Relationship trajectories and psychological well-being among sexual minority youth

Abstract

Dating in adolescence plays an integral part in the development of sexual and social identities. This process is particularly salient for sexual minority youth who face additional obstacles to their identity formation due to their marginalized status. We investigated the influence of participating in a same-sex relationship (SSR) or an opposite-sex relationship (OSR) on sexual minority youths' psychological well-being (i.e., symptoms of depression, anxiety and internalized homophobia, and self-esteem) in an ethnically-diverse sample of 350 youth (55% male) between the ages of 15-19 years, recruited from three GLBT drop-in centers in the New York City area. Using longitudinal data, we examined youths' SSR and OSR over time. Multivariate regression analyses suggest that involvement in a SSR was positively associated with changes in self-esteem in males, and negatively correlated with changes in internalized homophobia in females. We discuss the implications for positive development in sexual minority adolescent populations.

Historically, stigma surrounding sexuality has placed a burden on individuals who do not self-identify as heterosexual (i.e., collectively referred to as “sexual minorities”; Meyer, 2003). Sexual minority youth experience greater psychological distress than heterosexual youth (Mays & Cochran, 2001; Meyer, Dietrich & Schwartz, 2007) and more victimization, especially related to their sexual identity (e.g., verbal taunts, physical threats, physical violence; Garofalo, Wolf, Kessel, Palfrey, & DuRant, 1998; Rivers & D'Augelli, 2001). Actual and anticipated prejudicial and discriminatory treatment may come from family, peers, and others in their communities (Bauermeister, Morales, Seda & Gonzalez-Rivera, 2008; Herek & Capitano, 1999). Sexual minority youth who conceal their sexuality may delay the formation of a positive self-concept (Bos, Sandfort, Bruyn, & Hakvoort, 2008), which may promote discomfort with or rejection of their sexuality (i.e., internalized homophobia; Shidlo, 1994). Sexual minority youth grappling with low self-esteem and internalized homophobia often also experience heightened levels of depression and anxiety (Bos et al., 2008; Igartua, Gill & Montoroyr, 2003; Szymanski, Chung, & Balsam, 2001). These experiences may affect sexual minority youths' adolescent development, and have implications for the 2-3% of high school age youth who self-identify as gay, lesbian, or bisexual (GLB), as well as the additional 11% who are unsure of their sexual orientation and explore their sexuality through same-sex relationships (SSR), opposite-sex relationships (OSR), or both (Graber & Archibald, 2001). Given the prevalence of SSR and OSR relationships during adolescence, we sought to examine in this study the association between psychological well-being and participation in SSR and OSR over time in a sample of sexual minority youth.

Experiences of psychological distress may diminish or disappear in the presence of interpersonal ties that offer social support (Peplau & Fingerhut, 2007). In this study, we pay attention to youths' participation in dating as it provides them with a unique and intimate source of social support, plays an integral role in the renegotiation of parent and peer relationships, and fosters a sense of independence and self-sufficiency (Furman & Shaffer, 2003). On the other hand, some researchers have suggested that participation in a romantic relationship may also increase anxiety and depressive symptoms among youth, particularly if youth are dissatisfied with these relationships, encounter conflict in their interactions with partners, and/or experience difficult relationship breakups (La Greca & Harrison, 2005; Russell & Consolacion, 2003). Consequently, participation in dating relationships may increase, rather than buffer, psychological distress. Given that the onset of dating behaviors is one of the milestones of adolescent development and may be associated with psychological well-being, we sought to examine the prevalence and characteristics of SSRs and OSRs in a sample of self-identified sexual minority youth during adolescence, and to test whether their participation in these relationships were associated with four psychological outcomes (i.e., self-esteem, internalized homophobia, and symptoms of depression and anxiety) over time.

The Role of Romantic Relationships During Adolescence

Researchers have recognized the importance of adolescent romantic relationships (Davila, Steinberg, Kachadourian, Cobb, & Fincham, 2004), and distinguished these relationships from other close relationships (e.g., parents and peers) in youths' lives (La Greca & Harrison, 2005). Dating and experiencing romance during adolescence are essential in identity formation (Furman & Shaffer, 2003), and may help increase youths' self-esteem by knowing others find them attractive. Adolescent relationships also allow youth to explore their romantic interests, develop and refine relationship scripts based on intimacy, passion, and commitment (Connolly, Furman, & Konarski, 2000), and explore romance with different types of partners in order to define what character traits they desire in a partnership (Bouchey & Furman, 2003). Given that romantic explorations may have long-lasting influences across the lifecourse, researchers have acknowledged the importance of examining the development and trajectory of adolescent romantic relationships (Furman & Shaffer, 2003).

Participation in romantic relationships during adolescence has been linked to positive and negative psychological outcomes among heterosexual youth. La Greca and Harrison (2005) found that heterosexual adolescents who were not dating were more likely to report anxiety symptoms compared to youth who were dating, irrespective of the qualities of these romantic relationships or youths' peer relationships. Conversely, Davila and colleagues (2004) found that heterosexual youth who were dating were more likely to report a higher number of depressive symptoms than non-dating peers; however, this association disappeared when youths' evaluations of these relationships were taken into account. Taken together, these findings would suggest that positive dating experiences may be beneficial for heterosexual youth during adolescence.

Beyond the developmental advantages of romantic relationships seen among heterosexual youth, forming a close bond with someone experiencing similar challenges may also serve to mitigate sexuality-related stress (Detrie & Lease, 2007). While participation in same-sex relationships has been posited to improve psychological well-being among adult gay men (Isay, 2009), findings linking psychological well-being with sexual minority youths' dating experiences are scarce (Russell & Consolacion, 2003). In a unique study using data from the National Longitudinal Study of Adolescent Health (AddHealth), Russell and Consolacion (2003) examined youths' psychological well-being over a 12 month period, and compared it across sexual attractions (i.e., heterosexual or sexual minority) and relationship status (i.e., single, same-sex, opposite-sex). In their analyses, Russell and Consolacion found that 9% of the nationally-representative sample reported having a same-sex relationship at the time of the study. When they examined the differences in psychological symptoms across the six attraction-relationship groups, they found that sexual minority youth in a SSR reported comparable anxiety scores as heterosexual-single youth; however, anxiety scores were higher among the remainder of the sample. Furthermore, when depressive symptom scores were examined, sexual minority youth, irrespective of their relationship status, reported higher depressive symptom scores than heterosexual singles; nevertheless, mean depression scores for sexual minority youth in a SSR were comparable to heterosexual youth in an OSR. Taken together, Russell and Consolacion's findings provide evidence for the salience of SSRs and OSRs among sexual minority youth as well as the potential risks and protections that they may offer.

Accumulating evidence suggests that many sexual minority youth participate in romantic relationships with same-sex (Russell & Consolacion, 2003), as well as with opposite-sex (Diamond & Savin-Williams, in press), partners during adolescence. Unfortunately, previous studies have not accounted for youths' involvement in SSRs and OSRs concurrently. As a result, it is difficult to ascertain whether the links between adolescents' dating and psychological well-being vary depending on whether sexual minority youth participate exclusively in same-sex or opposite-sex relationships, or in both. Sexual minority youth may engage in OSRs to reduce the stigma and negative peer responses associated with their evolving sexual identity, or simply as a reflection of their attractions to both males and females at any given point during adolescence (Glover, Galliher, & Lamere, 2009). In the absence of tolerant environments where sexual minority youth may explore their sexuality openly, these alternative relationships may help to fulfill emotional needs, facilitate youths' exploration of their sexual identity, and provide a rubric for adult relationships (Glover et al., 2009). Failure to account for these opposite-sex relationships may mask or overestimate the association between psychological well-being and participation in SSRs. Furthermore, participation in OSRs may uniquely contribute to sexual minority youths' psychological well-being. For example, while participation in OSRs may increase sexual minority youths' self-esteem, it may also be associated with increased internalized homophobia, depression, or anxiety – particularly if youth feel ashamed or unable to explore their same-sex or bisexual desires. Consequently, in this study, we also examined the prevalence of OSRs during adolescence, and assessed whether involvement in OSRs was associated with changes in psychological well-being over time.

Our study builds upon the existing literature in several ways. First, we acknowledge that dating during adolescence may offer opportunities for risk and protection. Sexual minorities in SSRs, for instance, may have the additional advantage of improving their self-esteem and decreasing their internalized homophobia. Consequently, we include self-esteem and internalized homophobia alongside the two indicators of psychological distress (i.e., depressive and anxiety symptoms) previously examined. Second, we examine the change in psychological outcomes from a longitudinal perspective (i.e., mean scores at Time 2 after accounting for Time 1) and account for potential confounders, including the availability of social support and youths' disclosure of sexual identity to family and peers. Third, given that youths' relationship status may change over time, we include both baseline and follow-up relationship status in our analyses. In order to examine these relationships in greater detail, we employed a trajectory measurement approach that considers the changes in youths' SSRs over time. This trajectory measure allowed us to examine whether the exposure (i.e., how often) and timing (i.e., when) of SSRs had differing effects on sexual minority youths' psychological well-being over time. For example, participation in SSRs earlier in adolescence may be more likely among youth with greater self-esteem and lower internalized homophobia than among youth who participate in SSRs later in adolescence. These associations, however, may also vary based on whether youth participate in SSRs consistently or episodically over time. For example, youth who report having engaged in SSRs consistently over time (termed SSR-SSR in analyses) may report fewer depressive symptoms than youth whose relationship dissolved (i.e., SSR1 for youth who engaged in a SSR at Time 1 [earlier in adolescence] and not at Time 2), who have recently started a relationship (i.e., SSR2 for youth who engaged in a SSR at Time 2 [later in adolescence] and not at Time 1), and even fewer symptoms than sexual minority youth who did not engage in a SSR over time (termed NoSSR). Finally, given the possibility that some youth engaged in both SSRs and OSRs over time, we examined whether the associations between SSRs trajectories and psychological well-being changed when OSRs were included in our analyses.

Sex differences in sexual minority youths' relationship experiences

Gender must be accounted for when examining youths' relationship experiences. Tolman's (2006) compulsory heterosexuality framework suggests that youth learn the societal expectations of male/female marriage and the socialization of both genders for these roles early on. Males are encouraged to foster an assertive sexuality in order to expedite the onset of sexual activity with their partners (Tolman, Striepe, & Harmon, 2003), whereas females are encouraged to avoid voicing their beliefs and default to male opinions and preferences instead. This socialization leads to increasing disadvantages for females and places them as socially inferior to males. These gender disparities may help to explain the decreases noted in females' self-esteem across adolescence (Bauermeister, Zimmerman, Caldwell, Xue & Gee, 2009; Galambos, Barker & Krahn, 2006), as well as the increase in gender-specific social demands (Galambos, Almeida & Petersen, 1990; Wingood & DiClemente, 2002). Consequently, we hypothesized that females would report lower psychological well-being than males.

Demands of gender scripts may also affect sexual minority youths' sexual identity (Diamond & Savin-Williams, 2003; Schneider, 2001). Traditional male gender roles do not allow for the expression of non-normative/same-sex emotional or sexual attraction (Russell & Consolacion, 2003), and may explain why male sexual minority youth delay disclosing their sexuality to others even though they report recognizing same-sex attraction at an early age (Diamond & Savin-Williams, 2003; Russell & Consolacion, 2003). Compared to their female counterparts, however, males may have greater losses in their social position if they express their same-sex desires. Given that females are more likely to be perceived in society as sexual objects and allotted greater opportunities to express both opposite and same-sex desires, they may be subjected to fewer social penalties for deviating from traditional social roles when they express interest in SSRs. This may explain why there is greater variation in reported age of first same-sex attraction and the claiming of different sexual identities over time among females (Diamond & Savin-Williams, 2003; Schneider, 2001). Consequently, given that dating relationships serve to socialize youth during adolescence and gender-based stressors may influence youths' participation in SSRs and OSRs and psychological well-being, respectively, we examined the relationship between psychological well-being and SSRs and OSRs separately for males and females.

Study Objectives and Hypotheses

This study explored the association between SSR and OSR trajectories and psychological well-being using longitudinal data collected from sexual minority youth recruited from three GLBT drop-in centers in the New York City area. This study had three objectives. First, we sought to determine the frequency of SSR and OSR among sexual minority youth. Consistent with previous research, we expected SSRs and OSRs would be common among sexual minority youth (Russell & Consolacion, 2003); however, consistent with Tolman's framework, we expected OSRs would be more likely among female youth (Diamond & Savin-Williams, 2003).

Second, we sought to examine the association between SSR trajectories and psychological well-being (i.e. self esteem, internalized homophobia, and depressive and anxiety symptoms) over time. Given that greater participation in SSRs could reflect (and reinforce) youths' comfort with their sexuality, we hypothesized that youth who reported consistently participating in SSRs over time would report greater psychological well-being (i.e., greater self-esteem, less internalized homophobia, and anxiety and depressive symptoms) than youth participating in SSRs at only one time point (e.g., SSR1 or SSR2), and even greater well-being than those who had not reported a SSR over either period (e.g., NoSSR). Given the exploratory nature of our study, however, we had competing hypotheses regarding the association between psychological well-being and the timing of SSRs. On the one hand, we hypothesized that youth in the SSR1 trajectory would report greater psychological well-being scores than those who engaged in the SSR2 trajectory because they would have had more time to grow comfortable with their sexuality. To support this hypothesis, youth in the SSR1 trajectory would have to note increases in self-esteem and decreased internalized homophobia, depressive and anxiety symptom scores at Time 2. As a counterargument, however, we found it equally plausible that youth in the SSR1 trajectory would report lower psychological well-being scores than those who engaged in SSR2 trajectory due to a negative SSR1 experience and/or the dissolution of the Time 1 relationship(s). This would be reflected by decreased self-esteem and/or increased internalized homophobia, depressive and anxiety symptoms at Time 2 among youth in the SSR1 trajectory. Furthermore, individuals in the SSR2 trajectory would report increased self-esteem and decreased internalized homophobia, and depressive and anxiety symptoms because these youth could be better equipped to have a SSR if they had postponed initiating a relationship until they were older. Consistent with Tolman's framework, we expected the magnitude of these relationships to be greater for females than males across psychological outcomes.

Finally, we examined whether the inclusion of sexual minority youths' participation in OSRs over time was associated with psychological well-being. Consistent with findings from the heterosexual literature, we hypothesized that participation in OSRs over time would increase youths' self-esteem. Departing from the heterosexual literature, however, we expected that sexual minority youth engaged in OSRs over time would be more likely to report increased internalized homophobia, and depressive and anxiety symptoms. We expected the magnitude of these associations to be greater for sexual minority males than females, as they may reflect the stricter social definitions ascribed to masculinity.

Methods

Participants

Participants were recruited in three New York area drop-in centers providing social, recreational, and educational services to GLBT youth (D'Augelli & Grossman, 2006). Once enrolled, participants completed a baseline interview, and were asked back for two follow up interviews each a year apart. There was some participant attrition across the first two waves: 528 interviews were completed at Time 1, 358 interviews at Time 2, and 225 interviews at Time 3. We excluded eight participants who identified as transgender at Time 2 as their experiences differ from same-sex attracted male or female identified youth. For the purposes of this study, we analyzed Time 1 and 2 data to ensure sufficient statistical power for our analyses (n=350). Study attrition was mostly due to loss at follow-up (i.e., telephone numbers had been disconnected or if listed contacts changed between interviews). There were no significant sociodemographic differences in the sample's composition across the two waves (see Results section) included in this analysis.

Procedure

Participants were recruited if they participated in one or more of the programs offered at the drop-in centers. Due to concerns that seeking parental consent from GLBT youth might inadvertently expose their sexual orientation and potentially lead to verbal or physical harm, a waiver of the parental consent requirement was obtained from the IRBs from New York University and Pennsylvania State University. In place of parental consent, eligible youth were provided with an advocate at each site to assist in understanding the nature of participation in the project and to answer questions. Once enrolled, a master's-level mental health clinician interviewed youth in a private room onsite. Bilingual interviewers were available for Hispanic/Latino youth who preferred to be interviewed in Spanish.

Interviewers followed a structured interview protocol. Interviews lasted from two to three hours. Given the length of the interview, data collection was divided into two sections. In the first section of the interview, youth completed a battery of standard mental health questions. Youth then received a 10 minute break. In the second section, youth answered a series of questions relating to their interpersonal relationships, sexuality, and health behaviors including substance use and sexual behaviors. Youth were offered two optional breaks during the second section of the interview. For sensitive information (i.e., drug use and sexual behavior), youth completed a paper-and-pencil self-reported questionnaire. Participants received an incentive of $30 for the first wave, and $40 and $50 for the second and third waves, respectively.

Measures

Self-Esteem

Participants completed the Rosenberg Self-Esteem Scale (Rosenberg, 1979). This 10-item scale asks participants to rate from 1 (“Strongly Agree”) to 4 (“Strongly Disagree”) how much they agree with statements assessing global self esteem (e.g., “All in all, I am inclined to feel that I am a failure” and “I feel that I have a number of good qualities”). Mean scores were used in our analyses (T1: α = .85, T2: α = .86). Where appropriate, items were reverse-coded so that higher scores reflected greater self-esteem.

Internalized homophobia

This construct was assessed using the 12-item scale developed by Shidlo (1994). We created a mean composite score (T1: α = .79, T2: α = .84) based on participants' agreement to statements such as “I am proud to be a part of the GLBT community” and “There have been times when I've felt so rotten about being GLBT that I wanted to be dead.” Participants rated the statements using a 4-point scale ranging from 0 (“Strongly Disagree”) to 3 (“Strongly Agree”). Where appropriate, items were reverse-coded so that higher scores reflected greater internalized homophobia.

Depression

Participants completed the Beck Depression Inventory (BDI-II; Beck, 1996). This scale includes 21 statements assessing different symptoms of clinical depression (e.g. loss of pleasure, crying, and changes in appetite). Participants chose the statements that best reflect the way they have felt over the past two weeks. We created a composite score where higher scores meant having a greater number of depressive symptoms (T1: α = .91, T2: α = .91).

Anxiety

Participants completed the Brief Symptom Inventory (Derogatis, 1993). We used the 6 anxiety-related statements (i.e., “Nervousness or shakiness inside,” “Suddenly scared for no reason,” “Feeling fearful,” “Feeling tense or keyed up,” “Spells of terror or panic,” and “Feeling so restless you couldn't sit still”) to compute a mean composite score for anxiety symptoms. Respondents rated how much each listed symptom distressed them in the past week using a 5-point scale ranging from 0 (“Not at all”) to 4 (“Extremely”). Higher scores reflected greater anxiety symptoms (T1: α = .79; T2: α = .83).

Demographics

Sociodemographic information was collected on sex, age, area of residence (city, suburb, or small town/rural), educational attainment (last year of education completed), and race/ethnicity.

Dating Relationships

Participants answered a series of questions about dating and relationships (D'Augelli, 2002). First, they were asked whether they currently had a boyfriend or girlfriend. If in a relationship, youth were asked follow-up questions about the partners' sex, age of their partner (“Is this person older than you, the same age, or younger than you?”), the duration of this relationship (“How long have you been together?”; 1 = Less than a month to 5 = six years or more), and their overall satisfaction with the relationship (“How satisfied are you with your relationship with this partner?”; 1=Satisfied to 4 = Dissatisfied). Based on the sex of the partner, we created dummy variables to indicate whether youth were currently in a same-sex (SSR) or opposite-sex relationships (OSR), respectively. If participants had missing data on either variable, we assigned them a value of ‘No’ to provide a more conservative estimate (i.e., decreasing our chances of detecting an association between relationship participation and psychological well-being).

Specific to same-sex dating behaviors, participants were asked at each interview: “How many serious same-sex relationships (lasting more than 3 months) have you had in the past year?” We used this variable to compute four possible SSR trajectories: no SSR over the two Waves (NoSSR); had one or more SSRs at Time 1 but none at Time 2 (SSR1); had one or more SSRs at Time 2 but none at Time 1 (SSR2); and, had one or more SSR at both Waves (SSR-SSR).

For opposite-sex dating behaviors, participants were asked at each interview: “How many serious opposite-sex relationships (lasting more than 3 months) have you had in the past year?” When we tried to create the four possible OSR trajectories, we found that few participants reported engaging in OSRs at either time point. This restricted the cell size of each OSR trajectory. To minimize statistical bias due to inadequate power, we decided to include a dichotomous variable indicating whether a participant had reported one or more OSRs at each Wave in our analyses (see Table 2).

Table 2.

Same-sex relationship characteristics among males and females

	Wave 1			Wave 2
	Males N(%)	Females N(%)	χ2	Males N(%)	Females N(%)	χ2
Currently in SSR (y/n)	69(36.3%)	61(38.1%)	0.12	71(37.4%)	47(29.3%)	2.48
Age of Current SSR Partner			0.12			2.93
Older	37(56.1%)	34 (55.7%)		40 (56.3%)	23 (48.9%)
Same Age	9 (13.6%)	7 (11.5%)		16 (22.5%)	7 (14.9%)
Younger	20 (30.3%)	19 (31.1%)		15 (21.1%)	16 (34.0%)
Duration of Current SSR			0.74			13.54**
< 1 month	25(37.9%)	19(31.1%)		14 (19.7%)	2(4.3%)
1-6 months	24(36.4%)	26(42.6%)		34 (47.9%)	14(29.8%)
≥ 7 months	17(25.7%)	15(24.6%)		23 (32.4%)	30(63.8%)
Satisfaction with Current SSR			1.42			0.91
Satisfied	46 (69.7%)	39 (63.9%)		52 (75.4%)	33 (70.2%)
Some sat.	14 (21.2%)	15 (24.6%)		10 (14.1%)	8 (17.0%)
Some dis.	6 (9.1%)	5 (8.2%)		4 (5.6%)	4 (8.5%)
Dissatisfied	0 (0%)	1 (1.6%)		3 (4.2%)	1 (2.1%)
Number SSR			6.09*			5.07
(past year)	117	77 (48.1%)		99 (52.1%)	99 (61.9%)
0	(61.6%)	60 (37.5%)		70 (36.8%)	51(31.9%)
1	51 (26.8%)	20(12.5%)		21 (11.1%)	9(5.6%)
2 or more	21 (11.1%)
SSR Trajectory						11.75**
(past year)
NoSSR				73 (38.4%)	61(38.1%)
SSR1				26 (13.7%)	40(25.0%)
SSR2				40 (21.1%)	20(12.5%)
SSR- SSR				46 (24.2%)	39(24.4%)
Currently in OSR (y/n)	12(6.3%)	32(20.0%)	14.80**	8(4.2%)	41(25.6%)	35.33**
Age of Current OSR Partner1			4.93			9.35**
Older	5(45.5%)	25(80.6%)		2(28.6%)	33(82.5%)
Same Age	4(33.3%)	4(12.9%)		2(28.6%)	5(12.5%)
Younger	2(16.7%)	2(6.5%)		3(42.8%)	3(7.5%)
Duration of Current OSR1			0.19			2.91
< 1 month
1-6 months	3(27.3%)	8(25%)		2(28.6%)	3(7.5%)
≥ 7 months	4(36.4%)	14(43.8%)		2(28.6%)	14(35.0%)
	4(36.4%)	10(31.2%)		3(42.8%)	23(57.5%)
Satisfaction with Current OSR1			1.45			.92
Satisfied	7(63.6%)	22(68.8%)		4(57.2%)	26(65.0%)
Some sat.	3(27.3%)	7(21.9%)		2(%)	9(22.5%)
Some dis.	1(9.1%)	1(3.1%)		1(%)	5(12.5%)
Dissatisfied	0	2(6.3%)		0	0
Number OSR			25.71**			18.35**
(past year) 1
0	166(87.4%)	106(66.7%)		165(88.2%)	126(78.8%)
1	20(10.5%)	46(28.9%)		11(5.9%)	29(18.1%)
2 or more	4(2.2%)	7(4.4%)		11(5.8%)	3(1.9%)
OSR Trajectory						37.48**
(past year)
NoOSR				173(91.1%)	108(67.5%)
OSR1				10(5.3%)	11(6.9%)
OSR2				5(2.6%)	20(12.5%)
OSR- OSR				2(1.1%)	21(13.1%)

^*p ≤ .05,

^**p ≤ .01

Note. ‘SSR’ = same-sex relationship; ‘OSR’ = opposite-sex relationship; ‘SSR1’ = youth report one or more serious (more than 3 months) SSR in the past year at baseline and no SSR in the past year at follow-up; ‘SSR2’ = youth report no serious SSR in the past year at baseline and one or more SSRs in the past year at follow-up; ‘SSR-SSR’ = youth report one or more SSRs in the past year at both baseline and follow-up. Similar trajectories were computed for OSRs.

¹Percentages computed based on cases with complete data.

Sexual attraction

Given that sexual orientation and attraction do not always map perfectly onto one another (Coleman, 1987), participants were asked define their sexual identity and attraction toward people of the same or opposite sex using two indicators used in previous studies with this population (D'Augelli, 2002). The Kinsey scale (Kinsey, Pomeroy & Martin, 1948) was used to ask participants to place themselves on a range from 0 (“Totally straight”) to 6 (“Totally gay or lesbian”). We also used a 5-point item asking participants rating their level of attraction to males over the past year on a range from 0 (“Not at all”) to 4 (“Extremely”).

Social Support

This construct was assessed through the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet, Dahlem, Zimet & Farley, 1988). This 12-item scale asks participants to rate how much each descriptive statement (e.g., “I can count on my friends when things go wrong” and “My family is willing to help me make decisions”) is applicable to them from 1 (“Strongly Disagree”) to 5 (“Strongly Agree”). Four items refer to the presence of a “special person,” and were excluded to avoid confounding overall social support with social support garnered through a relationship with a partner. Consequently we created an 8-item mean composite score (T1: α = .87, T2: α = .87) where higher scores reflected greater social support.

Disclosure of Sexual Identity

Considering the potential confounding between sexual identity disclosure and social support, we also included an index of Sexual Identity Disclosure (D'Augelli, 2002), as measured by two items indicating participants' self-report of “how ‘out’” they were to family (e.g., parents, grandparents, brothers and sisters) and to friends, respectively, using a 5-point scale: 1 (“Out to no one in your family/none of your friends”) to 5 (“Out to everyone in your family/all of your friends”).

Data Analytic Strategy

First, we verified whether our study variables were normally distributed. We found the scores of the internalized homophobia scale were skewed; thus, we applied a log-10 transformation to alleviate the violation to the normality assumption. We conducted attrition analyses across the primary study measures by comparing participants with Time 1 and 2 with those who only participated in Time 1. We then compared the study variables by sex using t-tests for continuous variables and χ² tests for categorical variables. Rather than pooling the sample in subsequent analyses, we stratified the sample by sex to acknowledge differences in the gendered construction of sexuality and romantic pursuits (Diamond & Savin-Williams, 2003; Tolman, Striepe & Harmon, 2003).

We then conducted a multiple regression analysis to test the association between psychological well-being over time and the four differing SSR trajectories: no SSR reported in the past year at either assessment (NoSSR); SSR reported only at the Time 1 or Time 2 interview (SSR1 or SSR2, respectively); or SSR reported at both assessments (SSR-SSR). These four relationship trajectories provided us with an opportunity to examine whether psychological well-being varied due to the presence and timing of a SSR relationship; we used the NoSSR category as the reference group. To avoid the risk of conceptual confounding in our analysis of SSR trajectories, we accounted for various indicators that could be associated with SSR trajectories, including youths' baseline psychological well-being score, age, Kinsey scale score, and disclosure of their sexual identity to parents and peers. We also included social support from youths' social networks as a covariate. This approach further strengthened our ability to determine whether participation in a SSR and/or OSR relationship was associated with psychological well-being, beyond the availability of support in participants' lives.

We next included whether participants had reported participating in one or more OSR at Time 1 and Time 2 given our interest to examine the association of OSRs and psychological well-being, net of youths' SSR trajectories. Finally, we tested for the interaction between OSR at Time 1 and 2 and psychological well-being in our analyses, yet found no support for an interaction effect. For parsimony, these interaction effects are not presented in the text. We present standardized beta coefficients in the text and unstandardized coefficients in the tables.

Results

Attrition Analyses

We found no mean differences between participants included in the analysis (i.e., had Time 1 and 2 data) and those excluded (i.e., only Time 1 data) by sex (χ² (1, N = 514) = 2.02, ns), age (t (512) = .51, ns), educational attainment (t (503) = -1.02, ns), residence (χ² (2, N = 514) = .65, ns), or race/ethnicity (χ² (3, N = 514) = .99, ns). We also found no differences in their Kinsey scale score (t (510) = -.34, ns), disclosure of their sexual identity to family (t (290.34) = .09, ns) or friends (t (509) = .39, ns), and social support (t (510) = -1.08, ns). We observed no differences in participation in SSRs (χ² (1, N = 514) = .46, ns) or OSRs (χ² (1, N = 514) = .69, ns) at Time 1, nor mean score differences on self-esteem (t (512) = -1.52, ns), internalized homophobia (t (511) = -.15, ns), anxiety (t (512) = .77, ns) or depressive symptoms (t (512) = .89, ns).

Sample Description

Slightly over half the sample was male (54%). Youth reported being between 15 and 19 years-old (M = 17 years, SD = 1.27). As noted in Table 1, most participants lived in a city or suburb. We found no sex differences by age, educational attainment, or residence. As for race/ethnicity, the sample composition was as follows: 42% Latino/Hispanic, 21% Black/African-American; 24% White; and 13% other race/ethnicity. There were more Latino/Hispanic males (52%) than females (31%) represented in the sample (χ² (3, N = 350) = 22.49, p ≤ .001).

Table 1.

Descriptive Statistics of Baseline Demographic Variables and Psychological Well-being Indicators by Sex

	Males M(SD)/N(%)	Females M(SD)/N(%)	Total M(SD)/N(%)	t/χ2
Age	16.97(1.28)	17.05(1.26)	17.01(1.27)	-0.60
Residence				2.05
City	139(73.2%)	107(66.9%)	246(70.3%)
Suburb	42(22.1%)	41(25.6%)	83(23.7%)
Small Town/Rural	9(4.7%)	12(7.5%)	21(6.0%)
Educational Attainment	11.17(1.49)	11.43(1.50)	11.29(1.50)	-1.63
Race				22.49***
Hispanic/Latino	99(52.1%)	49(30.6%)	148(42.3%)
African American/Black	38(20.5%)	35(21.9%)	73(20.9%)
White/Caucasian	39(20.5%)	43(26.9%)	82(23.4%)
Other	14(7.4%)	33(20.6%)	47(13.4%)
Kinsey Scale	4.48(1.35)	4.00(1.40)	4.26(1.39)	3.28***
Same-Sex Attraction				0.09
Slightly	16 (8.4%)	13(8.1%)	29(8.3%)
Moderately	31 (16.3%)	28(17.5%)	59(16.9%)
Very	55 (28.9%)	46(28.8%)	101(28.9%)
Extremely	88 (46.3%)	73(45.6%)	161(46%)
Out to Family	3.17(1.48)	2.85(1.45)	3.02(1.47)	2.03*
Out to Friends	3.96(1.25)	4.30(1.04)	4.11(1.17)	-2.76**
Social Support
Time 1	3.63(.88)	3.39(.85)	3.52(.88)	2.55**
Time 2	3.75(.85)	3.55(.84)	3.66(.85)	2.21*
Self-Esteem
Time 1	3.27(.45)	3.10(.53)	3.19(.49)	3.15**
Time 2	3.28(.45)	3.19(.52)	3.24(.48)	1.72
Homophobia1
Time 1	0.53(.43)	0.40(.38)	0.47(.42)	3.11**
Time 2	0.43(.46)	0.29(.33)	0.36(.41)	3.24***
Depression
Time 1	9.63(8.63)	13.48(10.34)	11.39(9.63)	-3.79**
Time 2	9.01(8.27)	11.30(9.04)	10.06(8.69)	-2.45*
Anxiety
Time 1	0.70(.68)	0.86(.81)	0.77(.75)	-1.93
Time 2	0.59(.69)	0.76(.77)	0.67(.73)	-2.13*

^*p ≤ .05,

^**p ≤ .01,

^***p ≤ .001

¹Raw means are presented across groups, yet the t-test statistic is computed with the log-10 transformation.

Overall, the sample self-identified as “mostly gay or lesbian” in their Kinsey scale score (M = 4.26, SD = 1.39). As shown in Table 1, males were more likely than females to describe their sexual orientation as gay rather than bisexual (t (348) = 3.28, p ≤ .001). Across both genders, over 75% categorized their level of same-sex attraction as “Very” or “Extremely” attracted, (χ² (3, N = 350) = 0.09, ns). Males had a higher degree of disclosure of their sexual identity to family than females (M = 3.17, SD = 1.48 for males; M = 2.85, SD = 1.45 for females; t (348) = 2.03, p ≤ .05); females had higher degrees of disclosure of their sexual identity to friends than males (M = 3.96, SD = 1.25 for males; M = 4.30, SD = 1.04 for females; t (348) = -2.76, p ≤ .01). As shown in Table 1, males also reported more social support than females.

Males scored higher than females in their Time 1 self-esteem score, yet no mean differences were noted at Time 2 (see Table 1). Compared to females, however, males reported higher scores on the internalized homophobia scale. Females reported more depressive symptoms than males across both waves. We found no sex differences in youths' anxiety scores at Time 1; however, females reported more anxiety symptoms than males at Time 2.

Characteristics of Youths' Same-Sex Relationships

Over a third of youth reported currently being in a same-sex relationship at both Time 1 and Time 2; no sex differences were found in the proportion of youth reporting currently being in a SSR.

When we examined the characteristics of youths' current same-sex relationships, we found approximately 55% of the sample reported older partners, 30% reported younger partners, and 15% reported same-aged partners. Most youth reported moderate to high satisfaction with their SSR. We found no sex differences in partners' age or youths' satisfaction with their SSR (see Table 2). The median duration of youths' current relationship ranged between 1 and 6 months at Times 1 and 2. We did not find sex differences in relationship duration at Time 1; however, females were more likely to report being having been in a relationship longer than males at Time 2 (Males: Less than one month = 19.7%, 1-6 months = 47.9%, 7 months or more = 32.4%; Females: Less than one month = 4.3%, 1-6 months = 29.8%, 7 months or more = 63.8%; χ² (2, N = 117) = 13.54, p ≤ .001).

Given the variability regarding being in a SSR at Times 1 and 2, we ascertained the number of serious SSRs defined as lasting more than 3 months in the past year. Females reported a greater number of serious SSRs than males at Time 1 (χ² (2, N = 346) = 6.09, p ≤ .05). This sex difference in number of SSRs disappeared by Time 2 (see Table 2). We then used these data to group youth into four SSR trajectories: NoSSRs (38.4% of males, 38.1% of females); SSR1 (13.7% of males, 25.0% of females); SSR2 (21.1% of males, 12.5% of females); and SSR-SSR (24.2% of males, 24.4% of females). While we found no sex differences in the proportion of youth in the NoSSR or SSR-SSR trajectories, females were more likely to be represented in the SSR1 trajectory while males were more likely to be represented in the SSR2 trajectory (χ² (3, N = 350) = 11.75, p ≤ .01). Given the number of sex differences, we stratified the sample by sex (N = 190 males; N = 160 females) in subsequent regression analyses (see Table 3).

Table 3.

Multivariate linear regressions of relationship trajectories on psychological well-being over time for male (N = 190) and female (N = 160) sexual minority youth

	Self-Esteem b(SE)	Homophobia1 b(SE)	Depression b(SE)	Anxiety b(SE)
MALES
Intercept	1.48(.48)**	.25(.45)	13.21(7.89)	1.29(0.73)
Score at T1	.47(.07)**	.55(.07)**	.43(.06)**	.34(.07)**
Age	-.02(.02)	.003(.02)	.20(.41)	-.01(.04)
Kinsey Scale	.004(.02)	.006(.03)	-.83(.42)*	-.06(.04)
Out to Family	-.01(.02)	-.003(.02)	.16(.42)	-.03(.04)
Out to Friends	.06(.03)*	-.01(.03)	-.01(.49)	.04(.04)
Social Support (T1)	-.02(.04)	.02(.04)	1.02(.72)	.08(.07)
Social Support (T2)	.09(.04)*	-.05(.04)	-3.06(.73)**	-.22(.07)**
SSR12	.09(.09)	.10(.09)	-2.09(1.62)	-.23(.15)
SSR22	.06(.08)	-.04(.08)	-.10(1.35)	-.08(.12)
SSR-SSR2	.15(.07)*	-.11(.08)	-1.25(1.33)	.06(.12)
In OSR (T1) 3	.14(.12)	.28(.13)*	-1.39(2.17)	-.09(.20)
In OSR (T2) 3	.16(.16)	-.001(.16)	2.43(2.81)	.35(.25)
R-Squared	28.1%	25.3%	32.0%	18.6%
FEMALES
Intercept	1.71(.51)**	.49(.25)*	-6.10(10.27)	-.10(.90)
Score at T1	.61(.06)**	.35(.07)**	.29(.07)**	.33(.07)**
Age	-.05(.03) †	.001(.01)	1.31(.54)*	.04(.05)
Kinsey Scale	.03(.03)	.00(.02)	-.02 (.62)	-.02(.06)
Out to Family	-.01(.03)	.01(.01)	.40(.56)	.01(.05)
Out to Friends	-.02(.03)	-.04(.02)**	-.53(.70)	.04(.06)
Social Support (T1)	-.04(.04)	.03(.02)	.95(.86)	.15(.08) †
Social Support (T2)	.12(.04)**	-.07(.02)**	-2.88(.86)**	-.20(.08)**
SSR12	.10(.09)	-.08(.05)†	-2.51(1.86)	-0.18(.16)
SSR22	.05(.11)	-.15(.05)**	-.65(2.20)	.03(.19)
SSR-SSR2	.06(.09)	-.07(.05)	-.53(1.87)	-.11(.17)
In OSR (T1) 3	-.07 (.10)	.02(.05)	1.88(2.03)	-.11(.18)
In OSR (T2) 3	.08(.08)	-.01(.04)	-.29(1.76)	.13(.16)
R-Squared	41.5%	24.7%	18.3%	11.2%

^†p ≤ .10,

^*p ≤ .05,

^**p ≤ .01

¹Internalized homophobia was computed using a log-10 transformation to alleviate skewness.

²NoSSR trajectory serves as reference group.

³‘No’ category servers as reference group.

Note. ‘SSR1’ includes youth who reported one or more serious (more than 3 months) SSR in the past year at baseline and no SSR in the past year at follow-up; ‘SSR2’ includes youth who reported no serious SSR in the past year at baseline and one or more SSRs in the past year at follow-up; ‘SSR-SSR’ includes youth who reported having one or more SSRs in the past year at both baseline and follow-up. ‘OSR’ indicates youth's participation in an opposite-sex relationship at time of interview.

Characteristics of Youths' Opposite-Sex Relationships

Twenty-three participants reported being in an opposite-sex relationship at both Time 1 and Time 2, with females being more likely to report currently being in a OSR (see Table 2). Approximately 14% of the sample reported currently being in an OSR. At both Time 1 and Time 2, females were more likely than males to report being in an OSR [Time 1: 6% of males vs. 20% of females; χ² (1, N = 350) = 14.80, p ≤ .001); Time 2: 4% of males vs. 26% of females; (χ² (1, N = 350) = 35.33, p ≤ .001)].

When we examined the characteristics of youths' OSRs, we found approximately that 70% of the sample reported older partners, 20% reported younger partners, and 10% reported same-aged partners. Although we found no sex differences at baseline, we did find that females were more likely than males to report older partners at Time 2 (χ² (2, N = 346) = 9.35, p ≤ .01). The median duration of youths' current OSRs ranged between 1 and 6 months at Times 1 and 2, with most reporting moderate to high satisfaction with their OSR. We found no sex differences in relationship duration or youths' satisfaction with their OSR (see Table 2).

Given the variability regarding being in an OSR at Times 1 and 2, we ascertained the number of OSRs defined as lasting more than 3 months in the past year. Females reported a greater number of serious OSRs than males at Time 1 (χ² (2, N = 346) = 25.71, p ≤ .01) and Time 2 (χ² (2, N = 346) = 18.35, p ≤ .01). We then used these data to group youth into four OSR trajectories: NoOSR (91.1% of males, 67.5% of females); OSR1 (5.3% of males, 6.9% of females); OSR2 (2.6% of males, 12.5% of females); and OSR-OSR (1.1% of males, 13.1% of females). Males were more likely than females to be represented in the NoOSR trajectory, whereas females were more likely to be represented in the OSR2 and OSR-OSR trajectories (χ² (3, N = 350) = 37.48, p ≤ .01).

Relationship Trajectories and Psychological Well-being

Males

Self-esteem

Males who reported a SSR at both Times (SSR-SSR) reported greater self-esteem at Time 2 (β = 0.14, p ≤ .05) than males who did not report a SSR over time (NoSSR). We found no difference in males' self-esteem score at Time 2 in any other SSR trajectory (see Table 3). Males' self-esteem score at Time 2 was also positively associated with their baseline self-esteem score (β = 0.47, p ≤ .001), the number of friends who knew their sexual identity (β = 0.15, p ≤ .05), and the amount of social support reported at Time 2 (β = 0.17, p ≤ .05). Self-esteem at Time 2 was not associated with males' age, Kinsey scale score, number of family members who knew youths' sexual orientation, or having an OSR at baseline or follow-up.

Internalized Homophobia

We found no differences in males' internalized homophobia score at Time 2 based on their SSR trajectory; however, youth who reported an OSR at Time 1 reported higher internalized homophobia at Time 2 score (β = 0.15, p ≤ .05). Internalized homophobia at Time 2 was positively associated with males' baseline score (β = 0.53, p ≤ .001). We found no association between internalized homophobia and males' age, Kinsey scale score, number of friends or family who knew about their sexual identity, or social support at baseline or at follow-up.

Depressive Symptoms

We found no differences in males' depressive symptoms score at Time 2 based on their participation in SSRs or OSRs over time. Males who reported greater depressive symptoms at Time 1 were more likely to report greater number of symptoms at follow-up (β = 0.45, p ≤ .001). Males' depression score at Time 2 decreased; however, depressive symptoms scores were negatively associated with Kinsey scale scores (β = -0.14, p ≤ .05) and greater social support at follow-up (β = -0.31, p ≤ .001). No additional covariates were associated with depressive symptoms.

Anxiety Symptoms

We found no differences in males' anxiety score at Time 2 based on their SSR or OSR trajectories. Males who reported greater anxiety at Time 1 were more likely to report higher anxiety at follow-up (β = 0.34, p ≤ .001); however, anxiety decreased as participants reported greater social support at Time 2 (β = -0.27, p ≤ .01). We found no other association between males' anxiety score at follow-up and our study variables.

Females

Self-esteem

We found no differences in females' self-esteem scores at Time 2 based on their SSR or OSR trajectories. Females who reported higher levels of self-esteem at Time 1 were more likely to report higher levels of self-esteem at Time 2 (β = 0.63, p ≤ .001). Females' self-esteem was associated positively with social support at Time 2 (β = 0.20, p ≤ .01), and negatively with age (β = -0.11, p ≤ .10). No additional covariates were associated with self-esteem.

Internalized Homophobia

Females in the SSR2 trajectory reported less internalized homophobia at Time 2 (β = -0.21, p ≤ .01) than females in the NoSSR trajectory. We also found that females in the SSR1 trajectory were less likely to report internalized homophobia at Time 2 (β = -0.16, p = .06); however, this difference only approximated statistical significance. We found no differences in internalized homophobia for the SSR-SSR trajectory. Internalized homophobia at Time 2 was positively associated with females' baseline score (β = 0.38, p ≤ .001), and negatively associated with the amount of social support reported at Time 2 (β = -0.26, p ≤ .001) and participants' disclosure of their sexuality to friends (β = -0.20, p ≤ .01). We found no associations between internalized homophobia and females' age, Kinsey scale score, disclosure to family, or having an OSR at baseline or follow-up.

Depressive Symptoms

We found no differences in females' depressive symptoms score at Time 2 based on their SSR or OSR trajectory. Females who had greater depressive symptoms at Time 1 were more likely to report a higher number of symptoms at follow-up (β = 0.33, p ≤ .001). Females' depressive symptoms score at Time 2 were positively associated with age (β = 0.18, p ≤ .05), and negatively associated with social support at follow-up (β = -0.27, p ≤ .001). No additional covariates were associated with depressive symptoms.

Anxiety Symptoms

We found no differences in females' anxiety scores at Time 2 based on their SSR or OSR trajectory. Females who had greater anxiety scores at Time 1 were more likely to have higher anxiety scores at Time 2 (β = 0.35, p ≤ .001). Greater social support at Time 2, however, was negatively associated with symptoms of anxiety (β = -0.22, p ≤ .01). We found no other association with females' anxiety scores at follow-up.

Discussion

Youth explore and integrate aspects of their sexuality into their personal identity through involvement in sexual and romantic relationships (Furman & Shaffer, 2003). Dating and involvement in same-sex and opposite-sex relationships have been previously associated with both changes in psychological distress (i.e., anxiety and depression), but no conclusive information on the associations have emerged. Dating and involvement in same-sex and opposite-sex relationships may also be associated with sexual minority youths' self-esteem and internalized homophobia. Although involvement in SSRs have been posited to enhance self acceptance and reduce internalized homophobia among gay men (Isay, 2009), the role of sexual minority youths' participation in both same-sex and opposite-sex romantic relationships has received limited attention in the scientific literature (Diamond, Savin-Williams, & Dubé, 1999; Russell & Consolacion, 2003). Consequently, we sought to describe the prevalence and characteristics of SSRs and OSRs in a sample of self-identified sexual minority youth, and to examine whether their participation in these relationships was associated with four psychological outcomes (i.e., self-esteem, internalized homophobia, and depressive and anxiety symptoms).

Participation in Same-Sex Relationships

Consistent with previous findings of sexual minority youth (Russell & Consolacion, 2003), SSRs were common among this cohort. One-third of the sample reported currently being in a SSR when interviewed at baseline or follow-up, and approximately half of the sample reported at least one serious SSRs (i.e., lasting over 3 months in duration) in the past year. Furthermore, over one-half of all youth reporting a SSR at the time of their interviews also reported having an older partner, and another quarter reported having a younger partner. While research with heterosexual-identified adolescent samples has found that males are more likely to date younger partners and females are more likely to date older partners (Bauermeister et al., 2009; Ford & Lepkowski, 2004), we found no gender disparity in partners' age in this sample. Given that most youth in SSRs reported high satisfaction with these relationships, there was insufficient variability to test whether relationship quality was associated with psychological well-being, beyond youths' participation in SSRs in our multivariate analyses. Given the salience of romantic relationships during this developmental period, future research is necessary to understand in greater depth how sexual minority youths' relationship characteristics (i.e., partners' age, relationship satisfaction) may be associated with psychological well-being in samples having greater variability in relationship characteristics.

Participation in romantic relationships is closely linked to youths' perceptions of themselves (Furman & Shaffer, 2003), and may assist in identity exploration (Glover, Galliher, & Lamere, 2009). In our study, participation in SSRs was associated with self-esteem and internalized homophobia; however, the timing and sequence of SSRs had different effects on the self-esteem of males and females. Among males, youth participating in the SSR-SSR trajectory reported increased levels of self-esteem over time (i.e., adjusting for Time 1) than those in the NoSSR trajectory. Interestingly, males who reported episodic participation in SSRs (e.g., SSR1 or SSR2) did not report change in self-esteem over time. This finding suggests that prolonged participation in SSRs (at least as much as the one-year period covered by our assessment) may be necessary in order to increase self-esteem among males. Among females, however, we found that intermittent participation in SSRs was associated with decreased internalized homophobia over time. Females in the SSR1 and SSR2 trajectories reported less internalized homophobia than those in the SSR-SSR or NoSSR trajectories. One possible explanation is that females in these episodic SSR trajectories may experience less internalized homophobia because they adhere to scripted social norms regarding female sexuality (i.e., maintain their sexual fluidity; Peplau & Fingerhut, 2007). Compared to episodic SSRs, female youth in the NoSSR trajectory may fear deviating from compulsive heterosexuality norms and miss out on the protective effects associated with exploring and becoming comfortable with their sexuality. Furthermore, it is not immediately evident to us why females in the SSR-SSR trajectory would not report lower homophobia scores than those in the NoSSR trajectory. One plausible explanation for this finding is that females in the SSR-SSR trajectory may be more likely to deviate from the compulsory heterosexuality norm and to report comparable feelings of internalized homophobia if they are exposed to sexual prejudice. While plausible, we were unable to test these statements; additional qualitative and quantitative research is needed to describe, test and elucidate these mechanisms. Specifically, future research should examine whether internalized homophobia and the pursuit of SSRs are linked to salient compulsive heterosexuality norms in female youths' lives.

Participation in Opposite-Sex Relationships

Consistent with research indicating the presence of SSRs and OSRs among sexual minority youth (Diamond & Savin-Williams, in press), approximately one-quarter of sexual minority youth in our sample reported having serious OSRs in the last year, with females being more likely to report currently having an opposite-sex partner than males. Consistent with the heterosexual literature (Bauermeister et al., 2009; Ford & Lepkowski, 2004), females were more likely than males to report older opposite-sex partners.

In contrast to the existing literature on dating and psychological well-being among heterosexual youth, we found limited evidence to suggest that being involved in OSRs was associated, either positively or negatively, with psychological well-being – with one notable exception. Males who reported participating in an OSR at the first interview were more likely to report greater internalized homophobia at follow-up. We found no other association between participation in OSRs and psychological well-being among males or females. These findings underscore how similar activities (e.g., dating in adolescence) may have different influences on heterosexual and sexual minority youth. Specifically, our findings reflect the inequality in permissible sexual and gender identity expressions among male and female youth (Russell & Consolacion, 2003). Among males, the deviation from rigid social norms, including masculine gender scripts and a heterosexual identity, may produce more discomfort with their sexual identity (Russell & Consolacion, 2003; Tolman, Striepe, & Harmon, 2003; Tolman, 2006). As a result, their participation in OSRs may reflect their negotiation of these gender scripts and their exploration of their sexual identity and desires. Among females, the deviation from heteronormativity may be perceived as less of a social grievance given the eroticization of same-sex behaviors among women; however, they still take their toll on psychological well-being by disrupting females' ability to integrate a stable sexual identity (Meyer, 2003). We found some support for these perspectives in our sample. Males were more likely to self-identify as gay than bisexual, and to be known as gay or bisexual by more family members than females; however, they were also more likely to report more internalized homophobia over time than females. Sexual minority females, on the other hand, were more likely than males to report higher symptoms of depression and anxiety over time. Females also reported lower self-esteem at baseline than males; yet, consistent with research examining change in self-esteem across adolescence (Galambos, Barker & Krahn, 2006), this difference disappeared at Time 2. Taken together, these findings tentatively support Tolman et al.'s (2003) argument for gender-specific effects of compulsive heterosexuality norms. Unfortunately, we were unable to measure and test whether youths' compliance with these compulsive heterosexual norms was associated with psychological well-being.

We found the lack of differences between males and females across anxiety and depressive symptoms to be noteworthy. Although Russell and Consolacion (2003) reported differences in depression and anxiety due to attraction and relationship status, we were unable to replicate their findings as the timing and sequence of SSRs or OSRs among sexual minority youth was not associated with symptoms of distress. This difference in findings may be attributable to multiple factors including not having a heterosexually-identified comparison group in our sample, accounting for concurrent participation in OSRs and SSRs, and inclusion of other covariates including social support. For example, while romantic partners are considered an important source of social support, it is plausible that we did not detect an effect because youth already had sufficient social support from individuals other than a romantic partner (i.e., social support was associated with improved psychological well-being for both males and females). This argument is consistent with past research suggesting that the presence of a “significant other” such as a dating partner may be unnecessary to decrease psychological distress if youth have access to social support from others (Goodenow, Szalacha, & Westheimer, 2006). These findings underscore the need to replicate these findings in other samples of sexual minority youth.

Limitations and Conclusions

This study has some limitations that should be mentioned. First, our findings may not be generalizable to all sexual minority youth. The youth in this sample, having been recruited at drop-in centers specifically designed for GLBT identified youth, may already possess strong coping skills and be capable of utilizing resources to seek out environments affirming their sexual identities. Nonetheless, the lack of representativeness of samples of sexual minority youth is a methodological concern in most studies with this population (Binson, Blair, Huebner, & Woods, 2007). Additionally, this study sample is comprised of sexual minority youth living in close proximity to urban areas. Consequently, these youth may be able to engage with a GLBT-friendly community more readily than youth living in other parts of the United States (e.g., rural areas), may have access to more gay-affirming social support, and may exist in an environment that better provides for the possibility of SSRs, including increased access to potential dating partners. Even so, the presence of less psychological distress suggests that the influence of SSRs on well-being may be greater in less supportive environments. Future research should replicate our findings with other samples of sexual minority youth. Furthermore, the inclusion of other SSR characteristics in future studies (e.g., relationship satisfaction and/or age of partner trajectories), as well as sexual minority youths' future orientation regarding their adult relationships (e.g., having children, settling down with a same-sex partner), may provide additional information regarding the role of SSR trajectories on psychological well-being. These limitations notwithstanding, our findings provide a unique opportunity to examine the role of SSRs and OSRs on psychological well-being. This study also has several strengths, including its prospective design, the inclusion of the participants' status at Time 1, and the use of highly reliable and valid measures of psychological well-being. Future research should include more observations across different developmental periods (e.g., middle adolescence, late adolescence, and the transition to young adulthood), as the aforementioned relationships may differ across these developmental periods.

Overall, our findings provide an important contribution to the growing literature on the mental health of GLB youth. Similar to heterosexual youth, sexual minority youth engage in relationship exploration. Our results also suggest that involvement in a SSR or OSR may be associated with greater self-esteem and lower internalized homophobia. While these effects may vary based on sex, for both male and female sexual minority youth, there is evidence to suggest that within supportive contexts, exploration of emergent sexual identities can be a beneficial experience (Peplau & Fingerhut, 2007). In order to thrive emotionally and psychologically, sexual minority youth may require an avenue for their identities to develop safely. Providers within supportive environments (e.g., drop-in centers, support groups, schools, shelters) may be well-poised to offer sexual minority youth support as they explore their sexuality. Furthermore, service providers should evaluate their on-going programs, including whether particular programs enhance sexual minority's experiences with SSRs and OSRs. These findings may help inform how specific services lead to better mental health outcomes for sexual minority youth. Future research is needed to explore how to cultivate positive support for sexual minority youth, as well as to understand better what characteristics of their relationships with others lead to beneficial outcomes.