & Rennels, 2017; Topkaya, 2014). Indeed, a meta-analysis found that conformity to masculine norms was modestly negatively associated with psychological help-seeking, with items assessing self-reliance, playboy, and power over women most strongly associated with unfavorable attitudes toward help-seeking (Wong et al., 2017).

Most research examining gender and gender role espousal has only examined reported attitudes toward help-seeking rather than actual help-seeking behaviors. Juvrud and Rennels (2017), however, used a novel puzzle task in which, before participants began the task, the researcher offered to help participants if they requested it. Results indicated that men and women performed similarly on the number of puzzles solved, but women sought help more than men. Men with more masculine stereotypes and gender-stereotyped attitudes were less likely to seek help than other men, both showing decreased frequency of help-seeking and a longer wait prior to seeking help. Women’s masculine stereotypes also negatively predicted the frequency of help requests and positively predicted the latency of their help requests. Interestingly, there were no gender differences in reported help-seeking for physical or mental health concerns, but feminine stereotype espousal did predict reported help-seeking for health problems. Thus, it is important to understand both gender and gender role espousal in predicting people’s help-seeking behaviors. This research also suggests that men may report similar levels of help-seeking as women but may not actually seek help as often.

The gender difference in attitudes toward help-seeking appears to exist from a young age. Chandra and Minkovitz (2006) examined mental health stigma, attitudes toward help-seeking, and use of social support among eighth-grade boys and girls from two different Canadian middle schools. Girls had more knowledge about, and experiences with, mental health than boys. Further, girls were twice as likely as boys to express the likelihood of using mental health services (Chandra & Minkovitz, 2006). Parental disapproval and perceived mental health stigma partially explained the relationship between gender and the children’s willingness to use mental health services. These results imply that there is a connection between mental health knowledge, experience, and willingness to seek treatment. In addition, parental views about mental health and gender-related perceptions of stigma likely shape youths’ mental health help-seeking attitudes and behaviors.

The Media’s Role in Mental Health Stigma and Attitudes toward Help-Seeking

People learn sociocultural norms, such as those about gender roles and mental health stigma, from a range of sources, including parents and other adult role models, peers, mainstream media, and social media (e.g., Gagrčin et al., 2022; Santarossa & Woodruff, 2017; Ward & Grower, 2020). Mainstream and social media may be a particularly powerful conduit of sociocultural values, partially because people in the media are seen as more “attractive” than people in one’s actual day-to-day interactions, making it more likely that they will be viewed as aspirational role models (Greenwood, 2016; Ho et al., 2016; Ward & Grower, 2020).

A recent systematic review of the impact of media reports of serious mental illness (SMI) on mental health stigma found that across a number of experimental studies, negative stories of those with SMI worsen individuals’ mental health stigma (Ross et al., 2019). Although the number of studies was limited, news reports that aimed to challenge mental health stigma that were informative or focused on recovery were associated with lower levels of mental health stigma (Ross et al., 2019). Additional non-experimental studies support this finding. Textual analysis of readers’ responses to positively framed versus negatively framed news stories found that readers’ comments largely matched the valence of media portrayals of people with SMI (as cited in Ross et al., 2019). Ross and colleagues (2019) noted a dearth of studies examining the role of social media content on users’ mental health stigma, noting that users’ comments generally reflected the tone of the posts’ content. This highlights the potential importance of the tone and valence of social media posts about mental illness.

Social media may serve a particularly important role in shaping individuals’ mental health stigma, perceptions of stigma, and attitudes toward help-seeking, given its synchronicity and users’ ability to comment, read, and respond to others’ comments (Pavlova & Berkers, 2020). For example, a sample of Twitter users rated 1,101 posts that were “scraped” from Twitter related to seven commonly stigmatized disorders (schizophrenia, depression/anxiety, autism, eating disorders, obsessive-compulsive disorder, and addiction) for stigmatizing content (Jansli et al., 2022). A large proportion of the posts were rated as stigmatizing, with some conditions more often stigmatized (schizophrenia) than others (depression/anxiety). Other studies conducting content analyses of mental health posts on Twitter have found a mix of content, with “inspirational” posts shared the most, and stigmatizing content the second-most shared, of seven types of posts that were identified (Saha et al., 2019). Indeed, researchers note that the rising popularity of social media platforms “creates opportunities as well as threats for mental health discourse” (Pavlova & Berkers, 2020, p. 113, 250). Among adolescents, greater time spent on social media predicted greater rumination about social media posts (e.g., worry about whether posts would be “liked,” shared, etc.; see also Büchi et al., 2019), which then predicted anticipated self-stigma related to seeking help for mental health concerns.

The Current Study

Prior research shows that there is an association between high levels of mental health stigma and negative attitudes toward help-seeking (Jennings et al., 2015). In addition, gender role espousal and gender are both related to mental health stigma and attitudes toward seeking psychological treatment (Cole & Ingram, 2020; Topkaya, 2014; Vogel et al., 2011). Limited research, however, has examined the role of social media in influencing users’ mental health stigma and attitudes toward help-seeking. In addition, to date, the role of participants’ gender and gender role espousal on attitudes toward help-seeking has never been tested as a moderator of the relationship between social media and users’ self-stigma and attitudes toward help-seeking. In the current research, we tested the effect of exposure to stigmatizing social media posts on mental health self-stigma and attitudes toward help-seeking, and how those effects may be moderated by gender and gender role espousal.

We hypothesized three main effects: First, consistent with scholarship showing general effects of media on help-seeking attitudes (e.g., Ross et al., 2019), we hypothesized a main effect of stigmatizing social media posts, such that people who viewed such posts would report higher levels of mental health self-stigma and more-negative attitudes toward help-seeking compared to a control condition. Second, consistent with prior research (e.g., Ricciardelli et al., 2020) we predicted a main effect of gender, such that men would express more mental health self-stigma and more-negative attitudes toward help-seeking overall. Next, we hypothesized that participants higher in feminine gender role espousal would show less mental health self-stigma and more-positive attitudes toward help-seeking (e.g,. Juvrud & Rennels, 2017). Finally, we hypothesized an interaction, such that the effect of the manipulation on mental health stigma would be stronger for those low in feminine gender role espousal than among those high in feminine gender role espousal.

Method

Participants

One hundred eighty-six individuals were recruited through Amazon’s Mechanical Turk participant-procurement platform to complete a study ostensibly about perceptions of posts on social media. Data from nine participants were incomplete so they were removed from the sample. In addition, one person did not disclose their gender and was therefore also removed from analyses. Thus, a total of 176 participants were included in the data analysis, with 57 (32.4%) identified as female and 119 (67.6%) identified as male; none identified as non-binary or “other.” The mean age of participants was 37.19 (SD = 11.27) years.

Measures

Gender Role Espousal

Five items from the expressive subscale of Spence et al.’s (1974) Personal Attributes Questionnaire assessed the extent to which participants espoused traditional gender roles of femininity. Participants were asked to rate the degree to which different traits represented the typical man or woman (cf. Spence & Buckner, 2000). Sample traits included “aware of others’ feelings” and “emotional.” Although the Personal Attributes Questionnaire was developed on a 5-point scale, due to a programming error, the items were assessed on a four-point scale anchored with 1 (man much more than woman) to 4 (woman much more than man). Items were coded such that higher numbers indicated greater espousal of traditional feminine gender roles. Internal reliability was acceptable; Cronbach’s α = .80.

Mental Health Self-Stigma

The Self-Stigma of Seeking Help scale (SSOSH; Vogel et al., 2006) consists of 10 items assessing the extent to which participants’ self-esteem would be impacted by their decision to seek psychological help, such as “If I went to a therapist, I would be less satisfied with myself” and “My view of myself would not change just because I made the choice to see a therapist.” Responses for each item were on a five-point scale from 1 (strongly disagree) to 5 (strongly agree). The SSOSH uniquely predicted attitudes toward seeking help for mental health concerns and help-seeking over a 2-month follow-up period in the validation sample (Vogel et al., 2006). Items were averaged; higher numbers indicate greater self-stigma. The scale demonstrated acceptable internal reliability; Cronbach’s α = .89.

Attitudes toward Help-Seeking

The 10-item Attitudes Toward Seeking Professional Psychological Help-Short Form scale (ATSPPH-SF; Fischer & Farina, 1995) is a commonly used measure that assesses attitudes toward seeking psychological help. Example items are: “The idea of talking about problems with a psychologist strikes me as a poor way to get rid of emotional conflicts” and “I might want to have psychological counseling in the future.” Responses for each item were rated on a four-point scale from 1 (agree) to 4 (disagree). Items were averaged, with higher scores reflecting more-negative attitudes toward psychological help-seeking. The scale manifested acceptable internal reliability; Cronbach’s α = .75.

Procedure

Participants were recruited via Amazon Mechanical Turk (MTurk; www.mturk.com) for cash compensation. MTurk is an online crowdsourcing platform that compensates workers who voluntarily complete online human intelligence tasks, including surveys (e.g., Paolacci et al., 2010). MTurk allows for prescreening participants who successfully complete tasks, typically yielding diverse samples (Paolacci et al., 2010). The current sample was restricted to participants whose responses had been approved 95% of the time, were 18 years of age or older, and were located in the United States of America. We did not screen for usage of social media. Bot-check and attention-check questions were used to screen bots and invalid responses (c.f., recommendations by Burnette et al., 2022). MTurk participants have been found to respond with equivalent honesty and consistency as do participants from convenience-sampling techniques (Chandler & Shapiro, 2016).

After providing informed consent, participants completed the expressive subscale of Helmreich et al.’s (1981) Personal Attributes Questionnaire to report their levels of feminine gender role espousal. Participants then viewed 10 fictitious yet ostensibly real posts to the Twitter social media platform (“tweets”). We chose Twitter as our social media platform because user content is generally public to all (as opposed to Facebook), and is generally text-based (as opposed to Instagram). Participants were randomly assigned to view either 10 innocuous tweets (the control condition) or five innocuous tweets and five tweets that stigmatized mental health diagnoses or services (the stigma condition). The author of the tweets was identified only by name and username, and there were no images associated with the tweets. An example of an innocuous tweet is “It snowed soo much last night! I think @jasonc694 and I are gonna go skiing today,” while an example of a stigmatizing tweet is “My friend is feeling sad again today. It’s not ‘depression’ or ‘bipolar’—those aren’t real. STOP WHINING” (Figure 1 presents this tweet as it was presented to participants). In a pilot test, 78 additional MTurk workers perceived that the 10 posts used in the stigma condition expressed more negativity toward mental health treatment, M(SD) = 2.28 (1.11) than did the 10 posts in the control condition, M(SD) = 3.18 (1.29), t(76) = 3.34, p = .001.

Figure 1.

Example Tweet derogating mental health treatment.

After viewing the tweets, participants were asked a series of filler questions about the posts, such as how interesting they found the posts and how interesting they thought the author of the posts might be to corroborate the supposed purpose of the study. In an ostensibly unrelated study, participants then completed the SSOSH (Vogel et al., 2006) and Attitudes toward Help-Seeking (Fischer & Farina, 1995) scales.

Results

Data are available at https://osf.io/szn2v/?view_only=c164ae6f3139478589799b08225988ce.

Preliminary Analyses

The two dependent variables, mental health self-stigma and attitudes toward help-seeking, were very strongly associated, r = .77, p < .001. Because of this strong correlation, we assessed whether the two dependent variables might better be considered as a single construct that assessed overall stigma toward mental health treatment. Entering all 20 items from both measures into a single analysis yielded a strong reliability score, Cronbach’s α = .90, and entering all 20 items from both measures into an exploratory factor analysis yielded a single factor, Eigenvalue = 7.41, which explained 37.04% of the variance. Together, these analyses suggested that all 20 items were better conceptualized as indicators of a single underlying construct than of two separate constructs. We therefore combined the 20 items into a single composite variable, which we will refer to as the stigma score. Participants with higher stigma scores manifested higher mental health self-stigma and more-negative attitudes toward help-seeking.

Primary Analyses

These stigma scores were entered into a linear regression in which participants’ gender, participants’ feminine gender role espousal score, and condition (stigma vs. control tweets) were predictors, along with the three two-way interaction terms and the one three-way interaction term. As hypothesized, feminine gender role espousal predicted stigma scores, t = −2.83, beta = −0.88, p = .005, such that participants higher in feminine gender role espousal manifested lower stigma scores. Contrary to hypotheses, however, there was no association between gender and stigma scores, t = −1.56, beta = −.79, p = .12, nor between condition and stigma scores, t = −1.34, beta = −1.27, p = .18.

Among the two-way interactions, the Gender x Condition term approached statistical significance, t = 1.83, beta = 1.92, p = .07, as did the Gender x Gender role espousal term, t = 1.73, beta = 1.13, p = .09, while the Condition x Gender role espousal term did not, t = 1.54, beta = 1.51, p = .13.

Importantly, though, these effects were qualified by a significant three-way Condition × Gender × Gender role espousal term, t = −1.98, beta = −2.25, p = .05. ¹

To interpret this three-way interaction, we first explored the extent to which feminine gender role espousal and condition predicted self-stigma scores among men. This regression analysis yielded a main effect of feminine gender role espousal, t = −4.07, beta = −.47, p < .001, such that, as was the case across the full sample, men higher in feminine gender role espousal manifested lower stigma scores. There was, however, no effect of condition, t = 0.12, beta = .04, p = .91, and no interaction between condition and feminine gender role espousal, t = 0.22, beta = .08, p = .83. Thus, men’s stigma scores were not impacted by the condition to which they were assigned, and, regardless of condition, men lower in feminine gender role espousal manifested higher stigma scores than did men higher in feminine gender role espousal.

We then explored the extent to which feminine gender role espousal and condition predicted stigma scores among women. As suggested by the initial three-way interaction, women showed a different pattern of results than men. Among women, there was no association between feminine gender role espousal scores and stigma scores, t = −0.33, beta = −.06, p = .75, but there was an effect of condition on stigma scores, t = 2.14, beta = 1.37, p = .04, indicating that women exposed to the stigmatizing tweets reported more stigma than did women assigned to the control condition. More importantly, though, these main effects were qualified by a Condition x Gender role espousal interaction, t = −2.18, beta = −1.47, p = .03. To understand this two-way interaction, we compared women high in feminine gender role espousal (scores on the gender role measure that were above the median) with those low in feminine gender role espousal (scores below the median). Among women high in feminine gender role espousal, there was no effect of condition on stigma scores, t(25.48) = 1.69, p = .10, such that those exposed to stigmatizing tweets, M(SD) = 37.55(9.11) and those in the control condition, M(SD) = 44.58(15.01) demonstrated equivalent stigma. Conversely, among women low in feminine gender role espousal, there was an effect of condition on stigma scores, t(18) = −2.49, p = .02, such that those exposed to stigmatizing tweets demonstrated significantly higher stigma, M(SD) = 57.80(10.04), than did those in the control condition, M(SD) = 46.80(9.69).

In sum, then, men and women were impacted differently by stigmatizing tweets. Men were not impacted by the tweets they viewed: Rather, men higher in feminine gender role espousal showed lower self-stigma, regardless of exposure condition. Women, on the other hand, showed a more-complex pattern of results: Whereas women high in feminine gender role espousal were not influenced by the stigmatizing tweets, women low in feminine gender role espousal were. Among these latter women, the stigmatizing tweets increased the extent to which they derogated mental health services.

Discussion

The discrepancy between the number of individuals who experience mental illness and the number of individuals who seek treatment for those mental health concerns is striking (e.g., Aguirre Velasco et al., 2020; Ebert et al., 2019; Mojtabai et al., 2011). This discrepancy stems from a variety of reasons, with people’s attitudes toward mental health services and toward mental health professionals cited as the most common barriers (Aguirre Velasco et al., 2020). In addition, attitudinal/evaluative factors are more important than structural barriers to initiating and continuing treatment among those with mental health disorders (Mojtabai et al., 2011). For a variety of reasons, negative attitudes toward mental health treatment continue to prevent people from seeking the help they need.

In the current research, we tested the extent to which exposure to social media posts endorsing such stigma might impact participants’ attitudes toward seeking mental health treatment. Women low in feminine gender role espousal who were exposed to ostensibly real Twitter posts derogating mental health treatment held more-negative attitudes toward such treatment than did participants who viewed neutral tweets. To our knowledge, this is the first study to examine whether gender role espousal moderates the impact of social media messages on individuals’ attitudes toward mental health treatment.

Implications

At a practical level, our findings demonstrate the power of social media in creating and perpetuating negative attitudes toward mental health treatment, at least among a subset of the population. That such a modest manipulation—viewing five brief social media posts stigmatizing treatment for mental illness—would impact people’s self-reported attitudes toward mental health treatment speaks to the power of social media in perpetuating these negative attitudes. At the same time, though, this effect only emerged among women low in feminine-typed gender role espousal. It did not emerge among men or among women high in feminine gender role espousal. Thus, the effect of stigmatizing social media posts appears to be both powerful and nuanced.

At a broader level, our scholarship underscores the extant research across a variety of contexts, showing the negative effects that social media can have on people’s cognitive and attitudinal processes. For example, our findings are consistent with those arguing the potentially harmful effects of social media on disordered eating and body image dissatisfaction (e.g., Harriger et al., 2022; Logrieco et al., 2021; Roberts et al., 2022), suicidal ideation (e.g., Spitzer et al., 2023), and overall subjective well-being (e.g., Brooks, 2015), among others. Similarly, research suggests that social media plays a role in shaping attitudinal factors associated with people’s health behaviors, such as vaccine hesitancy (e.g., Moran et al., 2022; Wilson & Wiysonge, 2020), with tweets about health-related information often spreading rapidly online (Wang et al., 2021) and individuals with greater health-related conspiracy beliefs reporting using social media as a source of health information (Allington et al., 2020).

Finally, our research speaks to the gendered nature of attitudes toward mental health treatment. Our finding that feminine gender role espousal was associated with lower mental health stigma is aligned with prior research reporting that women and those higher in feminine gender role espousal report lower mental health stigma and more favorable attitudes toward help-seeking (Bradbury, 2020; Chandra & Minkovitz, 2006; Cole & Ingram, 2020; Juvrud & Rennels, 2017; Topkaya, 2014; Wong et al., 2017). Our findings suggest nuance in the way that gender, gender role espousal, and exposure to messages stigmatizing mental health have on attitudes toward help-seeking.

Interventions aimed at counteracting public stigma at the individual level have largely focused on education regarding mental illness and intergroup contact, whereas structural interventions typically focus on mass media education campaigns and legislation to protect those with mental illness (McCullock & Scrivano, 2023). Both education and intergroup contact interventions generally result in small-to-moderate short-term effects with concerns regarding the maintenance of those effects (McCullock & Scrivano, 2023), with limited conclusions available regarding the efficacy of self-stigma-based interventions. In their meta-review of meta-analyses on stigma-reduction interventions, McCullock and Scrivano (2023) noted that, in line with Stangl et al.’s (2019) Health Stigma and Discrimination Framework model, it is likely important to focus on intersecting stigmatized identities and to include multicomponent interventions that may target public and self-stigma in addition to structural elements of stigma. Notably, the current study suggests the importance of structural facets of stigma. Stricter policy regulations regarding social media messaging could potentially have synergistic effects when paired with existing individual-level contact and education-based methods to limit mental health stigma.

Limitations and Directions

Future research should delve deeper into understanding why the effect emerged only among women low in feminine gender role espousal. Several possible reasons may underlie this effect. For example, among men, feminine gender role espousal may play a stronger role in influencing help-seeking attitudes, overriding any effect of short-term social media exposure. Or it may be possible that, across gender, feminine gender role espousal may be stronger than the effects of stigmatizing content in influencing people’s attitudes toward help-seeking, such that those who are high in feminine gender role espousal may be likely to reject or ignore such messages.

Additional ideas for future research relate to the study’s operationalizations. For example, although our manipulation of exposure to social media posts derogating mental health treatment was effective, stigma was assessed very soon after exposure to the posts. This suggests that, at the very least, it would be interesting to test the extent to which the effects of tweets on attitudes endure beyond the relatively brief duration of the study. In addition, the social media posts were ostensibly written by an author with whom participants were not familiar. It would therefore be worthwhile to test whether the source of such social media posts might further moderate the effect: It is conceivable that posts derogating mental health treatment from well-liked celebrities or from people familiar to the participants might yield stronger effects on stigma that might overwhelm the moderating influence of gender and gender role espousal that we observed.

In addition, other factors are known to predict a person’s attitudes toward help-seeking, such as the individual’s own treatment history and whether or not they are referred to help by a family member or other trusted individual (Wahto et al., 2016). These factors were not assessed in the current study. Future research would benefit from assessing and controlling for the effects of these factors, which are associated with help-seeking attitudes.

Another direction for future research would be to assess the extent to which social media usage moderates the observed effects. It may be that heavy users of social media are especially impacted by exposure to social media posts derogating mental health treatment. Such people may have come to rely on social media for much of their knowledge about the world and may be more likely to trust the information they see on such platforms, as a result. This heightened trust might in turn yield particularly strong effects of exposure, perhaps so strong as to overwhelm the interactions we found. Conversely, it is conceivable that such heavy users might be less impacted by posts derogating mental health treatment. Being more experienced with social media, heavy users might perceive social posts more skeptically. Such skepticism might yield weak or nonexistent effects of exposure to derogatory posts. Some research supports this, finding that greater social media consumption is negatively associated with belief in misinformation (Halpern et al., 2019). Additional research should incorporate the Uses and Gratifications Theory (Katz et al., 1973) as a framework to examine whether individuals’ social media usage moderates the effects observed in the current study.

Conclusion

This is the first study to examine whether gender and gender role espousal moderate the impact of social media messages on individuals’ stigma toward mental health help-seeking. For women low in feminine gender role espousal, viewing social media posts that derogated mental health treatment yielded higher stigma. At a broad level, our research speaks to the potential harms of social media content that derogates mental health treatment. But perhaps more importantly, our research makes clear the importance of understanding how this effect is moderated by participants’ gender identification and gender role espousal.