Prevalence and Correlates of Sexting Behaviors in a Provincially Representative Sample of Adolescents

Abstract

Objectives:

To examine the prevalence and correlates of sending and receiving sexts (i.e., sexually explicit images) in a provincially representative sample of adolescents in Canada.

Methods:

Data from the 2014 Ontario Child Health Study, a provincial survey of households with children in Ontario, which includes a sample of 2,537 adolescents aged 14 to 17 years (mean age = 15.42, male = 51.6%) were used to address the research objectives.

Results:

The past 12 months prevalence of sending and receiving sexts was 14.4% and 27.0%, respectively. In unadjusted logistic regression analyses, non-White adolescents and those living in low-income households were less likely to send or receive sexts compared to White and non-low-income adolescents. Adolescents who disclosed their sexual and/or gender minority identities were 3 to 4 times more likely to send and receive sexts than youth who had not disclosed these identities. Higher levels of mental health problems generally observed among adolescents who sent or received sexts. In fully adjusted models, low income and ethnic minority status were associated with reduced odds of sending and receiving sexts, while sexual and/or gender minority disclosure status was associated with increased odds. Social anxiety was associated with reduced odds of sending and receiving sexts, while conduct disorder was associated with elevated odds.

Conclusion:

The prevalence of sexting behavior was higher among adolescents who disclosed their sexual or gender minority identities. Sexting behaviors were associated with higher levels of mental health problems. Identifying vulnerable populations and the potential mental health ramifications associated with sexting behavior is vital to mitigating negative sequelae.

Introduction

Due to the increased popularity of smartphones (which typically include built-in cameras) since the early 2010s, dramatic changes have been observed on how adolescents communicate with each other. Most U.S. and Canadian adolescents now own a smartphone¹ and engage in texting, online gaming, and social media daily,² although adolescents in the U.S. spend more time on screen-based media compared to Canadian adolescents.³ Among other consequences of this dramatic change in technology and communication, sexting, defined as sending and receiving sexually explicit messages and/or nude or seminude photos (i.e., sexts) via texting, social media, and other Internet services, is raising a public health concern. The prevalence of sending and receiving sexts among adolescents has been estimated to be as high as 14.8% and 27.4%, respectively.^4,5 Despite the considerable interest in this topic, including recent meta-analyses, considerable research gaps support further investigation. In particular, in a recent meta-analysis by Madigan et al.,⁵ only one study from Canada was included in comparison to 22 studies from the United States. Canadian youth differ on important demographic characteristics such as family income, parent education, and ethnic composition.^6–9 Moreover, Canadian youth hold a more positive attitude toward sexual minorities compared to youth in the United States.¹⁰ These differences may contribute to varying patterns of sexting behavior among Canadian youth, compared to their U.S. counterparts.

The primary concern of adolescents’ sexting behavior is perhaps the link between sexting behavior and emotional problems. A recent meta-analysis documented a significant, positive association between sexting behaviors and adolescent emotional problems (odds ratio [OR] = 1.79. 95% confidence interval [95% CI], 1.41 to 2.28),⁴ while earlier review articles failed to find a consistent association.^11–13 Variation in these findings may reflect methodological differences in approaches to measuring sexting and emotional problems as well as variation in the sample composition such as age groups.¹²

Fewer studies have examined the association between sexting and behavioral problems (i.e., conduct disorder [CD], oppositional defiant disorder). A recent meta-analysis, based on three studies,^14–16 reported a significant pooled effect size (OR = 2.50, 95% CI, 1.29 to 4.86) for the association between sexting and delinquent behavior.⁴ Based on problem behavior theory,^17,18 adolescents who engage in problem behaviors such as impulsive behaviors, aggressive behaviors, or substance use are more likely to report higher numbers of sexual partners and earlier age of sexual debut.¹⁹

The literature on differences between boys and girls in the prevalence of sexting behavior (sending and receiving) so far is inconsistent.^5,11 Some studies have found that boys are more likely to receive sexts and to request them, while girls are more likely to send sexts.^12,20,21 However, a recent meta-analysis indicated no significant differences between boys and girls for sending or receiving sexts.^5,22

Previous research as to whether sexual minority adolescents (e.g., adolescents with sexual identities other than heterosexual, youth who report same-sex sexual or romantic attractions or behaviors) do^23–25 or do not²⁵ report higher levels of sexting behaviors when compared with their heterosexual peers is also mixed. Moreover, the limited number of studies that do exist have focused almost exclusively on sexual minority identity (and not gender identity) in association with sexting behavior, without information regarding outness (i.e., if the individual had disclosed their sexual or gender minority status to others in their environments). Indeed, outness is an important dimension for understanding the links between sexual minority status and psychosocial outcomes^26–28 but has not been previously examined in relation to sexing behaviors. Focusing on the outness of sexual or gender minority status provides a new way of understanding sexting among sexual and gender minority adolescents.

Differences observed in sexting across race/ethnicity are also inconclusive^16,24,29,30 with some racial/ethnic minority adolescents, such as Asian/Pacific Islander adolescents report lower levels of sexting compared to White adolescents,³¹ while other racial/ethnic minority adolescents such as Black/African American adolescents report higher rates of sexting compared to White adolescents²⁹ and Hispanic adolescents.²⁴ The limited contexts in which this question has been studied (e.g., most studies are conducted in the United States where the demographic composition is very different from other countries) suggest the need to examine differences in sexting behaviors across racial/ethnic subgroups and across national contexts.

Notably, across studies that have examined the association between sexting behavior and adolescent mental health, only about half ask about receiving sexts, and in most of the studies, authors combine sending and receiving sexts as a composite variable in their models. To date, evidence suggests that adolescents who both send and receive sexts are more likely to be older in age and report higher levels of delinquent behaviors (i.e., substance use, property crime), compared to adolescents who do not send and receive sexts.⁵ The limited number of studies that have examined and differentiated between receiving and sending sexts points to the importance of examining these behaviors separately.

Despite the extant studies conducted on the prevalence and correlates of adolescent sexting behavior, the focus on American samples, as well as the broader discrepancies in the findings, suggests that evidence is needed regarding these behaviors among Canadian adolescents. The current study examined (1) the past 12-month prevalence of sexting behaviors (both sending and receiving sexual images via Internet-based communication technology) and (2) sociodemographic (sex, age, low-income status, ethnicity), identity (sexual and/or gender minority status), and mental health-related correlates (i.e., generalized anxiety disorder, major depressive disorder, social anxiety disorder, CD, and oppositional defiant disorder). The present study adds to the current literature by (1) using a well-validated, dimensional measures of five child psychiatric disorders based on Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria encompassing both emotional and behavioral problems, (2) focusing on the link between sexual/gender identity and race/ethnicity in sexting behavior, and (3) using a household-based, provincially representative sample of adolescents.

Methods

Participants and Procedures

Data from the 2014 Ontario Child Health Study (2014 OCHS), a province-wide, cross-sectional study on child and adolescent mental disorders, were used. Data were collected in the home by trained Statistics Canada interviewers from the person most knowledgeable about the child and directly from any adolescents aged 12 to 17 years. Data collection occurred from October 2014 to October 2015. Detailed accounts of the survey design, content, training, and data collection are available elsewhere.³² The present study included adolescents aged 14 to 17, who responded to the questions regarding sexting behavior. Among the eligible 3,028 adolescents, 16.2% (n = 491) were excluded from this analysis due to missing responses. The final sample for analyses include 2,537 adolescents (mean age = 15.42, male = 51.6%). No statistically significant differences were observed when comparing those included versus excluded from the analyses on all study variables. The 2014 OCHS was a voluntary survey conducted under the Statistics Act, which provides respondents with guarantees of their privacy and confidentially. Parents and children were asked for their consent to participate. The study procedures were approved by the Hamilton Integrated Research Ethics Board at McMaster University and Research Ethics Committees at participating school boards.

Measures

Dependent Variables

Sexting

Adolescents were asked two questions regarding sexting behaviors in the past 12 months: (a) sending sexts: “Have you ever sent sexual pictures of yourself to someone else over the Internet or by text (e.g., ‘sexting’)?” and (b) receiving sexts: “Has someone ever sent you sexual pictures of themselves over the Internet or by text (e.g., ‘sexting’)?” Response options included never (0); yes, more than 12 months ago (1); yes, once in the last 12 months (2); yes, more than once in the last 12 months (3); valid skip (6); and not stated (9). Response options were coded for sending (no, never sent [=0]; yes, I have sent [=1]) and receiving sexts (no, never received [=0]; yes, I have received [=1]). Original response options 1 to 3 were collapsed into the category: Yes, have received/sent sexts. While this measure does not include all elements of sexting (i.e., sexually explicit texts), the focus on image sharing is pertinent regarding the legal and social consequences of sexting behaviors among adolescents.

Sociodemographic characteristics

Standard Statistics Canada questions were administered to the parent about the child’s age (in years), sex (0 = male, 1 = female), and household income below the low-income measure (0 = not low income, 1 = low income) based on 2013 before tax cutoffs.³²

Predictor Variables

Ethnicity

Adolescents were asked about their ethnic background (i.e., White, aboriginal, South Asian, East and Southeast Asian, West Asian and Arab, Black, Latin American, and other and multiethnicities). Based on the frequency of response options (i.e., small cell sizes), ethnic groups were collapsed into two categories: White (=0) and non-White (=1). Among adolescents who were classified as “non-White (37.23% of participants),” the most common ethnic groups were East and Southeast Asians (10%), South Asians (8.3%), and Black (5.1%).

Sexual and/or gender minority disclosure status

Adolescents were asked whether they had ever “come out” to someone as gay, lesbian, bisexual, or transgender. The responses were coded into a binary variable (0 = no disclosure; 1 = disclosed). While sexual and gender identity is traditionally assessed using questions regarding identity, or in the case of sexual minority status attraction or behavior, having come out to at least one person reflects an important milestone for sexual and gender minority adolescents³³ and is an important and understudied aspect of sexual minority status within the developmental literature.³⁴

Mental health problems

Adolescents completed the OCHS Emotional Behavioral Scales (OCHS-EBS),³⁵ a 52-item symptom checklist that assesses selected DSM-5 disorders in the past 6 months. Each item is scored on a 3-point frequency scale, summed to generate scaled scores: Major Depressive Disorder (9 items), Generalized Anxiety Disorder (6 items), Social Anxiety Disorder (5 items), Conduct Disorder (11 items), and Oppositional Defiant Disorder (6 items). The items and scales of the OCHS-EBS meet adequate psychometric requirements of validity and reliability for use as dimensional measures of child and adolescent psychiatric disorders.^36,37

Statistical Analyses

Differences between adolescents who sent/received sexts versus those who did not engage in sexting behavior in mean levels of mental health problems, sexual/gender minority disclosure status, and sociodemographic correlates (i.e., age, sex, low-income status) were examined using unadjusted binary logistic regression analyses. Two regression models (one for sending and another one for receiving sexts) in which all variables were entered simultaneously were conducted to determine the independent associations between each correlate and sending and receiving sexts. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated and reported. All tests were two-tailed with a significance level set at 0.05. Logistic regression analyses were performed using statistical software STATA Version 15.0.³⁸ Sampling weights were applied to obtain population estimates, and the bootstrap method was used to obtain accurate variance estimation by taking into account the design features of the study and weight adjustments.³²

Results

The lifetime prevalence of adolescent sending and receiving sexts was 14.4% and 27.0%, respectively. Table 1 presents the prevalence of sexting (sending and receiving) by levels of mental health problems (depression, generalized anxiety disorder, social anxiety disorder, CD, and oppositional defiant disorder), ethnic minority status, gender and sexual minority disclosure status, and sociodemographic covariates (age, sex, and low income). Unadjusted OR and 95% CI are also indicated in Table 1. Notably, levels of all mental health problems, except social anxiety disorder, were consistently elevated among adolescents who endorsed sending or receiving sexts, compared to those who did not. Furthermore, no sex difference in the prevalence of either sending or receiving sexts was observed. Sexual or gender minority identity disclosure status was significantly associated with increased odds of sending and receiving sexts (OR = 4.14 [2.27 to 7.54], P < 0.001; OR = 3.44 [1.87 to 6.35], P < 0.001, respectively), while ethnic minority status (non-White) was associated with decreased odds of sending or receiving sexts (OR = 0.32 [0.21 to 0.50], P < 0.001; OR = 0.49 [0.35 to 0.68], P < 0.001, respectively). A subsequent sensitivity analysis focusing on differences between Asian, Black, and other ethnicity youth revealed that Asian adolescents were significantly less likely to report sending or receiving sexts than White youth (OR = 0.19 [1.11 to 0.34], P < 0.001). Black youth were not significantly different in their sexting behavior than White youth and other ethnicity youth (Appendix Table A1). Older age was positively associated with sending and receiving sexts (OR = 1.38 [1.16 to 1.64], P < 0.001; OR = 1.33 [1.15 to 1.53], P < 0.001, respectively), while poverty was associated with a decreased odds of sending and receiving sexts (OR = 0.57 [0.37 to 0.89], P < 0.05; OR = 0.63 [0.44 to 0.91], P < 0.05, respectively).

Table 1.

Prevalence (%/x¯) and Unadjusted Odds Ratio (95% CI) of Sociodemographic and Mental Health-Related Correlates of Lifetime Sexting Behaviors.

	Sent Sexts			Received Sexts
Study Characteristics	Yes	No	OR (95% CI)	Yes	No	OR (95% CI)
Sex (%)
Male	13.4	86.6	Ref	26.9	73.1	Ref
Female	15.5	84.5	1.18 (0.78 to 1.80)	27.2	72.8	1.02 (0.75 to 1.38)
Low-income status (%)
No	15.7	84.3	Ref	28.7	71.3	Ref
Yes	8.6	91.4	0.50 (0.33 to 0.77)**	19.1	80.9	0.59 (0.43 to 0.80)**
Ethnicity (%)
White	18.8	81.2	Ref	32.0	68.0	Ref
Non-White	7.1	92.9	0.32 (0.21 to 0.50)***	18.7	81.3	0.49 (0.35 to 0.68)***
Sexual/gender minority disclosure status (%)
No	13.2	86.8	Ref	25.6	74.4	Ref
Yes	38.6	61.4	4.14 (2.27 to 7.54)***	54.2	45.8	3.44 (1.87 to 6.35)***
Mental health problems (M, SE)
Major Depressive Disorder	4.82 ± 0.34	2.72 ± 0.11	1.16 (1.11 to 1.22)***	4.21 ± 0.23	2.59 ± 0.12	1.14 (1.10 to 1.18)***
Generalized Anxiety Disorder	4.82 ± 0.30	3.19 ± 0.12	1.16 (1.10 to 1.23)***	4.09 ± 0.21	3.18 ± 0.13	1.09 (1.04 to 1.15)***
Social Anxiety Disorder	3.32 ± 0.24	3.06 ± 0.10	1.04 (0.97 to 1.10)	3.12 ± 0.18	3.09 ± 0.11	1.00 (0.95 to 1.06)
Conduct Disorder	2.05 ± 0.17	0.89 ± 0.05	1.41 (1.29 to 1.54)***	1.87 ± 0.12	0.76 ± 0.05	1.55 (1.38 to 1.75)***
Oppositional Defiant Disorder	3.19 ± 0.22	2.00 ± 0.08	1.22 (1.14 to 1.31)***	3.02 ± 0.17	1.86 ± 0.07	1.24 (1.17 to 1.31)***

Note. CI = confidence interval; M = mean; SE = standard error; OR = odds ratio.

*P < 0.05. **P < 0.01. ***P < 0.001.

Table 2 presents the adjusted ORs (aORs) and 95% CI for the associations between sending and receiving sexts, race/ethnic minority status, sexual/gender minority disclosure status, and mental health problems. After adjusting for covariates, sexual and gender minority disclosure status remained significantly associated with an increased odds of sending and receiving sexts (OR = 2.26 [1.10 to 4.68], P < 0.05; OR = 2.30 [1.27 to 4.17], P < 0.01, respectively) compared to adolescents who had not disclosed these statuses. Being non-White was associated with a reduced odds of sending and receiving sexts (OR = 0.35 [0.22 to 0.55], P < 0.001; OR = 0.49 [0.34 to 0.70], P < 0.001, respectively), in comparison to adolescents who identified as White.

Table 2.

Associations between Sexting and Adolescent Mental Health Problems, Adjusted Odds Ratios and 95% Confidence Intervals.

	Sending	Receiving
Sociodemographic
Age	1.38 (1.16 to 1.64)***	1.33 (1.15 to 1.53)***
Female sex	0.98 (0.62 to 1.57)	0.98 (0.69 to 1.39)
Low-income status	0.57 (0.37 to 0.89)*	0.63 (0.44 to 0.91)*
Identity
Non-White	0.35 (0.22 to 0.55)***	0.49 (0.34 to 0.70)***
Sexual/gender minority disclosure status	2.26 (1.10 to 4.63)*	2.30 (1.27 to 4.17)**
Mental health problems
Major Depressive Disorder	1.07 (0.98 to 1.16)	1.09 (1.01 to 1.17)*
Generalized Anxiety Disorder	1.10 (0.98 to 1.24)	0.98 (0.89 to 1.07)
Social Anxiety Disorder	0.86 (0.80 to 0.94)**	0.88 (0.82 to 0.95)**
Conduct Disorder	1.35 (1.20 to 1.52)***	1.41 (1.20 to 1.64)***
Oppositional Defiant Disorder	1.00 (089 to 1.12)	1.10 (0.98 to 1.22)

*P < 0.05. **P < 0.01. ***P < 0.001.

Adolescents with conduct problems were more likely to send and receive sexts (OR = 1.35 [1.20 to 1.52], P < 0.001; OR = 1.41 [1.20 to 1.64], P < 0.001, respectively), whereas social anxiety disorder symptoms were associated with a decreased odds of sending and receiving sexts (OR = 0.86 [0.80 to 0.94], P < 0.01; OR = 0.88 [0.82 to 0.95], P < .01, respectively).

Discussion

The current study examined the lifetime prevalence and correlates of sending and receiving sexually explicit images in a provincially representative sample of adolescents in Canada. Prevalence estimates of sending (14.4%) and receiving (27.0%) these images are remarkably similar to rates of sexting reported in a recent systematic review (14.8% and 27.4%).⁵ Consistent with past studies,⁵ the prevalence of sending and receiving sexts was similar for boys and girls, while older age was associated with a higher prevalence of sexting behavior.^5,39 Ethnic minority status was associated with lower rates of sexting behavior compared to White adolescents, and adolescents who had disclosed a gender and/or sexual minority identity reported higher rates of sexting behaviors.

In unadjusted analyses, sexting behavior was associated with higher levels of depression in adolescents, consistent with prior studies.^29,40,41 Dake et al.,²⁹ however, found a positive association between sending sexts and feeling sad or hopeless for more than 2 weeks in the past year, and the association remained significant after adjusting for demographic factors (i.e., age, sex, Socioeconomic status, and ethnicity). Besides the different coverage of depressive symptoms measured between the two studies (comprehensive vs. sad or hopeless), it is essential to note that while Temple et al.⁴² included anxiety disorder in the regression model, Dake et al.²⁹ did not control for anxiety symptoms. Similarly, Frankel et al.⁴⁰ and Van Ouytsel et al.⁴¹ found a significant association but did not include anxiety or sexual behaviors in their models. These findings underscore the importance of using comprehensive depression measures and controlling for anxiety in future studies examining the association between sexting behavior and adolescent mental health problems.

Current findings contribute to an emerging body of evidence suggesting a significant association between sending and receiving sexts and behaviors associated with CD such as property destruction and fighting.^4,14–16 Specifically, the current results indicate that sending and receiving sexts are significantly associated with higher symptoms of CD. The association between sexting and CD may also be explained with the problem behavior theory,^17,18 in that different problem behaviors (e.g., sexting, violence, substance use, earlier sexual debut, more sexual partners) are the consequences of a similar mechanism/origin such as impulsivity. While communicating about sexuality, either online or in-person is normative for adolescents,^11,22 youth with CD may be more likely to engage in sexting, similar to the ways in which they are more likely to engage in higher rates of other types of normative sexual behavior at an earlier age, and with more partners, when compared to youth without these problems.^19,43,44

The significant association between sexual and gender minority disclosure status and sexting behavior is in line with the previous studies that report more prevalent sexting behavior among sexual and gender minority adolescents.^{23–25,45–47} Online communication about sex, including sexting, may differ among sexual and gender minority populations compared with heterosexual populations, due to fewer available partners, as well as more privacy concerns about partners met in offline environments.^45,47 Hence, sexual and gender minority adolescents may engage more in sexual exploration and sexting behavior online as an alternate way of seeking intimacy.^11,22 The relatively higher prevalence of sexting behavior in sexual and gender minority adolescents compared to heterosexual adolescents speaks to the need to address the specific needs of these youth pertaining to sexting.^22,45,46

Regarding the association between ethnic minority and sexting,^{5,24,25,29,30,48} results indicate that nonWhite adolescents (who in this study were mostly Asians) engaged in less sexting behavior than White adolescents. Current findings may reflect significant differences in the race/ethnic composition of populations in Canada versus the United States, where most of the previous research was conducted. More research is needed to understand why some race/ethnic minority adolescents in Canadian contexts report lower levels of sexting behaviors compared to their White peers and if variation is observed among different race/ethnic minority groups.

The present results have important implications not only for parents but also for policy makers, educators, and clinicians working with adolescents. Given the prevalence of sexting found in the current study, it is essential to acknowledge that sexting can be a common part of peer interactions for adolescents. Parents, educators, and clinicians should actively monitor adolescents and be aware of the possible adverse correlates of sexting behaviors. Furthermore, advising adolescents of the legal consequences of nonconsensual sharing of sexually explicit content is critical to protect adolescents. We recommend that clinicians be informed of the concerns associated with sexting behavior and ask adolescents questions about sexting, including ongoing monitoring if they are following them.⁴⁹ From a policy perspective, one specific recommendation is to develop and provide position statements to address sexting, and perhaps the overall landscape of Digital Health and Safety Literacy for adolescent.

There are several limitations to the current study. The cross-sectional design precludes our ability to understand the temporal ordering between sexting and adolescent mental health problems. Furthermore, the sexting items used in the current study only include sexual pictures and do not include sexually explicit texts. Also, because the current data were collected in 2014 and 2015, our findings do not reflect the most up-to-date prevalence. The relational context in which sexting takes place was also not assessed in the current study. As such, it is not possible to understand whether adolescents were reporting nonconsensual sexting, sexting with adults, or sexting in the context of romantic relationships, which should be addressed in future studies. Although this study provides important preliminary information regarding sexting, further research is necessary to better understand how the correlates identified in this study are related to different types of sexting, and if these different types of sexting have different consequences for adolescent well-being. The current study assessed sex by asking participants if they were male or female. The use of these restricted options may have been discouraging for youth with other gender identities and may have influenced their disclosure of sexual or gender identity later in the survey. Although 37.23% of the participants in the current study identified as having race/ethnicities other than White, significant variation was observed regarding the race/ethnicities endorsed. Due to limited cell size, examining differences among non-White participants was not possible, but future research, using more targeted participant recruitment, may clarify the current findings. Finally, the current study focused on disclosure status of sexual or gender minority identity, and no information about the extent of disclosure, or sexual or gender minority adolescents who had not disclosed their identity to others is available.

Conclusion

Despite the limitations noted earlier, the current study extends the existing sexting literature in several important ways. First, we used a household survey that involved cluster sampling and stratification by residency and income. The sample thus represents the general adolescent population in the province of Ontario, Canada, and can be generalized to a large sample of adolescents. Second, the current study also investigated both sending and receiving sexts extending a previous literature which focused primarily on sending sexts. Most importantly, the current study included ethnic, gender, and sexual minority disclosure status to provide a deeper understanding of the extent to which sexting behaviors vary across potentially vulnerable subgroups in the population. Our findings suggest that prevalence and mental health correlates of sexting behavior among Canadian adolescents are parallel to the American counterparts. However, visible ethnic minority status was significantly associated with a reduced odds of sexting behavior in Canadian adolescents compared to the previous findings mostly with American adolescents. Furthermore, sexual and/or gender minority adolescents are at a higher risk of sexting behaviors. Parents, educators, and clinicians in Canada should acknowledge sexting as a common part of peer interactions and regularly communicate with adolescents regarding sexting and its possible implications.

Appendix

Table A1.

Associations between Sexting and Adolescent Mental Health Problems, Adjusted Odds Ratios and 95% Confidence Intervals.

Study Characteristics	Sending	Receiving
Sociodemographic
Age	1.39 (1.17 to 1.65)***	1.33 (1.18 to 1.54)***
Female sex	0.98 (0.61 to 1.55)	0.98 (0.69 to 1.38)
Low-income status	0.56 (0.36 to 0.87)*	0.61 (0.42 to 0.88)**
Identitys
Asian	0.19 (0.11 to 0.34)***	0.31 (0.17 to 0.58)***
Black	0.73 (0.23 to 2.36)	0.98 (0.47 to 2.06)
Other ethnicity	0.48 (0.27 to 0.83)**	0.65 (0.43 to 0.97)
Sexual/gender minority disclosure status	2.24 (1.10 to 4.58)*	2.29 (1.25 to 4.17)**
Mental health problems
Major Depressive Disorder	1.06 (0.97 to 1.16)	1.09 (1.01 to 1.17)*
Generalized Anxiety Disorder	1.11 (0.99 to 1.24)	0.98 (0.90 to 1.08)
Social Anxiety Disorder	0.87 (0.80 to 0.94)**	0.89 (0.82 to 0.96)**
Conduct Disorder	1.34 (1.19 to 1.51)***	1.41 (1.19 to 1.66)***
Oppositional Defiant Disorder	0.99 (0.89 to 1.12)	1.09 (0.98 to 1.23)

*P < 0.05. **P < 0.01. ***P < 0.001.