Analysis of Social Media Use, Mental Health, and Gender Identity Among US Youths

Key Points Question Is gender identity associated with social media use and mental health among youths? Findings This cross-sectional study of 1231 transgender, gender nonbinary, and cisgender youths found that gender identity moderated both the effect size and the direction of the association between social media use and mental health. Meaning These results suggest that gender identity is a key moderator when examining youth social media use and mental health and should be included in studies moving forward.


Introduction
The percentage of transgender and nonbinary (TGNB) youths coming out in the US has doubled in the last decade, 1 with 1.4% as transgender and 3.0% as nonbinary. 2 Despite studies documenting an association between gender, social media use (SMU), and indicators of mental health, [3][4][5] and that TGNB youths are at a higher risk for mental health issues than cisgender youths, [6][7][8] relatively little research has examined the association between SMU and mental health among TGNB youths. 9 Consequently, this study aimed to directly explore the interplay between gender identity, SMU, and indicators of mental health, including internalizing (emotional problems, depression, and body image), and externalizing problems (conduct problems).
Gender identity is an individual's deeply felt sense of being a man, woman, or an alternate gender (eg, nonbinary), which sometimes may not align with the sex assigned at birth. 10 TGNB individuals may experience gender dysphoria, an intense distress because of the disconnect between one's assigned sex and internal gender identity. 10,11 TGNB individuals may also experience high levels of minority stress, 12,13 the stress of being a minority in a majority social environment [14][15][16] fostered through social processes, institutions, and structures that harass and/or discriminate. 13,17 For TGNB youths, minority stress can manifest through violence due to gender nonconformity, [18][19][20] gender dysphoria, 21 family tension, and emotional distress from fear of rejection 21,22 and is associated with increased mental health struggles and greater risk for suicide. 17,[23][24][25] Currently, 25% to 32% of TGNB youths attempt suicide. 18,26 Though there are public concerns about the effects of SMU on adolescent mental health, research on TGNB youths suggests social media may be a protective factor instead of a risk for mental health. 9 The research on SMU and mental health tends to be mixed. Some studies suggest time spent on social media is associated with mental health problems, 4,27,28 while others find no link. [29][30][31] Certain variables may moderate the association between SMU and mental health, including SMU context and content, problematic behaviors, 32 and gender (associations greater for girls than boys 33 ), suggesting gender identity as a potential moderator. 34 School media literacy may also help youths use social media in ways that might benefit their mental health; however, this is understudied with gender minority youths. 35 TGNB individuals engage with media in various ways for multiple reasons. 9,12 While general media represents TGNB individuals less frequently and accurately than cisgender individuals, [36][37][38][39] social media allows TGNB individuals to portray themselves how they see fit. TGNB youths access social media for a variety of content, building positive connections 38 and creating support systems protective against mental illness, based on common interests and experiences. 9,[40][41][42] Little is known, however, about how TGNB youths use social media in general outside of gender identity exploration and development.
Acknowledging continued interest in the association between social media and mental health, elevated mental health risks relevant to TGNB youths, and the potential effects of social media in the lives of TGNB youths, our study aimed to examine the association between SMU and mental health as moderated by gender identity. Following minority stress theory, we hypothesized that positive social media practices may be more protective for TGNB youths (compared with cisgender youths) as it may reduce feelings of minority stress.

Methods
The survey was approved by the Brigham Young University institutional review board and participants were treated under the human participants' guidelines from the American Psychological Association. Informed consent was obtained online. Parents gave consent for their minor children to participate. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies by having multiple TGNB individuals as authors and addressing potential bias via qualitative focus groups with TGNB youths.
Participants included a national quota sample of 1231 youths aged 10 to 17 years obtained using a Qualtrics panel and collected between May and August 2021. The panel consisted of youths from all 50 states. We gave Qualtrics quotas for race and ethnicity based on recent US Census data, and they recruited based on those estimates. We chose this age group to examine because many youths begin using social media during this age. Additionally, there is substantial gender identity development during adolescence.
Race and ethnicity were self-reported; categories included American Indian or Alaska Native, Asian, Black, Hispanic or Latinx, Pacific Islander, White, and mixed or another race or ethnicity. Selfreported gender categories included cisgender female, cisgender male, and transgender, nonbinary, or other. Household income data were also collected.

Measures
The data were skewed for outcome measures. Descriptions of transformations and reliability statistics are included in Table 1.

Outcome Variables
Depression | The patient health questionnaire (PHQ-8) measured participant depression. 43 Youths reported how frequently they experienced 8 symptoms within the last 2 weeks. A sample item of the PHQ-8 included "Feeling down, depressed, or hopeless." Responses ranged from 1 (not at all) to 4 (nearly every day).
Emotional Problems | Five items measuring emotional problems were completed from the Strengths and Difficulties Questionnaire (SDQ). 44 Items were rated on a 3-point scale, 1 (not true) to 3 (certainly true), and a sample item included "I worry a lot." Conduct Problems | Five items measuring conduct problems from the SDQ were also completed.
The same rating scale was used, and a sample item included, "I take things that are not mine." Body Image | Youths reported how often they agreed with 3 statements about body image 45 using a 5-point Likert-scale: 1 (never) to 5 (always). A sample item included, "I'm pretty happy about the way I look."

Media Variables
Social Media Time | Participants were first asked if they had ever used social media. Seventy-two participants (5.8%) reported they had never used social media and were omitted from all future analyses. Participants who reported they used social media estimated the hours they spent on social media in a typical day. This was measured on a scale of 1 (none) to 8 (more than 8 hours). Though only moderate indicators of SMU, self-reports of screen time provide information about youths' perceptions of their screen time and differentiates between light and heavy users. 46 Age at First Smartphone | For youths who had a smartphone, they were asked how old they were when they got their first smartphone. These answers ranged from age 5 to 17 years. Taking Intentional Breaks | Youths were asked how often they take intentional breaks from their smartphones (eg, by putting their smartphone on airplane mode or leaving it in another room).
Responses were measured on a 6-point Likert scale, 1 (never or rarely) to 6 (every day or almost every day).
Problematic SMU | Youths were first asked if they used either social media or video games more frequently. Then, youths who reported that they used social media more responded how much they agreed with 7 items related to their social media habits (those who chose video games were excluded from this scale only). This scale was adapted 48 from a scale that originally assessed problematic cell phone use. 49 Participants were asked to rate how much they agreed with a series of statements regarding their SMU (eg, "When I am not using social media, I am thinking about using it or planning the next time I can use it"). Items were rated using a 5-point Likert scale, 1 (strongly disagree) to 5 (strongly agree).
Digital Well-Being in Schools | Youths rated on a 5-point Likert scale how much they agreed with the statement, "My school tries to help us learn how to use our phones or social media in healthy ways." Items were rated from 1 (strongly disagree) to 5 (strongly agree).
Mindful Media Use | A modified Mindfulness Attention Awareness Scale 50 was used to measure mindfulness around SMU. Youths were asked to think about the last time they were on social media and how much certain behaviors were present (eg, "I was engaging with social media without really paying attention"). Items were rated on a 6-point Likert scale, 1 (not at all) to 6 (very much). This was recoded so a higher score indicated more mindful media use.
Cleaning and Curating | Participants were asked 2 items about how regularly they cleaned or curated their social media feed or followers (eg, by muting or unfollowing certain accounts). Items were rated on a 6-point Likert scale, 1 (never) to 6 (about once a week).

Qualitative
Member checking is often considered a hallmark of careful and culturally considerate research. 35 To aid with interpretation of the quantitative results, 7 adolescents were recruited as an advisory board via 2 focus groups centered on cointerpreting the findings. The advisory group was not designed for further data collection nor to change study results. Rather, the process emphasized cointerpretation with adolescents who have relevant lived experience, allowing insider perspectives to expand the interpretation and potential directions for future research.

Statistical Analysis
Basic descriptive statistics were first conducted for all major variables based on gender identity. This was done using multivariate analysis of variance (MANOVA

Preliminary Analyses
A series of MANOVAs explored gender identity differences in outcomes and SMU. There was a significant multivariate effect of gender identity differences for all outcomes measured in the study

Main Analyses
In general, time spent on social media and age at receiving first smartphone were not associated with any outcomes. Attending a school with what students perceived as strong digital literacy training, active SMU, low levels of social comparisons, and low levels of problematic SMU (eg, depression: B = 1.02; 95% CI, 0.66-1.38; P < .001) were associated with lower risk of mental health problems ( Table 2). Four results were significantly moderated by gender identity; given the focus of the study, we focus on these 4 results.
First, there was a significant interaction between gender identity and active SMU for emotional problems (B = 1.82; 95% CI, 0.16 to 3.49; P = .03). Specifically, active media use was more associated with lower emotional problems for TGNB youths than for cisgender youths (Figure 1). Specifically, depression and conduct problems were lower for TGNB youths when they reported regularly cleaning or curating their social media feed, but both depression and conduct disorders were higher for cisgender youths when they engaged in this same activity (Figure 3). Finally, we found a significant interaction between school media literacy and gender identity for depression (B = −1.07; 95% CI, −1.98 to −0.15; P = .02) with school media literacy being more associated with lower rates of depression for TGNB youths vs cisgender youths.

Discussion
There were significant gender identity differences for all health outcomes measured in the study.
TGNB youths had the highest levels of depression, emotional problems, conduct problems, and negative body image compared with cisgender youths. However, TGNB youths' use of social media was differentially associated with mental health.
Aligning with previous research on SMU, 51,52 this study found that active media use was associated with lower rates of mental health problems, especially for TGNB youths. As our advisory board suggested, TGNB youths may be more intentional about creating online spaces that are free from the negative interactions that can plague them in school 53 or at home. 9 One TGNB youth shared, "On social media, I am able to choose to be around the people that don't make me uncomfortable, that don't make me hate myself." TGNB youths can actively be themselves (present themselves) in a way that aligns with their identity via pictures and pronouns.
Relatedly, cleaning and/or curating feeds was associated with lower levels of depression for TGNB youths but higher levels for cisgender youths. TGNB youths are more likely to be bullied or harassed online and offline and may therefore have a greater need to curate safe spaces for themselves on social media. Illustrating this point, one TGNB youth said, "Real life isn't safe for   LGBTQ people, but online there is more control where I can find people who have similar beliefs."

JAMA Network Open | Pediatrics
Conversely, cleaning and/or curating feeds appeared to have an opposite pattern for cisgender youths, a finding which merits further investigation: perhaps cisgender youths fear offending peers by unfollowing those in their social circle, and thus benefit from a social media break by uninstalling apps entirely rather than curating feeds.
Taking intentional technology breaks was significantly associated with increases in depression and emotional problems for TGNB youths but not for cisgender youths, again suggesting the balance of risks and benefits of youths' SMU differs by gender identity. For TGNB youths for whom social media is a key venue for social acceptance, breaks could cut this off and potentially be detrimental to health. 9 As a TGNB advisory board adolescent explained, "I'm fine taking breaks because I already have a support group that is super nice to me. For others, when they delete it, [they delete] their safe place. That's why they feel bad…they don't have that community anymore." Attending a school with a perceived strong media literacy program was also associated with positive outcomes for all youths, 54,55 and again particularly so for TGNB youths. Given the apparent importance of online spaces for TGNB youths, these programs may contribute to protective practices and facilitate even greater intentionality around SMU. 38

Implications
Parents could be less concerned about screen time potentially causing mental health struggles in their TGNB youth and instead focus on how social media may be a resource for their children in the face of everyday minority stress. Policy makers and school officials worried about the link between social media and mental health should consider the differential associations of social media by gender identity and take a more person-centered approach. Blanket policies that severely limit SMU among youths may have different (and more negative) impacts for TGNB youths. We encourage policies (at school, state, and national levels) that focus on supporting school media literacy programs as opposed to only limiting screen time. Pediatricians might consider asking detailed questions around media use beyond screen time at well-child checkups. Clinical and medical professionals treating adolescents might consider discussing social media practices and may take a more nuanced approach depending on patient gender identity.

Limitations
This study had limitations. Major limitations included the self-report and cross-sectional nature of the data, and there were a relatively low number of TGNB youths.