Identity Development and Exploration Among Sexual Minority Adolescents: Examination of a Multidimensional Model

Abstract

Based on a multidimensional model of sexual identity development, the current study examined self-labeling, romantic, and sexual experiences among middle and late adolescent sexual minorities. Eighty-two youths between the ages of 14 and 21 were recruited from gay, lesbian, bisexual, and transgender (GLBT) community centers and associated Internet groups. Differences between males and females and between middle and late adolescents were first examined. Males reported markedly different patterns of self-labeling and romantic experiences, relative to females, but few age differences emerged. Additionally, patterns of association between self-labeling and sexual or romantic experiences provided support for the multidimensional model of identity development and exploration. Finally, significant associations emerged between relationship experiences and relational competency, such that adolescents who described themselves as isolated from intimate interpersonal relationships reported less relational esteem and more relational depression than those who reported predominantly engaging in same-sex romantic relationships. In addition, alternative styles of engaging in interpersonal relationships (e.g., same-sex intimate friendships, heterosexual dating) also were also associated with increased relational competency.

KEYWORDS:

• adolescent
• sexual minority
• romantic relationships
• identity development
• sexual orientation
• multidimensional model
• social constructionist

Adolescent sexual minorities represent youth who are gay, lesbian, bisexual, transgendered (GLBT), or questioning their sexual orientation (Fisher & Akman, 2002). It is currently estimated that the prevalence of youth who identify as a sexual minority is between 4% and 17% of the U.S. adolescent population (Anhalt & Morris, 1998). The dominant culture influences perceptions of sexual orientation by promoting the assumption of heterosexuality as the only normal developmental outcome, thereby creating potential challenges for sexual minority youth who are attempting to explore, accept, and integrate sexual orientation into their lives (Tharinger & Wells, 2000). For this reason, sexual minority youth may experience a different identity development process in which the knowledge that they are different is ever present (Striepe & Tolman, 2003). Although there are many contributing variables that influence sexual minorities' development, the emphasis of the current study is on two essential components of experience that collectively influence this development: identity development and identity exploration (e.g., romantic and sexual experiences).

IDENTITY DEVELOPMENT

Perhaps the most salient of the many tasks faced during adolescence is the formation of individual identity (Erikson, 1968). During this critical period, adolescents are expected to negotiate developmental tasks including forming a cohesive sense of self, achieving autonomy while maintaining belongingness, and demonstrating independence while balancing being supported (Tharinger & Wells, 2000). Often these important tasks are achieved through social comparison or identifying with the standards and norms of the main population. Those who are unable to identify with the dominant culture may experience a different developmental trajectory. Adolescent sexual minorities differ from the dominant culture in sexual orientation, and this divergence from the main culture's assumed developmental path creates a novel point of reference in identity formation. Striepe and Tolman (2003) offered further insight into the unique developmental challenges experienced in identity formation by sexual minority adolescents.

Few adolescents worry that they will have to sit down with their parents and confide what they have come to realize about their sexual identity, that is to say “Mom, Dad, I'm straight.” In our society, heterosexuality is assumed from birth. It is when adolescents show signs of being different than the heterosexual norm that sexual identity becomes a visible aspect of development. (p. 523)

Recent research on sexual identity development among GLBT youths draws from a social constructionist theoretical perspective, of which a key feature is emphasis on how individuals interrelate, how culture and society impact this process, and finally, the phenomenological reality taken from these integrated parts (Owen, 1992). Social constructionism underscores that human behavior and development cannot be understood in a social or cultural vacuum (Schaller, 2002), but must be understood through the meaning given to different outcomes via personal, social, and cultural perspectives. In regards to sexual orientation, the social constructionist perspective holds that identity is not essential or innate but rather is “constructed” from social and cultural frameworks (Broido, 2000). From this viewpoint, researchers have recognized that sexual orientation is impacted by personal, social, and cultural factors, and have now begun to generate and examine models of identity development that seek to explore how this collective experience influences identity development in adolescent sexual minorities (Cox & Gallois, 1996).

One recent model of sexual minority identity development is concerned with the interactive influences between the individual and social environment (Horowitz & Newcomb, 2001). This model offers a unique advantage to past models by specifically examining desire, behavior, and identity as three separate constructs related to sexual orientation and identity, emphasizing the salience of the individual meaning ascribed to each of these constructs.

Once the categories of desire, behavior, and identity are separated, it's theoretically possible for a person to change his or her sexual identity or behaviors while maintaining an underlying sexual orientation … in this sense social constructionist perspective empowers the individual's choice in sexual expression, while recognizing that there may not be a choice in the orientation of sexual desire. (p. 16)

This multidimensional explanation of sexual orientation recognizes the importance of society's influence while maintaining that individuals are able to choose how their identity will be defined within culture rather than being defined solely by their sexual orientation. The overall purpose of this model is to create an inclusive description of identity that is not restricted to limited and insufficient labels generated by past models of identity development.

The multidimensional model lends itself to the notion that sexual orientation and identity do not exist as fixed points (i.e., homosexual or heterosexual), but rather represent a continuum of experience. However, despite the recent emergence of these models, the conceptualization of sexual orientation as a continuous variable is not new. In his landmark research over 60 years ago, Kinsey (1941) concluded that it is erroneous to conceptualize sexual orientation as dichotomously distributed but rather it should be approached as continuum based. Although several criticisms have been raised against Kinsey's work regarding the integrity of his methods and results (Brecher & Brecher, 1986; Ericksen,1998), he initiated an interesting body of research that has not been revisited until recent years. More recent research has addressed many of the methodological issues associated with Kinsey's work and extended the literature on sexual orientation and identity development.

Perhaps the greatest area of growth in this body of research is the recognition of a need to move away from categorical labels to continuous definitions, with the resultant development of measures aimed at assessing sexual orientation differently (Balter, 1998; Holden & Holden, 1995; Rothblum, 2000). Sell (1996) conducted a large scale review of different measures of sexual orientation examining various dichotomous, bipolar, orthogonal, and multidimensional scales. The most representative scales capturing the full range of experiences were those that gave participants the flexibility to identify their sexual attractions, behaviors, attitudes, and orientation on multiple dimensions. Johns (2004) used multidimensional scales to measure identity development in a sample of 143 adult sexual minority participants. Participants in her study viewed identity formation occurring in different phases rather than linear stages. By using a continuum-based measure, participants had the opportunity to describe their experiences in a manner that provided support for a multidimensional approach rather than a stage-based process.

A multidimensional approach was also supported by Kinnish's (2003) work with 762 heterosexual and sexual minority participants (i.e., bisexual, gay, lesbian). Participants were asked to retrospectively report their sexual behavior, fantasy, romantic attractions, and identity for a five-year period beginning at the age of 16. Individuals who identified as bisexual reported the most transitions in sexual identity (e.g., changing self-label of sexual orientation to bisexual from straight, or gay to bisexual) over the five-year period, gay men and lesbian women reported the second most, and heterosexual individuals reported the fewest. It is interesting to note that lesbian women reported a greater number of transitions in sexual identity compared to gay men. These results suggest important gender differences between different sexual minority groups in the development of identity that could be better understood upon additional investigation. Also, the limitations of retrospective studies, such as uncertain accuracy of memory accounts, suggest that research using adolescent samples could help clarify this body of research. Specifically, examining changes in sexual identity, behavior, and attractions throughout adolescence or examining differences between younger and older adolescents in the self-descriptions of their sexual orientation would begin to address remaining questions about the development of sexual identity. In addition, including an examination of the differences between genders would provide an extension to the research cited above and help clarify differences between or within these two sexual minority groups (i.e., gay men and lesbian women).

Due to the fairly recent emergence of the multidimensional models, research that utilizes this perspective has been limited and restricted primarily to adult samples. To address these gaps, the current study asked sexual minority adolescents to define their sexual attraction utilizing dimensional continuums of sexual orientation, traditional categorical descriptions (e.g., gay, lesbian, bisexual, heterosexual), and open-ended formats (i.e., “What words would best describe the way you see your sexual orientation?” “What words do you use to describe your sexual orientation to others?”). In addition, associations among self-labeling and same- and opposite-sex romantic and sexual attractions or behaviors were examined.

IDENTITY EXPLORATION

Interpersonal processes and relationship experiences in adolescence are hypothesized to serve as the foundation for future romantic relationships. Adolescents' exploration and experiences in peer relationships help build interaction skills and provide the opportunity for romantic connections to develop (Furman, Simon, Shaffer, & Bouchey, 2002). The assumption of heterosexuality influences adolescents' relationship development because a majority of heterosexual peer groups discourage and disapprove of homosexual relationships (Connolly, Furman, & Konarski, 2000). Rejection of alternative sexual orientation development among adolescent peer groups suggests that adolescent sexual minorities may be forced to establish meaningful interpersonal, sexual, and romantic relationships independent of peer support. This pattern stands in sharp contrast to the process for heterosexual youth, who work to develop these relationships within the supportive context of their peer groups. Thus, many adolescent sexual minorities may be forced to either abstain from romantic relationships or engage in alternative relationships in place of romantic relationships (Diamond, Savin-Williams, & Dube, 1999).

Relationship Withdrawal

The stigma surrounding the development of romantic relationships with same-sex partners and the difficulty of being able to identify other youth with same-sex romantic interest contribute to adolescent sexual minorities being less likely to have any type of romantic relationship experience during their middle and high school years compared to their heterosexual counterparts (Diamond & Dube, 2002). The difficulty in identifying any other sexual minority peers, especially a desired romantic partner, limits the opportunity for these students to learn and practice important interpersonal relationship skills that are critical in the development of adult romantic relationships (Connolly et al., 2000). These obstacles often result in varying degrees of isolation and withdrawal, leading to sexual minority youth having smaller peer groups, hindering the closeness of friendships, and cultivating negative expectations about romantic relationships and control over their romantic lives (Diamond & Lucas, 2004). This social and emotional isolation is concerning and research has documented the associated negative outcomes (e.g., compromised self-esteem, high-risk behaviors, mental health problems) for adolescent sexual minorities who are isolated from their peers (Alexander, 2002; Diamond, 2003; Martin & Hetrick, 1988; Quinn, 2002; Savin-Williams, 1988; van Heeringen & Vincke, 2000). Due to the difficulties in creating and maintaining a same-sex romantic relationship, many adolescent sexual minorities may seek alternative relationship contexts.

Alternative Relationships

As adolescent sexual minorities seek to fulfill their emerging social need for romantic and sexual involvement while navigating through the constraints of dominant heterosexual cultural assumptions, three main types of alternative relationships emerge. These types include: (1) commitment to a same-sex intimate friendship, (2) engagement in exclusively sexual relationship with a same-sex partner, and (4) participating in heterosexual dating.

Same-sex alternatives

The many impediments faced by sexual minority adolescents in forming romantic relationships often force these youth to adapt same-sex relationships that may provide limited benefits typically obtained in the context of a romantic partnership.

The difficulty inherent in simply identifying other sexual-minority youths creates onerous risk; in response to risk sexual minorities may strike a tenuous balance between risk and reward pursuing exclusively emotional or exclusively sexual relationships that allow them a measure of same-sex intimacy without placing them in jeopardy. (Diamond et al., 1999, p. 177)

Sexual minority adolescents may seek to develop intimate same-sex friendships in an attempt to cultivate a relationship that satisfies emotional needs that are usually fulfilled in a romantic relationship. The dynamics of these “passionate friendships” are marked by intense emotional investment, providing intimacy and support that is not dependent on sexual consummation (Diamond et al., 1999). These types of friendships provide closeness, intimacy, and often the exclusiveness of romantic relationships, but are devoid of other aspects such as sexual intimacy. Typically this pathway of emotional fulfillment through intimate friendships is more likely to occur among young women (Savin-Williams & Diamond, 2000). In contrast, young men appear more likely to pursue a pathway that focuses on exclusively sexual relationships (Diamond, 2003).

Due to social norms and gender expectations, young men are generally not afforded the opportunity to develop deeply intimate same-sex friendships characterized by mutual self-disclosure, affection, and tenderness; therefore, relationships that are based on exclusive sexual intimacy may be the only avenue perceived by adolescent boys to have a same-sex relationship (Diamond et al., 1999). Such relationships may be important in serving as a means to validate same-sex attractions and confirm aspects of the adolescent's sexual orientation; however, such relationships are limiting due to the lack of emotional intimacy that is characteristically tied to more intense physical intimacy in romantic relationships.

Heterosexual alternatives

There are numerous reasons that contribute to adolescent sexual minorities' decisions to engage in heterosexual dating and relationships. Perhaps the most salient factor is the normative pressure toward opposite-sex dating during adolescence. Such pressure may provide a myriad of motivational aspects for adolescent sexual minorities to participate in heterosexual dating (e.g., exploration, concealment of orientation, romantic fulfillment). It is recognized that the majority of sexual minority youth date heterosexually throughout adolescence (Diamond et al., 1999). However, the limited amount of research focused on this topic has yet to address several possibilities of both negative and positive outcomes of heterosexual dating among adolescent sexual minorities. Thus, despite the importance of the dating experience during adolescence, only limited emerging research has sought to address the diverse roles that sexual minorities assume in the context of dating relationships. The current study examined the function and importance of these various relationship types and how they are associated with psychosocial functioning and relationship competency in middle and late adolescents.

In summary, two sets of hypotheses related to sexual identity development and romantic or sexual exploration were presented in this study.

1. Identity Development. We hypothesized age and gender differences in adolescents' use of both traditional labels and scores on continuous measures of heterosexual attraction and homosexual attraction. Given evidence for transition across adolescence in self-labeling and for greater fluidity of attraction among women, we predicted that there would be more variability in self-identification among younger adolescents and young women, relative to older adolescents and males. In addition, the multidimensional model of identity development predicts that, given the opportunity to describe their sexual orientation in their own words, adolescents would go beyond the use of traditional labels (e.g., gay, lesbian, heterosexual) and provide more individualized, meaningful descriptions of their sexual orientation. Finally, the multidimensional model would also predict that patterns of association between self-labeling, attraction, and behavior would be somewhat fluid, with adolescents reporting a range of attractions and behaviors, regardless of their stated self-labels.

2. Identity Exploration. We hypothesized age and gender differences in the types of romantic and sexual relationships/experiences reported by adolescents. Differences were examined in terms of dating and relationship experiences (having a crush, dating, going steady) and sexual behaviors (affectionate behaviors such as hand holding, petting behaviors, or sex) with both same- and opposite-sex partners. As with hypotheses regarding self-identification, we predicted that women and younger adolescents would report greater variability in same- and opposite-sex attractions and behaviors. We also examined differences between individuals with different predominant dating styles (i.e., same-sex romantic relationships, same-sex passionate friendships, same-sex exclusively sexual relationships, heterosexual relationships, or no participation in relationships) in terms of self-esteem and relationship competence. Given the salience of romantic experiences in adolescence, those who reported no participation in romantic or intimate relationships were expected to report the lowest levels of relational competence and self-esteem. However, differences between those who reported participation predominantly in same-sex romantic relationships and those who engaged in alternative relationship styles remained exploratory.

METHODS

Participants

Participants were 82 adolescents recruited using three strategies. First, participants were recruited from events and activities sponsored by local sexual minority community groups (approximately 76%). Second, Internet groups were contacted using listservs provided by the community centers (approximately 10%). Finally, participants who completed the study were given referral cards to distribute to other interested individuals. The remaining 14% of participants were recruited from referrals from previous participants and project staff.

Participants were between 14 and 21 years of age and were divided into two groups: Middle adolescents were defined as those who were still in middle or high school and late adolescents were participants who were no longer in high school but less than 22 years old. Table 1 provides a summary of sample characteristics including age, gender, and sexual orientation label participants used for themselves as well as what label they told others. The racial background of participants was self-identified as 87% White, 2% Asian, 8% Latino/Hispanic, 1% Native American, and 2% identified race as other. The religious affiliation was 48% Mormon (Latter Day Saints), 4% Catholic, 1% Protestant, 1% Jewish, 22% other, and 24% identified no religious affiliation. Fifty-two percent of participants' parents were married to each other, 38% had divorced or separated parents, 4% of the parents had never married, 4% were widowed, and the remaining 2% were unspecified. Parents' educational status suggested a wide range of socioeconomic backgrounds; 37.8% of fathers and 30.5% of mothers had a high school education or less; 20.7 % of fathers and 29.2% of mothers had completed technical school or some college; 39.1% of fathers and 39.0% of mothers had a college degree or higher. Data were not reported for 2 fathers and 1 mother.

TABLE 1 Percentages and Means of Male and Female and Middle and Late Adolescents Self-Labeling Using Dichotomous and Continuous Variables (N = 82)

	Male	Female	Middle	Late	Total
Categorical variables N, (%)
Gay/lesbian	30 (83.3)	18 (39.1)	18 (51.4)	30 (63.8)	48 (58)
Straight	3 (8.3)	11 (23.9)	5 (14.3)	9 (19.1)	14 (17)
Bisexual	2 (5.6)	13 (28.3)	9 (25.7)	6 (12.8)	15 (18)
Questioning/other	1 (2.8)	4 (8.7)	3 (8.6)	2 (4.3)	5 ( 7)
Continuous variables M (SD)
Homosexual attraction	8.4 (2.6)	6.1 (3.4)	6.8 (3.5)	7.4 (3.1)	7.1 (3.3)
Heterosexual attraction	2.8 (2.4)	5.5 (3.3)	4.6 (3.4)	4.0 (3.2)	4.3 (3.2)
Age	18.7 (1.4)	17.8 (1.7)	16.8 (1.1)	19.2 (1.2)
Note. Middle = adolescents not graduate from high school or dropped out, Late = adolescents graduated from high school.
Continuous attraction variables were measured on two 10-point likert scales (1 = not at all homosexual to 10 = highly homosexual; 1 = not at all heterosexual to 10 = highly heterosexual).

Procedures

Participants under the age of 18 were required to have written parental consent in addition to providing written assent, while those who are were 18 or older provided only their own signature. The consent form was phrased generically to indicate that the study was designed to learn more about the development of dating and relationships among adolescents, so that adolescents who had not disclosed to their parents would not be precluded from participation. Participants either collected survey packets at community events or received packets in the mail after providing their addresses to the first author or research assistants. Questionnaire measures took 20–30 minutes to complete and were returned by participants in a self-addressed, stamped envelope. Participants were compensated for participation with $10; the incentive was sent to participants by mail as soon as completed questionnaires were received.

Questionnaire Measures

Demographic information

The demographic section assessed race, age, gender, educational attainment, educational goals, religiosity, and educational attainment of parents.

Sexual development/identity survey

For the current study, a 25-itemed measure was created to obtain information on sexual identity development, identity disclosure, and identity exploration. No established measures were found that captured the range of behaviors, attractions, and aspects of identity that were relevant to the current study. The created measure was based on Horowitz and Newcomb's (2001) multidimensional model and provided a series of questions pertaining to attractions, relationships, behaviors, and disclosures using categorical responses, as well as continuous and open-ended questions. Participants were first asked to report if they had engaged in a range of dating and sexual behaviors (had a crush on, went steady with, said “I love you”, affectionate behaviors, petting, sexual intercourse) with a male or female partner. Response options were never, at least once in my life, at least once in the past year, and currently. Other items asked respondents to separately describe their heterosexual and same-sex attraction on continuous scales (1 = not at all, 10 = very much), to report which traditional labels (gay/lesbian, straight, bisexual, questioning) they used for themselves and which they told others, and to describe their sexual orientation in their own words. An additional item asked adolescents to select their predominant dating style from a list of descriptors (same-sex romantic relationships, same-sex exclusively sexual relationships, same-sex passionate friendships, heterosexual dating, and relationship withdrawal). Finally, participants were also asked to describe their disclosure history as part of a larger study.

Rosenberg self-esteem scale

The 10-item Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965, 1989) assesses global self-esteem. Items are answered on a 4-point Likert-type scale (1 = strongly agree, 4 = strongly disagree) and are averaged to create a global self-esteem score. Rosenberg demonstrated concurrent validity comparing its relationship to depressive affect, psychosomatic symptoms, nurses' ratings, peer ratings, and a number of other constructs. Additionally, Hagborg (1993) compared the RSES to nine separate self-esteem domains to determine the unidimensional nature of the RSES. Hagborg found that the RSES was highly correlated with other measures of self-esteem. Cronbach's alpha was .91 for this study.

Relational assessment questionnaire

The Relational Assessment Questionnaire (RAQ; Snell & Finney, 1993) is a 26-item self-report questionnaire that assesses different areas of romantic relationship functioning. Items are answered on a 5-point Likert type scale (1 = not at all characteristic of me to 5 = very characteristic of me). Three different components of relationship functioning are assessed. Relational-esteem is the tendency to positively evaluate one's capacity to relate intimately to another person (7 items; e.g., “I am a good partner for an intimate relationship”). Relational-depression is the tendency to feel depressed about the status of one's intimate relationships (10 items; e.g., “I am disappointed about the quality of my close relationship”). Finally, relational preoccupation is the tendency to be highly obsessed with thoughts about intimate relationships (9 items; e.g., “I think about intimate relationships all the time”). Assessment of convergent and discriminant validity of the RAQ found that the three relational indexes were related in predictable ways to relationship involvement and attraction. Snell and Finney observed Cronbach's alphas for relational esteem, depression, and preoccupation of .81, .88, and .85, respectively, with eight-week test–retest reliabilities of .71, .73, and .70. Cronbach's alpha for the three subscales was .80, .89, and .89, respectively, for participants in this study.

RESULTS

Identity Development

Table 1 summarizes the number and percentages of adolescents' endorsement of traditional labels of sexual orientation. Additionally, means and standard deviations are presented for continuous ratings of same- and opposite-sex attraction by gender and age. Analyses were conducted to assess for differences between males and females or between middle and late adolescents in self-identification using either traditional labels or scores on continuous measures of heterosexual attraction and homosexual attraction. None of the methods of self-identification yielded significant differences between middle and late adolescents; thus, only analyses examining differences between the genders are presented.

A chi-square analysis was conducted to evaluate differences between males and females in self-labeling using traditional labels of sexual orientation (i.e., straight, gay/lesbian, bisexual). Gender and traditional labels were found to be significantly related, χ² (2, N = 77) = 15.12, p < .01, V = .443, with males more likely to identify as gay and females more likely to identify as straight or bisexual. Independent-sample t tests were conducted to evaluate differences on continuous ratings of homosexual and heterosexual attraction. The test comparing males and females was significant for both homosexual attraction, t(75) = 3.10, p = .001, d = .698, and heterosexual attraction, t(75) = -3.90, p <.001, d = -.866. Males reported more polarized attraction patterns with high scores for homosexual attraction and low scores for heterosexual attraction, while means for females were roughly in the center of each scale.

Table 2 presents percentages for same- and opposite-sex attractions, behaviors, and romantic experiences among adolescents who identified as gay/lesbian, straight, or bisexual. Chi-square analyses examining associations among self-labeling and behaviors were not possible because the assumption of an expected value of at least five in each cell was violated for many cells. As might be predicted, patterns suggested that the majority of those who identified as gay/lesbian reported current or recent involvement in the full range of same-sex romantic and sexual experiences. The majority of those who identified as straight reported current or recent opposite-sex romantic and sexual experiences and those who identified as bisexual reported a range of romantic and sexual experience with both same- and opposite sex partners. However, adolescents who identified as gay or lesbian were very likely to report opposite-sex romantic and sexual experiences at some point in their lives and a noteworthy minority of adolescents who identified as straight reported a history of same-sex romantic or sexual experiences.

TABLE 2 Percentages of Reported Attractions and Behaviors With Same- and Opposite-Sex Partners in Individuals Who Self-Label as Gay/Lesbian, Straight, and Bisexual (N = 82)

Experiences/behaviors	Same-sex			Opposite-sex
	G/L	Straight	Bi	G/L	Straight	Bi
Crush
Never	0.0	78.6	0.0	20.8	0.0	0.0
At least once in my life	0.0	7.1	6.7	64.6	0.0	0.0
At least once in the past year	12.5	7.1	13.3	8.3	7.1	46.7
Currently	87.5	7.1	73.3	2.1	92.9	53.3
Gone steady
Never	10.4	92.9	13.3	50.0	14.3	26.7
At least once in my life	4.2	7.1	33.3	33.3	7.1	26.7
At least once in the past year	39.6	0.0	13.3	10.4	35.7	26.7
Currently	45.8	0.0	40.0	2.1	42.9	20.0
Saying “I love you”
Never	10.4	57.1	13.3	54.2	21.4	46.7
At least once in my life	6.3	21.4	13.3	22.9	7.1	20.0
At least once in the past year	33.3	0.0	20.0	12.5	21.4	20.0
Currently	50.0	21.4	53.3	6.3	50.0	13.3
Affectionate behaviors
Never	2.1	42.9	6.7	16.7	0.0	0.0
At least once in my life	2.1	42.9	6.7	60.4	14.3	33.3
At least once in the past year	29.2	14.3	33.3	10.4	14.3	40.0
Currently	66.7	0.0	53.3	8.3	71.4	26.7
Petting
Never	6.3	78.6	13.3	47.9	7.1	20.0
At least once in my life	4.2	7.1	20.0	37.5	7.1	33.3
At least once in the past year	39.6	7.1	20.0	8.3	42.9	20.0
Currently	50.0	7.1	46.7	2.1	42.9	26.7
Sexual intercourse
Never	12.5	85.7	33.3	68.8	42.9	60.0
At least once in my life	6.3	14.3	13.3	14.6	7.1	13.3
At least once in the past year	29.2	0.0	26.7	10.4	28.6	13.3
Currently	52.1	0.0	26.7	2.1	21.4	13.3

One-way analyses of variance were (ANOVA) conducted to evaluate the relationships between traditional labels of sexual orientation and ratings on continuous measures of heterosexual and homosexual attraction. The independent variable, sexual orientation, included three levels: gay/lesbian, bisexual, and straight. The dependent variables were the ratings on the two continuous measures of attraction. The ANOVAs were significant for both heterosexual attraction, F (2, 76) = 196.4, p <.001, and homosexual attraction, F (2, 76) = 131.7, p <.001. Games-Howell post hoc tests were conducted to evaluate pairwise differences among the means. There were significant differences between the means for all three groups on the continuous measure of homosexual attraction: straight M = 1.21, SD = .57; gay/lesbian, M = 9.19, SD = 1.02; bisexual M = 6.67, SD = 2.35. All three pairwise comparisons for heterosexual attraction were also significant: straight M = 9.86, SD = .36; gay/lesbian M = 2.29, SD = 1.69; bisexual M = 5.00, SD = 1.73. Cohen's d for the six pairwise comparisons ranged from 1.39 to 9.66.

Open-ended items

In addition to continuum and categorical questions for reporting sexual orientation, participants were asked to answer “What word(s) would best describe the way you see your sexual orientation?” Answers were reviewed and combined thematically. Four different patterns emerged from participants' responses. Sixty-one percent of participants reported a traditional mainstream label (e.g., gay, lesbian, straight). Fourteen percent reported themes of independence from being understood or categorized according to their sexual orientation (e.g., “It's an aspect of my life that does not define who I am”). Ten percent describe seeing their sexual orientation in a negative way (e.g., “conflicting,” “challenging”). Ten percent used words or phrases that denoted a positive theme in regards to their sexual orientation (e.g., “big pile of gay,” “here and queer,” “normal”). Five percent did not respond to this item.

Identity Exploration

Differences between males and females and between middle adolescents and older adolescents were examined with regard to the types of dating and romantic experiences reported. Table 3 presents percentages for same- and opposite-sex crushes, behaviors, and romantic experiences for males and females and for middle and late adolescents. Chi-square analyses were conducted to evaluate whether reports of crushes, romantic, and sexual experiences differed by gender or age. Of the 12 chi-square analyses examining associations between gender and the six behaviors (i.e., crush, gone steady, saying “I love you,” affectionate behavior, petting, sexual intercourse) with same-sex- and opposite-sex partners, four were statistically significant. Gender and opposite-sex crush were related, Pearson χ², (3, N = 80) = 19.59, p < .01, V = .495, as were gender and opposite-sex sexual intercourse, Pearson χ², (3, N = 80) = 8.09, p = .04, V = .318; females were more likely to report current or recent engagement in these behaviors with opposite-sex partners. Gender and same-sex sexual intercourse were significantly related, Pearson χ², (3, N = 82) = 9.80, p = .02, V = .346, as were gender and same-sex steady dating, χ², (3, N = 82) = 7.84, p = .05, V = .309. Males more likely to report current or recent engagement in same-sex steady dating and sexual intercourse. Of the 12 chi-square analyses examining associations between age and romantic or sexual experience with same- and opposite-sex partners, only one was statistically significant. Age and saying “I love you” to the opposite-sex were significantly related, Pearson χ², (3, N = 80) = 8.58, p = .03, V = .328, with late adolescents more likely to report saying “I love you” to the opposite-sex at least once in their lifetimes.

TABLE 3 Percentages of Reported Attractions and Behaviors with Same and Opposite-Sex Partners among Male and Female and among Middle and Late Adolescents (N = 82)

Experiences / behaviors	Same-sex		Opposite-sex		Same-sex		Opposite sex
	Male	Female	Male	Female	Middle	Late	Middle	Late
Crush
Never	8.3	17.4	19.4	6.5	11.1	15.6	8.3	15.6
At least once in my life	0.0	8.7	58.3	26.1	11.1	0.0	36.1	44.4
At least once in past year	11.1	13.0	5.6	23.9	5.6	17.8	22.2	8.9
Currently	77.8	60.9	11.1	43.5	69.4	66.7	33.3	26.7
Gone Steady
Never	22.2	34.8	47.2	32.6	27.8	24.4	44.4	20.0
At least once in my life	2.8	15.2	27.8	26.1	13.9	2.2	25.0	37.8
At least once in past year	38.9	17.4	11.1	23.9	22.2	31.1	16.7	17.8
Currently	36.1	32.6	8.3	17.4	36.1	42.2	13.9	20.0
Saying “I love you”
Never	16.7	23.9	41.7	50.0	16.7	24.4	52.8	40.0
At least once in my life	11.1	13.0	27.8	15.2	16.7	8.9	22.2	20.0
At least once in past year	30.6	17.4	13.9	15.2	25.0	22.2	19.4	11.1
Currently	41.7	45.7	11.1	19.6	41.7	44.4	5.6	24.4
Affectionate Behaviors
Never	8.3	13.0	13.9	6.5	11.1	11.1	13.9	6.7
At least once in my life	5.6	15.2	50.0	43.5	13.9	8.9	47.2	46.7
At least once in past year	30.6	26.1	13.9	19.6	19.4	35.6	13.9	17.8
Currently	55.6	45.7	16.7	30.4	55.6	44.4	25.0	24.4
Petting
Never	11.1	28.3	47.2	26.1	19.4	22.2	41.7	31.1
At least once in my life	2.8	13.0	27.8	32.6	8.3	8.9	22.2	35.6
At least once in past year	41.7	23.9	11.1	21.7	30.6	33.3	19.4	15.6
Currently	44.4	34.8	8.3	19.6	41.7	35.6	16.7	13.3
Sexual Intercourse
Never	16.7	43.5	72.2	54.3	38.9	26.7	58.3	66.7
At least once in my life	5.6	13.0	8.3	15.2	8.3	11.1	13.9	8.9
At least once in past year	30.6	17.4	13.9	13.0	13.9	28.9	19.4	8.9
Currently	47.2	26.1	0.0	17.4	38.9	33.3	8.3	11.1
Note: females N = 36; males N = 46; middle adolescents N = 35; late adolescents N = 47.

Table 4 presents means and standard deviations for the five relationship styles (i.e., close same-sex friendship, same-sex exclusively sexual, same-sex romantic relationship, heterosexual dating, and never participate in relationships) for relational esteem, relational depression, relational preoccupation, self-esteem. Differences among predominant dating styles with regard to self-esteem and relationship competence were assessed via four one-way ANOVAs. The ANOVAs for relational preoccupation, F(4, 75) = 2.07, p = .09 and self-esteem, F(4, 75) = 1.08, p = .38, were not significant. The ANOVA was significant for relational esteem, F(4, 75) = 4.47, p = .003. Sheffe post hoc tests indicated that those who engaged predominantly in either same-sex romantic relationships or heterosexual dating reported higher relational esteem compared to those who do not participate in relationships. Also, the ANOVA was significant for relational depression, F(4, 75) = 2.77, p = .03. Scheffe post hoc tests indicated that those not in relationships reported significantly higher rates of relational depression compared with those who reported same-sex close friends and same-sex romantic relationships. Table 5 presents the results of post hoc tests and Cohen's d estimates of effect size for each pairwise comparison. Examination of the Cohen's d effect sizes in Table 5 demonstrates large effect sizes for the differences between those who reported not participating in relationships and all other groups (i.e., same-sex romantic relationships, same-sex exclusive relationships, passionate friendships, and heterosexual dating), with all relationship styles having higher relational esteem compared to those not participating in relationships. Effect sizes for relational depression demonstrated moderate to large effect sizes for differences between those not participating in relationships and those who reported same-sex romantic relationships, passionate friendships, or heterosexual dating; again, all relationship styles reported less relational depression compared to those not participating in relationships.

TABLE 4 Means and Standard Deviations for Relational Competency and Self-Esteem Outcomes

Relationship Style	Psychosocial outcomes Mean (SD)
	Relational esteem	Relationaldepression	Relationalpreoccupation	Self-esteem
Close same-sex friendships (N = 14)	23.2 (3.6)	20.1 (5.8)	29.8 (7.7)	3.3 (.46)
Same-sex exclusively sexual (N = 5)	24.8 (5.8)	24.0 (13.9)	30.6 (11.8)	3.0 (.74)
Same-sex romantic relationship (N = 40)	24.8 (5.1)	21.7 (7.7)	31.9 (6.8)	3.0 (.62)
Heterosexual dating (N = 9)	24.6 (6.1)	21.1 (10.7)	27.9 (12.0)	3.1 (.50)
Never participate in relationships (N = 8)	16.3 (5.2)	31.1 (5.4)	23.5 (8.0)	2.8 (.49)
Note. Higher scores on the psychosocial subscales correspond to greater relation-esteem, relational depression, relational preoccupation, and self-esteem.

TABLE 5 Pairwise Comparisons and Associated Effect Sizes between Relationship Types for Relational Esteem and Relation Depression

Relationship types	Mean difference	p	Cohen's d
Relational Esteem
SSR-SSE	.00	1.00	.00
SSR-PF	1.58	.92	.30
SSR-HD	.24	1.00	.04
SSR-NP	8.55	>.01	1.65
SSE-PF	1.59	.99	.28
SSE-HD	.24	1.00	.04
SSE-NP	8.55	.11	1.53
PF-HD	−1.34	.97	−.24
PF-NP	6.96	.08	1.70
HD-NP	8.31	.04	1.46
Relational Depression
SSR-SSE	−2.30	.99	−.20
SSR-PF	1.56	.93	.23
SSR-HD	.59	1.00	.06
SSR-NP	−9.43	>.01	−1.41
SSE-PF	3.86	.97	.36
SSE-HD	2.89	.99	.23
SSE-NP	−7.13	.80	−.67
PF-HD	−.97	.99	−.11
PF-NP	−11.00	>.01	−1.95
HD-NP	−10.01	.16	−1.18
Note. SSR = same-sex romantic relationship, SSE = same-sex exclusively sexual, PF = passionate friendship, HD = heterosexual dating, NP = not participating in a relationship.

DISCUSSION

Stemming from social constructionist and multidimensional perspectives, current theories of sexual identity development conceptualize sexual orientation as a complex configuration of identity, attractions, behaviors, disclosures, and interpersonal explorations (Horowitz & Newcomb, 2001). Although recent research has begun to investigate the social constructionist and multidimensional models of identity development, little of this research has used adolescent samples. The current study targeted an adolescent and emerging adult population to evaluate the efficacy of the multidimensional approach in understanding sexual minority development.

Several limitations of the current project should be considered in the interpretation of results. First, consistent with much of the research on sexual minorities, the primary limitation of the current study is the use of a convenience sample in recruiting participants. The initial target population was local sexual minority adolescents; however, due to recruitment difficulties the sample targeted adolescent sexual minorities who were actively engaged with community supports. The difficulties that exist in trying to identify and contact sexual minorities who would be interested in participating are evident. The largest proportion of the sample was recruited from a sexual minority community center “queer prom” activity. Although this activity was open to the community, it is likely that those attending may have been more likely to be in a romantic relationship with the same-sex versus the general population of sexual minority adolescents. Also, the individuals at these activities identifying as straight likely do not represent the general population of those identifying as straight. Further, the geographical location of data collection resulted in a sample in which almost 50% of the participants reported affiliation with the Church of Jesus Christ of Latter Day Saints. The unique characteristics of the sample should be kept in mind during the following discussion of results and replication with broader and more diverse samples will be necessary. Finally, creating a new measure was necessary to answer the research questions for this project. Because of the diversity of research questions, no pre-existing measure would yield adequate information. A strength of the current measure was the addition of continuous measures and open-ended items that represented a multidimensional model approach, missing in existing measures. Limitations of the measure included a lack of questions eliciting specific details regarding motivation, intent, and purpose of different behaviors and experiences. These measurement issues provide directions for future research on this topic.

Identity Development

Age and gender differences

The results of analyses examining differences between middle adolescents (i.e., high school students) and late adolescents/emerging adults (i.e., post high school through 21 years) with regard to self-labeling using either traditional labels or continuum ratings of attraction were not significant. Thus, developmentally, adolescents appear to be identifying and labeling their attractions and sexual orientation at a consistent rate throughout their adolescent years and in to their young adult lives. This finding is reflective of the trend that adolescents are now beginning to identify their sexual orientation at earlier ages (Saltzburg, 2004). Historically, the age of recognition and disclosure was during young adulthood, which likely delayed or made difficult the rich opportunities for interpersonal and romantic development available to adolescents during their middle and high school years (D'Augelli, Hershberger, & Pinkington, 1998). Findings that support earlier recognition in self-labeling and identification are encouraging. As the age of identifying same-sex attraction and disclosure continues to lower to early and middle adolescence, possible developmental gaps between younger and older sexual minority adolescents and young adults in self-labeling and participation in intimate interpersonal relationships may decrease.

The analyses examining differences between males and females in regards to self-labeling using either traditional labels or continuum ratings of attraction were significant. Patterns of endorsement of traditional labels and ratings of heterosexual and homosexual attraction suggested that males' attraction patterns match their use of the traditional label “gay,” whereas female's variability in attraction to both sexes reflects a greater variability in traditional self-labels of sexual orientation. Females appear to be more likely to experience nonexclusive attractions, thus making sexual orientation and identification a much more fluid process compared to males. Past research has documented that females are likely to demonstrate more variability and transitions throughout their lifetimes in regards to their sexual orientation and identity, which is likely a result of more fluid and nonexclusive attractions compared to males (Diamond & Savin-Williams, 2000).

These findings are important and provide support for the social constructionist perspective and insight for future research in this area. First, the social constructionist viewpoint provides important new ways of understanding the meaning of sexual orientation. From this perspective, the use of categorical traditional labels such as gay, bisexual, straight are socially constructed and may therefore be an inadequate medium for capturing the true experience of an individual (Broido, 2000). Females may be more likely to show variability and transitions in the use of traditional labels throughout their lifetime as these labels do not accurately reflect their attractions. The picture with understanding male identity development through this theoretical perspective is more complicated. It is possible that male use of traditional sexual minority labels is more stable because these socially constructed labels fit male experience and attractions adequately. However, it may be possible that social influences impact how males conceptualize their attractions and that taking on a traditional sexual minority label may influence the level to which they commit to and report their attractions on continuous measures. Regardless of sexual orientation, stereotypical gender roles allow women a measure of flexibility and fluidity in their attractions and displayed affection toward both sexes that men are not afforded. Therefore, the differences displayed may represent unique differences between males and females or may be a result of gender role influences that constrict males' ability to demonstrate sexual lability.

Secondly, these findings are important in guiding future research in this area. Several research findings and models of identity development for sexual minority youth have primarily come from samples of gay young men. It has been suggested that sexual minority females are more similar to heterosexual females than they are to young gay males (Savin-Williams & Diamond, 2000). Findings from gay male populations may have limited utility in being generalized to females. Also, limiting sexual minority research to those who identify as a sexual minority will likely exclude numerous females who have same-sex attractions, behaviors, and romantic experiences but do not use the appropriate traditional label for inclusion. Understanding both heterosexual and homosexual attraction through male sexual development and female sexual development may prove more fruitful than grouping sexual orientation and sexual development into different bodies of research (Savin-Williams, 2005).

Labels versus experience

Patterns of association among adolescents' self-labeling of sexual orientation and reports of sexual attraction, sexual behaviors, and romantic experiences were consistent with traditional understandings of sexual orientation labels. Individuals identifying as straight were likely to engage in attractions, behaviors, and romantic experiences with opposite-sex partners, whereas those identifying as gay/lesbian were more likely to engage in these behaviors with the same-sex. Bisexuals were engaging in romantic experiences with both opposite-sex and same-sex partners. Although these results are intuitively logical, they prompt consideration of previous models and conceptualizations that extend from the essentialist perspective. Despite criticism targeted at the inflexibility of the essentialist perspective, the general trend of these results would suggest that traditional labels may have some utility in accurately representing attractions, behaviors, and romantic experiences for different sexual orientation groups (Broido, 2000). Therefore, despite a general push in the literature and by new models to pull away from the use of traditional labels, these labels demonstrate some inherent value in classifying and reflecting different contributing aspects of sexual orientation.

Another interesting aspect of self-identification and labeling was demonstrated through participants' responses to the open-ended questions inquiring how adolescents described their sexual orientation in their own words. A notable subset of respondents answered in a descriptive way, refusing to use labels in regards to their sexual orientation. Within these groups, positive, negative, and normalizing themes emerged in regards to describing sexual orientation as a part of their identity; many participants, however, made it clear that their sexual orientation was not representative of their overall identity. This suggests that labels that affix meaning to sexual orientation may not be representative of all sexual minority youth (Savin-Williams, 2005); however, it must still be recognized that a majority of the sample did use traditional labels in response to these questions, suggesting that these labels have some meaning and practical application. Thus, research is needed to identify the extent to which traditional labels will maintain utility and applicability to identity development in the future.

Notwithstanding both the quantitative and open-ended results, it is important to underscore that across all three groups of sexual orientation there were reports of same- and opposite-sex attraction, affectionate and sexual behavior, and romantic experiences. Sexual minorities (i.e., gay/lesbian, bisexual) were likely to engage in these experiences with an opposite-sex partner more frequently than straight individuals were to engage in them with a same-sex partner. These trends follow previous research that documents sexual minorities engaging in activities with opposite-sex partners for a variety of reasons which include: questioning or experimenting with sexual orientation, trying to hide or deny same-sex feelings, or lacking other alternatives for intimate relationships (Diamond et al., 1999; Zera, 1992). Less understood and researched are reports of same-sex attractions, behaviors, and romantic experiences among individuals who identify as straight. Like their sexual minority counterparts these individuals may be engaging in same-sex attractions and sexual behaviors to explore or question their sexual orientation; however, this knowledge remains a gap in the current study and literature. Future research should consider qualitatively examining motivation and intent behind attractions and engagement with same- and opposite-sex partners. Doing such would provide important information about the purpose and quality of these experiences and how they influence overall adolescent development. Taken together it seems that traditional labels do provide an avenue to accurately conceptualize sexual orientation; however, a diversity of experience continues to exist and measures of attractions and behaviors beyond traditional labels are also needed to compliment and better understand when traditional labels are and are not appropriate.

Identity Exploration

Age and gender differences

Overall, there were relatively few significant differences between males and females in their reports of engagement in same-and opposite-sex romantic and sexual behaviors and only one significant difference between middle and late adolescents. Females were more likely to report current or recent opposite-sex crushes and opposite-sex intercourse and males were more likely to report current or recent same-sex dating and sexual intercourse. This is consistent with other research which identifies females as being more nonexclusive in attractions and which documents that desire appears to be sufficient motivation for males' engagement in same-sex behaviors compared to females (Savin-Williams & Diamond, 2000). The differences that do exist suggest that females' nonexclusive attractions may translate into a greater fluidity in experiences across same- and opposite-sex partners. In consideration of the moderate to large effect sizes that were observed, future research with a larger sample size may find important gender differences.

Relationship styles and psychosocial outcomes

Differences in relational esteem and relational depression were observed between relationship styles (i.e., same-sex romantic relationships, same-sex passionate friendships, same-sex exclusively sexual relationships, heterosexual relationships, or no participation in relationships). Relational-esteem was significantly higher for those participating in same-sex romantic relationships and heterosexual dating compared to those who did not participate in relationships. These differences are understandable as those engaging in dating relationships will have more practice and greater opportunity to relate to another person, thus making their relational-esteem higher than adolescents who never have this opportunity. This provides additional evidence that youth learn important skills by engaging in dating and romantic relationships regardless if they are with the same- or opposite-sex (Diamond, 2003). Relational depression was significantly lower for those in same-sex romantic relationships and same-sex intimate friendships compared to higher rates for those who did not participate in relationships. Both these relationships provide emotional intimacy and social support that are important aspects in relationship satisfaction not enjoyed by those who are withdrawn from relationships.

Although nonsignificant, large effect sizes were observed for pairwise comparisons between all relationship styles (same-sex romantic relationships, same-sex exclusively sexual relationships, passionate friendships, and heterosexual dating) and those who did not participate in relationships. General trends suggested that those who did not participate in relationships had the most negative results across outcomes (e.g., relational esteem, relational depression, self-esteem). Several research studies have documented the negative effects for sexual minorities who become isolated (Alexander, 2001; Martin & Hetrick, 1988; Quinn, 2002; Sullivan & Wodarski, 2002), and it appears that isolation from relationship participation adds less positive relational outcomes to this list. Of course, it is important to recognize that those not participating in relationships likely represent a heterogeneous group of adolescents who are purposefully abstaining, as well as those who, despite desire, are unable to be in a relationship. Also, the large proportion of adolescent sexual minorities who reported engaging predominantly in same-sex romantic relationships in this study is encouraging; these findings suggest that although alternative relationships may be important avenues for developing the capacity to engage in future intimate relationships, adolescents may have more access to meaningful same-sex romantic relationships than previously thought.

Summary

Overall, the results of this study support the tenets of social constructionist theory and multidimensional models of identity development. Desire, behavior, and sexual orientation do not combine in a unitary construct; on the contrary, there is much fluidity in attraction, behavior, and self-identification. It seems important to note that although some benefits may be apparent in relationship styles that promote emotional intimacy, overall engagement in any interpersonal relationship style shows positive benefits compare d to those who do not participate. Taken together, adolescent sexual minority development is a complex process that cannot be understood through a generalized narrow model, but must rather acknowledge meaningful differences among groups and individuals in order to fully capture youths' experiences.

The greatest contribution of future research in this area would be to further explore trends and significant findings of this study with larger, more representative samples. Also, providing more in-depth assessment including qualitative measures that assess personal meaning of developmental events and the motivation and intent behind them will provide a rich source of data that will supplement the current knowledge base. Finally, it is necessary to revisit the importance of influential variables such as age and, especially, gender in regard to sexual development. Despite the importance of these factors, the current study was only able to assess the impact of these variables in a preliminary manner. Sample size limitations precluded the inclusion of gender and age as factors in many analyses. As future research continues, there will likely be more utility in disbanding heterosexual and homosexual research camps and looking at holistic sexual development. Such an approach will provide a potential twofold benefit. First, it is likely that important gender differences will continue to be identified and give greater understanding of the adolescent sexual developmental process that is likely unique for males and females. Secondly, it will help move away from the stigmatization and exclusiveness of conceptualizing sexual minority youth as different from heterosexual peers and provide a normalized picture of development appropriate for this population.

Portions of this manuscript were presented at the Biennial Meeting of the Society for Research in Child Development, Atlanta, GA.