The timing of sexual identity development milestones: Disentangling age from cohort influences

Abstract

Prior studies have reported generational declines in the ages at which sexual minority people first experience sexual identity development milestones (e.g., first awareness of same-sex attraction, self-realization of a sexual minority identity, same-sex sexual behavior, and disclosure); yet such studies have relied on retrospective data from adults that may conflate maturational age influences with cohort influences. In the current study, we disentangled age from cohort influences on milestone timing by harmonizing secondary data from three large and diverse datasets of cisgender sexual minority adolescents ages 15–19 years old (n = 1,310; M_age = 17.34, SD = 1.30) collected at three distinct sociohistorical eras of sexual diversity (the 1990s, the 2000s, and the 2010s). With three age-matched but sociohistorically distinct cohorts, we compared milestone timing and pacing across cohorts, as well as subgroup differences among adolescents with social identities at the intersections of cohort with sex, sexual identity, and race/ethnicity, respectively. A series of multiple regression models suggested that more recent generations of sexual minority adolescents reported earlier ages of first self-identification and disclosure than less recent generations; however, adolescents reported similar ages of awareness of same-sex attraction and same-sex sexual behavior across cohorts. Insights from this study build support for cohort differences in the timing of some sexual identity developmental processes and underscore the importance of accounting for both maturational age and sociohistorical change when measuring the sexual identity development processes of sexual minority youth.

Identifying, acknowledging, and expressing a sexual self are central tasks of identity development (e.g., Erikson, 1968; Hammack, 2014; Marcia, 1980). For decades, social and behavioral scientists have sought to discern how sexual minority people develop sexual identities and how sexual identity development is related to variability in health and thriving (Cass, 1979; Troiden, 1989). Numerous studies have shown that the development of same-sex attractions, sexual behaviors, and identity disclosures have implications for mental, physical, and behavioral health, family relationships, and academic performance (Corliss et al., 2009; Dirkes et al., 2016; Goldbach et al., 2014; Heatherington & Lavner, 2008; Katz-Wise et al., 2017; Legate et al., 2012; Watson et al., 2015). Measuring sexual minority identity development is fundamental to describing sexual minority populations, understanding the health disparities that too often plague them, and reducing the heterosexism that drives these disparities (Institute of Medicine, 2011; National Academies of Sciences, Engineering, and Medicine, 2020).

The developmental timing of sexual identity formation processes influences health and thriving. Numerous studies have shown that relatively earlier or later sexual identity development milestones (D’Augelli, 1994), key events in the formation and expression of a sexual minority identity, such as first same-sex attraction, behaviors, and disclosure, are differentially related to health-relevant outcomes including childhood maltreatment (Corliss et al., 2009), LGBT-related victimization (Rendina et al., 2019), depression, anxiety, and suicidality (Corliss et al., 2009; Dirkes et al., 2016; Katz-Wise et al., 2017), and alcohol use (Fish & Pasley, 2015). For this reason, understanding the factors that influence the timing of such processes is crucial.

Sociohistorical context may shape milestone timing. Given rapid sociocultural changes in the acceptance and visibility of sexual minority people in the past several decades, more recent cohorts of sexual minority people appear to be coming out earlier in the life course (Bishop et al., 2020; Dunlap, 2016). Although a simple comparison of mean ages of disclosure across publications lends face validity to this statement, such an approach confounds many important contextual influences such as the age of participants in the sample, the recruitment methods used, the distinct milestones examined, and whether other influences on milestone timing were accounted for (e.g., social identities such as gender, race/ethnicity, sexual identity). Future research must examine cohort differences in milestone timing in ways that account for age and the additional characteristics listed above to determine whether milestones occur earlier in the life course.

The goal of the current study is to document the timing and pacing (i.e., time between and order) of sexual identity development milestones across a diverse sample of age-matched sexual minority adolescents from three sociohistorically distinct periods. First, we review the theoretical impetus for understanding and describing sexual identity development milestones. We then review key findings on milestone timing across studies and discuss the limitations of this prior research. In response to these limitations, we strategically harmonize three age-matched, cohort-distinct samples of sexual minority youth to investigate the timing and pacing of milestones.

From Stage Models to the Life Course: Putting Sexual Identity Development into Temporal Context

For decades, scholars have sought to describe how sexual minority people form, acknowledge, and share their sexual identities, with important shifts in theoretical and methodological approaches. Early prominent theories of sexual identity development (e.g., Cass, 1979; Troiden, 1989) proposed stage-sequential, linear trajectories defined by phases of identity confusion, comparison, acceptance, and synthesis. The proposed pinnacle of sexual identity development was disclosing a sexual identity and ensuing pride related to this identity. Although formative to the study of sexual minority lives, stage models are widely critiqued as overly linear and acontextual (Diamond, 2006; Morgan, 2013; Mustanski et al., 2014). Tested on small samples comprised mainly of White, gay, adult men in the latter half of the twentieth century, stage models do not represent the diversity inherent to the sexual minority community. Further, the emphasis of stage models on the resolution of psychological conflict as the central task of sexual identity development speaks to their focus on individual psychological processes rather than the social-ecological contexts in which sexual identity development occurs (Mustanski et al., 2014). The assumption that sexual identity development is internal, linear, and stable across all non-heterosexual people has been contested by ample literature (Diamond, 2006; Diamond & Savin-Williams, 2000; Martos et al., 2015) and has given rise to theories that account for the influence of contextual diversity in addition to internal processes.

In response to the theoretical and empirical limitations of stage-sequential approaches to sexual identity development, more contemporary approaches center variability of sexual identity developmental processes and account for the role of social ecologies in forming a sexual minority identity. One contemporary line of research uses life course perspectives (Elder, 1998; Hammack & Cohler, 2011) to center the “when” of sexual identity developmental processes. Life course perspectives account for the role of time—both sociohistorical and developmental—in understanding how people develop sexual identities and how this development, in turn, influences stress and health. For example, historically-bound sociocultural contexts define possibilities for developing sexual identities in that those with same-sex attractions today have access to communities, lexicons, and education that were not available to a same-sex-attracted person several decades ago. Further, the developmental timing of these processes can exert influence on the life course. An adolescent who discloses a sexual minority identity is dependent on family members, teachers, and other adults for their basic needs, whereas an adult who discloses their identity may be relatively more independent (Russell & Fish, 2019). Therefore, the maturational age during which these processes unfold can shape future trajectories of health and well-being.

Life course approaches are particularly relevant for studying contextual influences on the sexual identity development of sexual minority people. In the past several decades, the pace at which public opinion has shifted regarding sexual minority identity and acceptance is unparalleled. The Stonewall Inn riots of 1969 (Carter, 2010) unearthed the reality of sexual and gender diverse identities to the public, shifting the concept of non-heterosexuality into a more mainstream historical narrative. Although the removal of homosexuality from repertoires of psychiatric illness in 1973 (American Psychiatric Association, 1980; Drescher, 2015) nudged sexual diversity away from associations with mental illness and towards normativity, the HIV/AIDS epidemic beginning in the 1980s recast shadows of illness onto sexual diversity (Hammack et al., 2018). Youth developing sexual minority identities in the 1980s and early 1990s heard HIV/AIDS described as the “gay plague” and watched “Don’t ask don’t tell” policies enter mainstream media (Bishop et al., 2021). The generation of sexual minority youth coming of age in the early 2000s experienced greater visibility than prior cohorts following cultural events such as Ellen Degeneres’ coming out on network television (Dow, 2001). However, in subsequent years, evidence of discrimination and persecution continued, including a rising tide of state constitutional amendments to ban same-sex marriage that swept across the United States (Soule, 2004). In the 2010s, sexual and gender diversity gained further legal ground and media visibility, with public advances such as federally-sanctioned marriage equality at the forefront. At the same time, 2023 was the worst year in recent history for LGBTQ legislative attacks, particularly for LGBTQ youth. In 2023 alone, more than 500 bills across 49 states have been proposed to limit LGBTQ youth’s rights to receive medical care, play sports, and have their identities reflected in school curricula (Trans Legislation Tracker, 2023). In sum, historically-bound social configurations of sexual identities have shaped how individuals and cultures understand and develop sexual identities, and studies of sexual identity development must account for these sociohistorical contexts.

Sexual Identity Development Milestones and Sociohistorical Change

Methodologically, the “when” of sexual identity development is operationalized by measuring sexual identity development milestones (i.e., “milestones”), the ages at which sexual minority people first experience key developmental events, such as first awareness of same-sex attraction, first self-identification with a sexual minority identity, first same-sex sexual experiences, and first disclosure of sexual minority identities to others (D’Augelli, 1994; Rosario et al., 2011). Studies measuring milestones have documented ample diversity in identity formation processes (Grov et al., 2006; Martos et al., 2015) that make clear the need to examine associations with health and wellbeing. Milestones also facilitate deeper understandings of sexual identity formation processes from life course approaches. If milestones shift to earlier developmental stages, new approaches to supporting sexual minority people as they develop identities are needed. School, parental, and peer relations vary across developmental contexts (Poteat & Anderson, 2012), and understanding the timing of milestones can provide insight into developmentally appropriate supports during particularly meaningful sexual identity developmental processes (e.g., first disclosures of sexual identity to family and friends).

Several studies have documented how milestone ages and pacing have shifted across sociohistorical time (e.g., Bishop et al., 2020; Calzo et al., 2011; Grov et al., 2018; Martos et al., 2015), yet key data constraints have limited their findings. These studies typically collect data from a sample of age-diverse sexual minority people grouped into distinct birth cohorts. Then, the mean age of milestones is determined and compared. For example, several studies have shown that milestone timing is earlier and accelerated among more recent cohorts relative to older cohorts by comparing the milestone ages of participants who fit different age groups and therefore represent different birth/historical cohorts (Bishop et al., 2020; Grov et al., 2018; Martos et al., 2015). Although such approaches may capture distinctions across age groups, they confound age and cohort influences. Because such studies examine differences across age groups in a cross-sectional design, there are different age ranges for each cohort in the sample and thus different possible response options. For example, study participants 18–24 years old can only report milestones that occurred at 24 years old or younger, whereas a 60-year-old participant can report milestones that occurred at a maximum age of 60 years old. Prior research suggests that sexual minority identity emerges across the life course (Diamond, 1998; Diamond, 2008), with one recent study finding that nearly one out of six respondents in a national probability sample of sexual minority adults reported first self-identifying, engaging in same-sex sexual behavior, and disclosing a sexual minority identity in adulthood (Bishop et al., 2022). Thus, studies that rely on distinct age cohorts to measure milestone ages may report diversity across ages rather than cohorts. Further, retrospective bias may operate differently across age-diverse samples (Hegarty, 2009). Although some research suggests that sexual minority adolescents and young adults’ recollection of milestones is reliable (Schrimshaw et al., 2006), recall bias of milestones has not been examined among relatively older sexual minority people. Studies with age-matched samples collected at distinct sociohistorical periods offer a necessary advancement for documenting whether shifting sociohistorical contexts contribute to earlier milestones.

Prior studies have also largely examined the earlier or later emergence of milestones, rather than attempting to understand whether and why specific milestones may occur progressively earlier, whereas others do not. For example, the emergence of same-sex sexual attraction may be linked with pubertal processes that are less influenced by historical changes than disclosure timing, as evidenced by the relatively less variability in milestone timing of same-sex attraction than other milestones (Bishop et al., 2022; Drasin et al., 2008; Hall et al., 2021; Herdt & McClintock, 2000). Understanding whether and how individual milestones are differentially influenced by sociohistorical contexts is important for understanding sensitive periods for developmentally appropriate support.

Although research on milestones has primarily examined mean ages of key events, other temporal indicators such as milestone pacing (i.e., the time between and order) have distinct and important implications for understanding sexual identity development (Costa et al., 2022). One prior study showed that in addition to the age of milestones such as the first self-realization and disclosure of a sexual minority identity, the time between these milestones was related to problem drinking behavior among sexual minority people (Parks & Hughes, 2007). Specifically, longer periods between first wondering and first deciding on a sexual minority identity were associated with more alcohol consumption. Another recent study found that anti-LGB structural stigma was associated with a longer duration between sexual minority self-identification and disclosure (Layland et al., 2023). Additional studies have described important demographic differences in milestone pacing by age cohort, sex, and race/ethnicity (Bishop et al., 2020; Grov et al., 2018). For example, “identity-centered” milestone ordering, characterized by self-identification with a sexual minority identity preceding same-sex sexual behaviors, was associated with lower internalized homophobia and less risky sexual behaviors than “sex-centered” development, in which same-sex sexual behaviors precede self-identification with a sexual minority identity (Dubé, 2000; Schindhelm & Hospers, 2004). Prior studies have also reported generational differences in the prevalence of “identity-centered” ordering of milestones, with more recent generations reporting higher rates of identity-centered ordering (Hall et al., 2021). Additional studies that account for life course influences are needed to understand better the degree to which milestone pacing has changed across sociohistorical time.

Sexual Identity Development and Subgroup Differences

Sexual minority people’s identity development is not homogeneous. Rather, the timing and pacing of sexual identity development processes is likely shaped by privilege and oppression stemming from mutually constitutive social identities such as gender, sexual identity, and race/ethnicity. Intersectionality theory (Cohambee River Collective, 1977; Crenshaw, 1989; Mallory & Russell, 2021) is an analysis and critique of the ways that interlocking systems of privilege and oppression, such as racism, sexism, and homophobia, shape identity-based inequalities. Acknowledging that experiences of oppression and privilege likely shape sexual identity formation processes, several studies have investigated subgroup differences in milestones by sex, sexual identity, and race/ethnicity and have found that male people tend to report earlier ages of self-identification and same-sex sexual behavior than female people (Hall et al., 2021; Herek et al., 2010; Martos et al., 2015). With regard to subgroup differences by sexual identity, gay/lesbian people report earlier average ages of first self-identification than bisexual people (Martos et al., 2015; Calzo et al., 2011). Prior research is mixed with respect to racial/ethnic variability in milestone timing. Some studies have found no differences in milestone timing by race/ethnicity (Grov et al., 2006; Martos et al., 2015), whereas others have found that sexual minority people of color report earlier ages of first same-sex sexual attraction, first self-identification as a sexual minority, and first same-sex sexual behavior than their White counterparts (Bishop et al., 2020; Parks et al., 2004).

Taken together, the timing and patterning of milestones are not uniform across those with diverse social identities. In conjunction with growing evidence that milestones vary by sociohistorical context, it is possible that both social identities and sociohistorical context, together, shape milestone timing and patterning. For example, growing cultural acceptance of bisexual identity and visibility in the past three decades may result in shifting timing of milestones across generations (Pollitt & Roberts, 2021), however, bisexual youth may still lag in milestone timing relative to monosexual youth due to biphobia, which arises from both sexual minority and sexual majority communities (Wandrey et al., 2015).

A burgeoning line of research suggests that identity-based subgroup differences may also be distinctive across cohorts. One recent study examined subgroup variability in milestone timing and pacing at the intersections of sociohistorical cohort with sex, sexual identity, and race/ethnicity, respectively, and found far more subgroup differences between generations than within generations, even when participants held distinct social identities (Bishop et al, 2020). Additionally, a recent meta-analysis of milestone timing and pacing showed that, on average, male respondents reported earlier ages of first same-sex attraction, self-identification, and sexual behavior than female participants, however, fewer differences were found among Millennial participants relative to participants from prior generations (Hall et al., 2021). In sum, these findings suggest that sociohistorical context may be a key contributor to the timing and pacing of milestones. At the same time, they did not used age-matched samples of sexual minority people, and therefore, implications for how subgroups of sexual minority people with distinct social identities vary in milestone timing across sociohistorical time are limited. Just as understandings of sexual identity have shifted in the past three decades, so have cultural narratives of gender and race/ethnicity (Doane, 2017; Forman & Lewis, 2015; Scott Carter et al., 2009). An examination of shifting milestone timing across the life course therefore also requires investigating these processes in the context of racial/ethnic and gender identities.

The Current Study

The current study aims to assess sexual minority youth’s milestone timing and pacing using methods that disentangle age and cohort influences. To do so, we harmonize three sociohistorically distinct, age-matched samples of sexual minority youth to describe: 1) the timing and pacing of sexual identity development milestones across three distinct sociohistorical periods; 2) whether these shifts vary among subgroups of sexual minority youth at the intersection of sex, race/ethnicity, and sexual identity with cohort. Consistent with prior research, we hypothesize that sexual minority youth in more recent cohorts will report relatively younger ages of milestones, as well as less time between milestones. Additionally, in line with prior studies, we expect that gay/lesbian youth, youth of color, and male youth from more recent cohorts will report relatively younger ages and less time between most milestones than bisexual youth, White youth, and female youth from less recent cohorts.

Methods

Data

Data come from three secondary datasets of sexual minority youth who were adolescents in sociohistorically distinct periods: the Challenges and Coping Study (i.e., CCS; D’Augelli et al., 2001), the Victimization and Mental Health among High Risk Youths Study (i.e., VMH study; D’Augelli & Grossman, 2006), and the Risk and Protective Factors for Suicide among Sexual Minority Youth Study (i.e., RPF; Baams et al., 2015). These datasets provide unique opportunities for comparison. The CCS contains data from sexual minority youth who were adolescents in the 1990s, the VMH sample were teens in the 2000s, and the RPF sample were adolescents in the 2010s. Additionally, overlapping research teams developed the surveys, each sample comprises sexual minority youth recruited primarily from community-based organizations, and each survey contains similar measures of sexual identity development milestones (see Supplementary Table 1).

The Challenges and Coping study (i.e., CCS; D’Augelli et al., 2001) is a cross-sectional sample of 569 sexual minority youth ages 14–31 who were surveyed about coping with challenges related to their sexual orientation. About 78% of the sample were between the ages of 14–22, and 93% were under 25 years old. Youth were recruited through community-based organizations (CBOs) and universities across the United States and Canada, and online. An adult study liaison was present at each site and supervised youth who opted to complete the survey after consenting (see D’Augelli et al., 2001 for additional details). Data were collected between 1995 and 1996. This historical period was characterized by structural and interpersonal stigma directed toward sexually minority people. For example, in 1996, former President Bill Clinton signed the Defense of Marriage Act, which federally banned marriage for same-sex couples. The notion of same-sex sexuality as indicative of physical and/or mental illness was evident in the cultural conflation of gay identity with the HIV/AIDS epidemic (Epstein, 1988). At the same time, sexual minority people were engaging in political organizing, such as the first March on Washington for LGB rights in 1993.

The Victimization and Mental Health among High Risk Youths Study (i.e., VMH study; (D’Augelli & Grossman, 2006) is a three-wave longitudinal study of sexual orientation-based victimization and health among 528 lesbian, gay, and bisexual (LGB) youth. In the current study, we use data from Wave 1. Youth were 15–19 years old at Wave 1 and were recruited from three community based organizations in New York City and its surrounding suburbs. Snowball sampling was also used to diversify the sample. Each study site had a contact person who was a liaison for the research project. Youth first completed consent materials and a self-administered questionnaire, and then a face-to-face interview with a trained interviewer (see D’Augelli & Grossman, 2006 for additional details). The first wave of data was collected between 1999 and 2001. Youth in this study were adolescents during growing visibility and shifting attitudes regarding minoritized sexual identity: in 1997, Ellen DeGeneres came out on national television, and the following year, Matthew Shepherd’s brutal bias-based murder captured national attention (Walters, 2003). These events made evident the continued stigma of sexual minority youth and led to several advancements such as the inclusion of sexual orientation as a protected status in hate crime legislation in several states.

The Risk and Protective Factors for Suicide among Sexual Minority Youth study (i.e., RPF; Baams et al., 2015) is a four-wave panel study of risk and protective factors for suicide among 1,061 lesbian, gay, bisexual, transgender, and queer (LGBTQ) youth ages 15–21 years old. Youth were recruited from three community-based organizations and college groups located in major cities in the Northwest, Southwest, and West Coast of the United States and via snowball sampling. After the initial screening, eligible participants contacted site coordinators to confirm an appointment to complete a survey packet. Participants provided consent and completed the survey packet at the study site (see Baams et al., 2015 for additional details). In the current study, we use data from Wave 1, which was collected from 2011–2012. Sexual minority youth in this period were adolescents during an age of greater visibility alongside ongoing stigmatization of sexual diversity. A growing number of state legislatures had legalized marriage for same-sex couples, yet there remained a lack of federal non-discrimination regulations for sexual minority people. The internet became increasingly available to youth, improving access to information and education regarding sexual diversity.

For clarity, we refer to the CCS dataset as the “1990s cohort”, the VMH sample as the “2000s cohort”, and the RPF sample as the “2010s cohort” moving forward. Participants had slightly different age ranges across cohort samples, with the 2000s cohort containing the most limited age range (ages 15–19). To disentangle age from cohort influences, we limited our sample to participants ages 15–19 years old. Additionally, we limited the sample to cisgender participants, as transgender adolescents are theorized to experience distinct identity development milestones (Puckett et al., 2021). Table 1 describes the demographic characteristics of our final analytic sample, which were 1,310 sexual minority youth (M_age = 17.34, SD = 1.30) with valid responses to age, sex, sexual identity, race/ethnicity, and at least one milestone question. Nine participants in the 2010s cohort reported a milestone age that was older than their age, and their response for that milestone was set to missing. About 46% of the sample were from the 2010s cohort, 39% were from the 2000s cohort, and 15% were from the 1990s cohort. The combined sample was diverse concerning race/ethnicity. Slightly under half (47%) of the combined sample were White, approximately 20% were Black, and 33% reported another race (i.e., Asian American, Native American, Native Hawaiian, East Indian, American Indian/Alaska Native, Another Race, Latinx, Multiracial, or race unreported). The sample was divided nearly evenly between male (48%) and female (52%) respondents. Slightly more participants in the sample identified as lesbian/gay (50%) relative to bisexual (47%), and a small group (3%) reported another sexual identity.

Table 1.

Demographic characteristics of participants

Variable	CCS (1990s) (n = 192)		VMH (2000s) (n = 512)		RPF (2010s) (n = 606)		Full analytic sample (n = 1,310)
	n/mean	%/SD	n/mean	%/SD	n/mean	%/SD	n/mean	%/SD
Sex
Female	87	45.31	245	47.85	353	58.25	685	52.29
Male	105	54.69	267	52.15	253	41.75	625	47.71
Race
Black/African American	16	8.33	128	25.00	121	19.97	265	20.23
White	147	76.56	319	62.30	146	24.09	612	46.72
Another Race	29	15.10	65	12.70	339	55.94	433	33.05
Sexual Identity
Lesbian/gay	131	68.23	238	46.48	285	47.03	654	49.92
Bisexual	55	28.65	274	53.52	284	46.86	613	46.79
Another sexual minority identity	6	3.13	−a	− a	37	6.11	43	3.28
Age (years)	18.21	.89	17.02	1.28	17.33	1.30	17.34	1.30

Note.

^aOption not provided. Column total percentages presented.

Measures

Milestone age.

Four items were used to assess the age at which participants reported milestones (D’Augelli et al., 2008). Age of first awareness of same-sex attraction was assessed in each study with responses to an open-ended question about the age at which participants first became aware that they were attracted to or felt different from people of the same sex, even though they may not have labeled these feelings. The item from the CCS reads, “How old were you when you first became aware that you were attracted to people of the same sex, even though you might not have labelled these feelings?” Age of first same-sex sexual behavior was assessed with responses to an open-ended question about the age at which participants had their first sexual experience with a male/female, respectively. A sample item from the RPF survey reads, “Did you engage in sexual activity with a [male/female] in the last year? If yes, how old were you the first time you engaged in sexual activity with a [male/female]?” Responses were retained for the item that matched the respondent’s sex, such that males’ responses were coded as the age at which they first engaged in sexual behavior with a male, and females’ responses were coded as the age at which they first engaged in sexual behavior with a female. Age of first sexual minority self-identification was assessed with responses to a question about how old participants were when they first self-identified with a minoritized sexual orientation. A sample item from the VMH survey reads, “How old were you when you first considered yourself LGBT?” Finally, age of first sexual minority disclosure was assessed with responses to an open-ended question about how old participants were when they first told someone that they identified with a sexual minority identity. A sample item from the CCS study reads, “How old were you when you first told someone that you were lesbian, gay, or bisexual?” All responses were coded in years. Participants who were missing responses, reported that they didn’t know, or had not experienced a given milestone were excluded from the analyses of the associated milestone.

Milestone pacing.

Four variables were generated to assess the pacing of (i.e., time between and order of) milestone ages. The differences between milestone ages were computed for: (1) same-sex sexual behavior and self-identification; (2) same-sex sexual behavior and disclosure; and (3) self-identification and disclosure; and (4) same-sex awareness of attraction and disclosure. Early sexual identity development theorists (e.g., Troiden, 1989) proposed that sexual identity processes follow a set order of same-sex awareness of attraction, same-sexual sexual behavior, self-identification, and disclosure. However, recent evidence suggests variability in the ordering of milestones (Bishop et al., 2020; Calzo et al., 2011; Floyd & Stein, 2002). Therefore, in line with prior studies of milestone pacing (Martos et al., 2015), we examine pacing variables that test several possible orders of milestones and retain negative values to represent participants who follow the ordering that diverges from that proposed by prior theorists.

Cohort.

The measure for the three distinct cohort groups was constructed from respondents’ participation in either the CCS study (i.e., 1990s cohort), the VMH study (i.e., 2000s cohort), or the RPF study (i.e., 2010s cohort).

Covariates.

Several demographic covariates were measured. Participants’ sex was coded as male or female according to their reported sex, and female was the reference group. Participants’ sexual identity was coded as lesbian/gay, bisexual, or another sexual identity (e.g., queer, pansexual, questioning, or another sexual minority identity), and lesbian/gay was the reference group. Finally, we created a race/ethnicity variable consisting of White, Black, and Another race/ethnicity (i.e., Asian American, Native American, Native Hawaiian, East Indian, American Indian/Alaska Native, Another Race, Latinx, Multiracial, or race/ethnicity unreported) given that races/ethnicities other than White and Black were not consistently measured across datasets, and White was the reference group.

Analytic Approach

Data were analyzed in Stata 18 (StataCorp, 2023). First, among those who reported milestone ages, we conducted a series of multiple regression analyses to test differences between milestone ages and pacing by cohort, sex, sexual identity, and race/ethnicity, as well as subgroups at the intersections of these identities, controlling for the effects of maturational age. For each milestone, we estimated separate models and employed robust standard errors to account for heteroskedasticity. We then examined Wald F tests to determine whether the parameter contributed significant variance to the model. Finally, we estimated marginal means for each demographic group, followed by post-hoc pairwise comparison procedures to test mean differences between groups, with Bonferroni adjustments to reduce type 1 error. For models that included interactions, we tested each interaction separately. For models testing the interaction of cohort with sexual identity, we used a two-category sexual identity variable (i.e., lesbian/gay vs. bisexual+) given that the 1990s dataset only contained response options of “lesbian/gay” and “bisexual”. We also conducted several sensitivity analyses, which we describe in the results.

Results

Table 2 reports estimated marginal means from a series of multiple regression models testing sociodemographic differences in the age of milestones by social identities including cohort, sex, sexual identity, and race/ethnicity, adjusted for each demographic identity and maturational age. With respect to ages of awareness of same-sex attraction, there were significant differences by groups defined by sexual identity [F(2,1284) = 38.98, p<.01]. Post hoc pairwise comparisons revealed that lesbian and gay participants (M = 9.52, SE = .13) reported earlier ages of awareness than both bisexual participants (M = 11.24, SE = .15) and those with another sexual minority identity (M = 11.90, SE = .71), who did not differ from one another. An F test of cohort differences in awareness of same-sex attraction was significant [F(2,1293) = 3.23, p = .04], but post-hoc pairwise comparisons showed no significant differences between groups. Finally, groups did not significantly differ by sex nor race/ethnicity in the age at which they reported awareness of same-sex attraction.

Table 2.

Marginal means from regression analyses comparing ages of milestones by cohort, sex, sexual identity, and race/ethnicity

	First Awareness			First Self-Identification			First Same-Sex Sex			First Disclosure
	M	SE	F	M	SE	F	M	SE	F	M	SE	F
Cohort			3.23*			9.45***			0.10			42.67***
CCS (1990s)	10.67	0.26		14.89ab	0.16		14.44	0.24		15.84ab	0.13
VMH (2000s)	10.08	0.16		14.10a	0.10		14.41	0.14		14.68ac	0.09
RPF (2010s)	10.58	0.16		14.07b	0.11		14.51	0.15		14.21bc	0.10
Sex			0.43			8.66**			6.92**			0.49
Female	10.46	0.13		14.38a	0.08		14.69a	0.12		14.59	0.08
Male	10.33	0.14		14.00a	0.10		14.22a	0.13		14.67	0.08
Sexual Identity			38.98***			27.76***			8.68***			26.60***
Lesbian/Gay	9.52ab	0.13		13.77a	0.09		14.20a	0.13		14.22a	0.08
Bisexual	11.24a	0.15		14.70a	0.08		14.85ab	0.13		15.08a	0.08
Another ID	11.90b	0.71		13.43	0.89		12.14b	1.06		14.42	0.70
Race/Ethnicity			1.78			1.43			0.34			.22
White	10.34	0.15		14.13	0.10		14.50	0.14		14.62	0.09
Black	10.75	0.21		14.41	0.14		14.32	0.19		14.70	0.13
Another Race	10.26	0.19		14.17	0.12		14.48	0.17		14.60	0.10

Note. The same superscripts indicate significant differences at the p<.05 level. Bonferroni post-hoc comparisons were used to minimize Type 1 error. Estimated marginal means adjusted for the effects of cohort, sexual identity, sex, race/ethnicity, and age are reported. Contrasts with significant omnibus F values bolded.

^*p < .05,

^**p < .01

^***p < .001.

In models testing sociodemographic differences in the age at which participants first self-identified with a sexual minority identity, timing varied by groups defined by cohort [F(2,1271) = 9.45, p<.01], sex [F(1, 1271) = 8.66, p<.01], and sexual identity [F(2, 1271) = 27.76, p<.01]. Post hoc pairwise comparisons revealed that the 2000s cohort (M = 14.10, SE = .10) and the 2010s cohort (M = 14.07, SE = .11) both reported significantly younger ages of self-identification than the 1990s cohort (M = 14.89, SE = .16) and did not differ from one another. Additionally, female youth (M = 14.38, SE = .08) reported later ages of self-identification than male youth (M = 14.00, SE = .10). With respect to differences by sexual identity, bisexual participants (M = 14.70, SE = .08) reported later ages of self-identification than lesbian/gay (M = 13.77, SE = .09) participants and those with another sexual identity did not differ from other groups. No significant differences in the age of self-identification by race/ethnicity were observed.

Differences in the ages of first same-sex sexual behaviors were present among groups defined by sex [F(1, 930) = 6.92, p<.01] and sexual identity [F(2, 930) = 8.68 p<.01]. Specifically, female adolescents (M = 14.69, SE = .12) reported later ages of same-sex sexual behaviors than male adolescents (M = 14.22, SE = .13). Bisexual youth (M = 14.85, SE = .13) reported significantly older ages of same-sex sexual behavior than lesbian/gay youth (M = 14.20, SE = .13) and youth with another sexual identity (M = 12.14, SE = 1.06), who did not differ from each other. There were no statistically significant differences in groups defined by cohort nor by race/ethnicity in the ages of first same-sex sexual behavior.

Finally, differences in the age at which participants first disclosed their identities emerged by groups defined by cohort [F(2, 1265) = 42.67, p<.01] and sexual identity [F(2, 1265) = 26.60, p<.01]. Specifically, the 2010s cohort reported the youngest mean ages of first disclosure (M = 14.21, SE = .10), which were significantly earlier than disclosure ages of the 2000s cohort (M = 14.68, SE = .09), which reported significantly earlier ages than the 1990s cohort (M = 15.84, SE = .13). With respect to differences by sexual identity, lesbian/gay participants disclosed their sexual identities at earlier ages (M = 14.22, SE = .08) than bisexual participants (M = 15.08, SE = .08) and those who were questioning did not significantly differ from other groups. There were no statistically significant differences in groups defined by sex nor by race/ethnicity in ages of first disclosure.

Table 3 presents estimated marginal means from a series of multiple regression models testing group differences in milestone pacing. Regarding mean differences in the pacing from same-sex sexual behavior to self-identification, F tests suggested differences across subgroups defined by cohort [F(2, 915) = 4.33, p=.01] and race/ethnicity [F(2, 915) = 5.15, p < .01], and no differences by sex nor by sexual identity. Regarding cohort differences, post hoc pairwise comparisons showed that participants in the 1990s cohort reported a positive difference in time between same-sex sexual behavior and self-identification (i.e., same-sex sexual behavior preceded self-identification on average; M = 0.44, SE = .25) that significantly differed from both the 2000s (M = −0.35, SE = .14) and 2010s (M = −0.45, SE = .17) cohorts, both of which reported a negative difference that did not differ from one another. Additionally, Black participants (M = 0.28, SE = .18) varied from White participants (M = −0.41, SE = .15) and participants of another race (M = −0.37, SE = .19) in that Black participants reported a positive difference, whereas other participants reported a negative difference in time between same-sex sexual behavior and self-identification.

Table 3.

Marginal means from regression analyses comparing pacing of milestones by cohort, sex, sexual identity, and race/ethnicity

	Sex to Self-Identify			Sex to Disclose			Self-Identify to Disclose			Aware to Disclose
	M	SE	F	M	SE	F	M	SE	F	M	SE	F
Cohort			4.33*			14.17***			14.34***			13.04***
CCS (1990s)	0.44ab	0.25		1.38ab	0.25		0.92a	0.12		5.23a	0.27
VMH (2000s)	−0.35a	0.14		0.17a	0.15		0.56b	0.09		4.62b	0.15
RPF (2010s)	−0.45b	0.17		−0.29b	0.16		0.14ab	0.08		3.70ab	0.16
Sex			0.22			6.95**			19.29***			1.13
Female	−0.20	0.13		−0.05a	0.13		0.20a	0.06		4.19	0.13
Male	−0.29	0.14		0.45a	0.13		0.67a	0.09		4.40	0.14
Sexual Identity			1.52			2.40			0.64			10.86***
Lesbian/Gay	−0.36	0.14		0.12	0.14		0.44	0.08		4.73a	0.14
Bisexual	−0.15	0.13		0.25	0.13		0.38	0.07		3.87a	0.14
Another ID	1.41	1.25		2.41	1.07		0.80	0.42		3.28	0.63
Race/Ethnicity			5.15**			1.19			1.40			1.36
White	−0.41a	0.15		0.16	0.15		0.50	0.07		4.36	0.14
Black	0.28ab	0.18		0.47	0.21		0.26	0.12		3.98	0.21
Another Race	−0.37b	0.19		0.09	0.18		0.41	0.10		4.39	0.18

Note. The same superscripts indicate significant differences at the p<.05 level. Bonferroni post-hoc comparisons were used to minimize Type 1 error. Estimated marginal means adjusted for cohort, sexual identity, sex, race/ethnicity, and age are reported. Contrasts with significant omnibus F values bolded.

^*p < .05,

^**p < .01

^***p < .001.

Concerning pacing of same-sex sexual behavior to disclosure, subgroup differences emerged among groups defined by cohort [F(2, 915) = 14.17, p<.01] and sex [F(1, 915) = 6.95, p < .01]. Regarding cohort differences, post hoc pairwise comparisons revealed that the pacing of the 1990s cohort (M = 1.38, SE = .25) differed from the 2000s (M = 0.17, SE = .15) and 2010s (M = −0.29, SE = .16) cohorts, respectively. The 2000s and the 2010s cohorts did not differ from one another. Regarding sex differences, male youth reported more time between same-sex sexual behaviors with disclosure on average (M = 0.45, SE = .13) than female youth (M = −0.05, SE = .13). There were no subgroup differences by sexual identity nor by race/ethnicity.

F-tests of a model examining pacing between the age of first self-identification as a sexual minority and age of first disclosure of a sexual minority identity revealed subgroup differences by cohort [F(2, 1254) = 14.34, p<.01] and sex [F(1, 1254) = 19.29, p<.01. The 2010s cohort spent less time between self-identification and disclosure (M = 0.14, SE = .08) than the 2000s (M = 0.56, SE = .09) and 1990s (M = 0.92, SE = .12) cohorts, who did not vary from one another. Male youth (M = 0.67, SE = .09) spent more time between self-identification and disclosure than female youth (M = 0.20, SE = .06). No subgroup differences emerged with respect to sexual identity nor race/ethnicity.

Finally, in a model testing the time between first awareness of same-sex attraction and first disclosure, F-tests revealed subgroup differences by cohort [F(2, 1248) = 13.04, p<.01] and sexual identity [F(2, 1248) = 10.86, p<.01]. The 2010s cohort (M = 3.70, SE = .16) spent significantly less time between first awareness and first disclosure compared to the 2000s (M = 4.62, SE = .15) and 1990s cohort (M = 5.23, SE = .27), who did not differ from one another. Lesbian/gay participants (M = 4.73, SE = .14) spent significantly more time between first awareness and first disclosure than bisexual participants (M = 3.87, SE = .14), and those with another sexual minority identity (M = 3.28, SE = .63) did not differ from other groups. No subgroup differences emerged by sex nor by race/ethnicity.

Intersectional Models

Table 4 presents estimated marginal means from a series of multiple regression models comparing mean differences in ages of milestones by subgroups at the intersection of cohort with sex, sexual identity, and race/ethnicity, respectively. F-tests revealed significant differences in the age of first same-sex sexual behavior by subgroups at the intersection of cohort and sex [F(2, 929) = 8.66 p<.01]. More specifically, males in the 2000s (M = 13.89, SE = .21) reported earlier ages of same-sex sexual behavior than females in the 2000s (M = 15.03, SE = .17). No other significant differences were observed across subgroups.

Table 4.

Marginal means from regression analyses comparing ages of milestones at the intersection of cohort with sex, sexual identity, and race/ethnicity

	Awareness of Attraction			Self-Identification			Same-Sex Sex			Disclosure
	M	SE	F	M	SE	F	M	SE	F	M	SE	F
Cohort x Sex			0.15			1.89			8.66***			0.18
1990s Female	10.72	0.40		15.17	0.22		14.74	0.36		15.73	0.18
1990s Male	10.62	0.31		14.60	0.22		14.06	0.32		15.92	0.16
2000s Female	10.18	0.21		14.44	0.12		15.03a	0.17		14.69	0.11
2000s Male	9.92	0.22		13.80	0.15		13.89a	0.21		14.71	0.12
2010s Female	10.62	0.20		14.12	0.14		14.37	0.20		14.16	0.13
2010s Male	10.60	0.25		14.02	0.18		14.81	0.22		14.25	0.16
Cohort x Sex ID			0.19			1.70			0.95			0.63
1990s Gay/lesbian	9.80	0.31		14.44	0.20		14.18	0.30		15.50	0.16
1990s Bi +	11.55	0.43		15.36	0.23		14.59	0.40		16.06	0.19
2000s Gay/lesbian	9.10	0.21		13.55	0.15		14.07	0.21		14.30	0.13
2000s Bi +	11.00	0.21		14.69	0.12		14.89	0.18		15.10	0.11
2010s Gay/lesbian	9.79	0.22		13.76	0.15		14.36	0.21		13.74	0.14
2010s Bi +	11.43	0.23		14.37	0.16		14.66	0.22		14.66	0.15
Cohort x Race/eth			2.22			1.88			0.99			0.37
1990s White	10.55	0.28		14.88	0.18		14.38	0.28		15.85	0.14
1990s Black	10.56	0.57		13.90	0.57		13.63	0.79		15.66	0.49
1990s Another race	11.19	0.69		15.13	0.25		15.04	0.47		15.79	0.22
2000s White	10.08	0.19		13.97	0.12		14.58	0.17		14.66	0.10
2000s Black	9.95	0.29		14.54	0.18		14.29	0.27		14.74	0.17
2000s Another race	10.41	0.40		14.14	0.26		14.19	0.37		14.85	0.20
2010s White	10.46	0.31		14.08	0.22		14.41	0.33		14.21	0.23
2010s Black	11.48	0.32		14.24	0.23		14.57	0.27		14.36	0.21
2010s Another race	10.31	0.21		13.99	0.14		14.47	0.19		14.13	0.12

Note. The same superscripts indicate significant differences at the p < .05 level. Each milestone outcome and interaction was tested in separate models. Bonferroni post-hoc comparisons were used to minimize Type 1 error. Estimated marginal means adjusted for cohort, sexual identity, sex, race/ethnicity, and age are reported. Race/ethnicity and sexual identity dichotomized to account for variability in items response across datasets. Only contrasts with significant omnibus F values bolded. POC = participants of color. Bi + = bisexual and another sexual identity.

^*p < .05,

^**p < .01

^***p < .001.

Table 5 displays estimated marginal means from a series of multiple regression models comparing mean differences in the pacing of milestones by subgroups at the intersection of cohort with sex, sexual identity, and race/ethnicity, respectively. F-tests revealed pacing differences among subgroups at the intersection of cohort with sex for pacing from same-sex sexual behavior to self-identification [F(2, 914) = 3.75, p=.02] and same-sex sexual behavior to disclosure [F(2, 914) = 9.10, p<.01]. Male adolescents from the 2010s cohort (M = −0.85, SE =.24) reported distinct pacing from same-sex sexual behavior to self-identification relative to female adolescents from the 1990s cohort (M = 0.54, SE =.35). Regarding pacing from same-sex sexual behavior to disclosure, female adolescents from the 1990s (M = 1.01, SE =.36) differed from female adolescents from the 2000s (M = −0.45, SE =.19) and male adolescents from the 2010s cohort (M = −0.63, SE =.21). Male adolescents from the 1990s (M = 1.80, SE =.35) differed from female adolescents from the 2000s (M = −0.45, SE =.19), female adolescents from the 2010s (M = −0.15, SE =.22), and male adolescents from the 2010s (M = −0.63, SE =.21). Finally, male adolescents from the 2000s (M = 0.73, SE =.21) differed from female adolescents from the 2000s (M = −0.45, SE =.19), female adolescents from the 2010s (M = −0.15, SE =.22).

Table 5.

Marginal means from regression analyses comparing pacing of milestones at the intersection of cohort with sex, sexual identity, and race/ethnicity

	Sex to Self-Identify			Sex to Disclose			Self-Identify to Disclose			Aware to Disclose
	M	SE	F	M	SE	F	M	SE	F	M	SE	F
Cohort x Sex			3.75*			9.10***			2.53			0.03
1990s Female	0.54a	0.35		1.01ab	0.36		0.55	0.13		5.13	0.44
1990s Male	0.40	0.35		1.80cde	0.35		1.28	0.18		5.33	0.32
2000s Female	−0.58	0.17		−0.45acf	0.19		0.23	0.12		4.51	0.21
2000s Male	−0.16	0.21		0.73fg	0.21		0.90	0.13		4.77	0.20
2010s Female	−0.16	0.23		−0.15dg	0.22		0.03	0.08		3.61	0.20
2010s Male	−0.85a	0.24		−0.63be	0.21		0.23	0.15		3.77	0.25
Cohort x Sexual ID			0.00			1.45			4.52*			1.17
1990s Gay/lesbian	0.37	0.33		1.40	0.32		1.02abc	0.15		5.80	0.33
1990s Bi +	0.59	0.35		1.38	0.38		0.69d	0.16		4.44	0.44
2000s Gay/lesbian	−0.47	0.22		0.20	0.23		0.73e	0.14		5.19	0.21
2000s Bi +	−0.23	0.17		0.13	0.17		0.40a	0.11		4.09	0.19
2010s Gay/lesbian	−0.56	0.23		−0.55	0.22		−0.02bde	0.13		3.97	0.23
2010s Bi +	−0.29	0.24		0.05	0.23		0.29c	0.10		3.38	0.21
Cohort x Race/eth			0.53			0.64			2.24			2.02
1990s White	0.42	0.29		1.43	0.29		0.94	0.13		5.42	0.29
1990s Black	0.18	0.79		1.95	0.93		1.77	0.48		5.11	0.83
1990s Another race	0.29	0.46		0.88	0.50		0.63	0.20		4.50	0.75
2000s White	−0.62	0.17		0.03	0.18		0.68	0.11		4.58	0.18
2000s Black	0.26	0.25		0.36	0.31		0.19	0.16		4.78	0.29
2000s Another race	−0.24	0.43		0.45	0.34		0.70	0.26		4.45	0.35
2010s White	−0.47	0.38		−0.20	0.35		0.17	0.12		3.90	0.30
2010s Black	0.07	0.26		−0.04	0.26		0.07	0.19		2.87	0.33
2010s Another race	−0.60	0.21		−0.43	0.21		0.13	0.12		3.92	0.20

Note. The same superscripts indicate significant differences at the p<.05 level. Each milestone outcome and interaction was tested in separate models Bonferroni post-hoc comparisons were used to minimize Type 1 error. Estimated marginal means adjusted for the effects of cohort, sexual identity, sex, race/ethnicity, and age are reported. Race/ethnicity and sexual identity dichotomized to account for variability in items response across datasets. Only contrasts with significant omnibus F values bolded. Bi + = bisexual and another sexual identity.

^*p < .05,

^**p < .01

^***p < .001.

F-tests also showed significant pacing differences in subgroups at the intersection of cohort and sexual identity for pacing from self-identification to disclosure [F(2, 1253) = 4.52, p=.01] . More specifically, lesbian/gay participants from the 1990s cohort (M = 1.02, SE = .15) reported distinct time between self-identification and disclosure relative to bisexual+ youth from the 2000s cohort (M = .40, SE = .11), bisexual+ youth from the 2010s cohort (M = .29, SE = .10), and lesbian/gay youth from the 2010s cohort (M = −.02, SE = .13). Additionally, bisexual+ youth from the 1990s (M = .69, SE = .16) reported distinct pacing relate to gay/lesbian youth from the 2010s (M = −.02, SE = .13). Finally, lesbian/gay youth from the 2000s cohort (M = .73, SE = .14) reported distinct time between self-identification and disclosure compared to lesbian/gay youth from the 2010s (M = −.02, SE = .13). No differences in milestone pacing emerged at the intersection of cohort with race/ethnicity.

Sensitivity Analyses

To examine the robustness of findings across a sample with an extended age range across datasets, we ran sensitivity analyses examining subgroup differences in the age and pacing of milestones among a sample with an extended age range (i.e., participants 21 years old and younger who reported milestones at or before age 19 years old; Table S2 and Table S3). This age range was selected because it includes the age range of the 2000s and 2010s study and was the age range used by the researchers who designed the 1990s study in prior published manuscripts (D’Augelli et al., 2001; D’Augelli et al., 2002; D’Augelli et al., 2003). Results revealed analogous substantive results relative to the main analyses, barring the following exceptions: In the sensitivity models examining subgroup differences in milestone ages, female participants reported later ages of first awareness of same-sex attraction than male participants and bisexual youth did not differ from youth with another identity in ages of first same-sex sex. In the sensitivity models measuring pacing of milestones, Black participants did not differ from those of another race in pacing from same-sex sexual behavior to self-identification, all cohorts differed from one another in pacing from self-identification to disclosure, and all cohorts significantly differed from one another and youth with another sexual identity differed from lesbian/gay youth in pacing from awareness to disclosure.

We then tested whether results differed if we censored lower age milestones to age eight, in line with prior studies (Tables S4 & S5; Bishop et al., 2020); that is, among those who reported a milestone of age eight or younger, milestone ages were recoded to a value of eight. Results were analogous, barring the following differences: the 2000s cohort reported younger ages of awareness of same-sex attraction than did the 2010s cohort, and the 2000s cohort differed from the 1990s cohort in pacing from self-identification to disclosure and pacing from awareness to disclosure.

Lastly, we tested the robustness of our results when we removed improbable responses of “0 years old” from our data (Tables S6 & S7). Results substantively matched our main analyses apart from two differences: the 2000s cohort reported younger ages of awareness of same-sex attraction than did the 2010s cohort, and those with another sexual identity significantly differed from those with a lesbian/gay identity in ages of first self-realization and first disclosure. There were no substantive differences in pacing between the main and sensitivity analyses.

Discussion

Prior studies have reported declining ages of key milestones related to sexual minority people’s sexual identity development, yet such studies have largely relied on retrospective data from adults that may conflate age with cohort influences on milestone timing. The goal of the current study was to disentangle age from cohort influences by comparing milestone timing and pacing across three age-matched but sociohistorically distinct samples of sexual minority youth to test whether maturational age differences in milestones were present across cohorts. Results replicated findings from prior studies showing that the ages of first self-identification and first disclosure occur progressively earlier among more contemporary cohorts. Conversely, in the current study, sexual minority youth reported similar ages of first awareness of same-sex attraction and first same-sex sexual behavior across cohorts. Findings extend prior research by demonstrating that sociohistorical influences on milestone timing may operate differently for milestones reflecting the development of sexual minority identity, relative to the development of sexual minority awareness of attraction or sexual behavior.

Our first aim was to take advantage of three age-matched sexual minority youth cohorts to describe whether and how the timing and pacing of milestones have shifted across sociohistorically distinct periods. Results were partially aligned with our hypotheses. After adjusting for demographic covariates, no subgroup differences emerged in the ages of first awareness of same-sex attraction nor among the age of first same-sex sexual behavior across cohorts; conversely, there was cohort variability in the ages at which participants first self-identified with a sexual minority identity and first disclosed a sexual minority identity to others. Cohort differences in self-identification and disclosure echo prior studies (Bishop et al., 2020; Grov et al., 2018; Hall et al., 2021; Martos et al., 2015; Meyer et al., 2021), which have shown that more recent cohorts of sexual minority people are reporting milestones at earlier ages than prior cohorts, and lend additional credibility to the notion that sexual minority youth are self-identifying and disclosing sexual minority identities at earlier ages across historical time. In the past several decades, there has been growing visibility of sexual minority people and their lived experiences. The proportion of Americans identifying with sexual minority identities is growing rapidly as sexual minority people become increasingly visible in media and politics (Gallup Inc., 2022; Hammack & Cohler, 2011). As young people’s access to sexual minority identity expands, the ability to name oneself—for oneself and others—is increasingly accessible. In safe and supportive contexts, access to authentic and affirming identities earlier in the life course is associated with higher self-esteem and better health, whereas, in stigmatizing contexts, earlier identification could lead to more discrimination and minority stress (Russell et al., 2014). As young people self-identify with and disclose sexual minority identities at increasingly younger ages, creating safe and supportive environments is more pressing than ever.

In the current study, the average age of first awareness of attraction to same-sex people emerged around 10 −11 years old for youth across cohorts. The “magical age of 10” (Herdt & McClintock, 2000) is a well-established pubertal marker of first memorable attraction that coincides with adrenal puberty, regardless of sexual identity. As such, the current study suggests that sexual minority youth’s timing of first attraction is aligned with that of the broader population and often aligned with maturational puberty. Although some studies of retrospectively recalled milestone ages among adults found cohort variability in the age of first awareness of same-sex attraction (Drasin et al., 2008), others (Grov et. al., 2018; Martos et al., 2015) match the current study’s findings, suggesting the utility of age-matching samples when examining cohort effects on milestone timing.

Across cohorts, participants reported similar mean ages of first same-sex sexual experiences (14–15 years old). Although this age is earlier than typical sexual debut among heterosexual samples (YRBS, 2020), it is consistent with the small body of extant studies of sexual minority youth’s milestone timing (Dube & Savin-Williams, 1999; Rendina et al., 2019; Rosario et al., 1996). Conversely, this finding diverges from prior studies of milestone timing that retrospectively sample sexual minority adults (Grov et al., 2006, 2018), and thus use broader age ranges when reporting cohort differences in same-sex sexual behavior. It is possible that such studies captured age differences rather than cohort differences in the timing of first same-sex sexual behavior. More specifically, the adult samples on which prior studies have relied may be measuring well-established developmental diversity (Bishop et al., 2022; Diamond, 2006) in milestone timing in older but not younger age cohorts. Therefore, our findings relative to prior studies may have differed due to parsing age from cohort influences.

It is notable that in the current study, cohort differences emerged with respect to self-identification with a sexual minority identity and disclosing a sexual minority identity to others, but not among awareness of attraction nor among sexual behavior. These findings indicate that for sexual minority youth, milestones related to attraction and sexual behavior may be less tied to sociohistorical change, whereas those related to identity formation are fundamentally social and interpersonal, and are therefore influenced by sociocultural contexts (D’Augelli, 1994; Hammack et al., 2018). Results suggest that altering the social contexts of sexual minority youth will not suppress their same-sex attractions or behaviors; rather, less favorable social contexts may impel youth to conceal their sexual identities for longer periods of time, resulting in barriers to community connection and support (Pachankis et al., 2020). In order to reduce sexual minority-related mental health disparities, building social contexts in which youth can experience sexual identity development milestones authentically and without fear of discrimination or victimization is crucial.

With respect to pacing (i.e., time between and order) of milestones, results suggested cohort variability across milestones. The cohort patterns among self-identification, same-sex behavior, and disclosure in the current study are consistent with prior studies. For example, both in the current study and in a national probability study of sexual minority adults from three different generations (Bishop et al., 2020), the least recent cohort reported same-sex sexual behavior before self-identification on average, whereas the two most recent cohorts reported self-identification before same-sex sexual behavior. It is possible that in less recent cohorts of sexual minority youth who had less access to sexual minority visibility, sexual behavior was a strategy for understanding and validating a sexual minority identity; in more recent cohorts, increased visibility and access to language, communities, and culture related to sexual diversity may have provided a foundation for understanding a sexual minority identity that is not reliant on sexual behavior. Qualitative studies with sexual minority people are needed to further explore this potential explanation.

Youth in the 2010s cohort spent less time, on average, between first self-awareness of sexual minority attraction and first disclosure than did those in the 2000s and the 1990s cohort. The time between first same-sex attraction and first disclosure can be conceptualized as a measure of sexual identity concealment (Mohr & Fassinger, 2000; Pachankis et al., 2020; Rosario et al., 2009). Concealment is complex in relation to health and well-being, with the potential to simultaneously generate stress from inauthenticity, protection from discrimination, and loneliness from disconnection (Huang & Chan, 2022; Pachankis et al., 2020). Moreover, concealment’s meaning and health relevance may be shifting across generations. A recent meta-analysis examining relations between concealment and mental health found that concealment was more strongly associated with internalizing mental health problems in studies conducted recently relative to less recent studies (Pachankis et al., 2020). Another recent study reported that more concealment time was related to more anti-LGB structural stigma (Layland et al., 2023). Additionally, the meanings and motives for concealment, non-disclosure, and disclosure may vary by cohort. For example, the HIV/AIDS epidemic impelled many sexual minority men to disclose their identities as a form of political activism; on the other hand, some contemporary young sexual minority people reject that LGBTQ+ people must come out (Wandrey et al., 2015). Further, youth who do not disclose a sexual minority identity may not always be actively concealing (i.e., effortfully hiding one’s sexuality), but rather, may be engaging in non-disclosure (i.e., lack of integration of one’s sexuality into one’s social network; Jackson & Mohr, 2016). Taken together, additional research is needed to examine relations between time concealed with minority stress and mental health in the context of social and historical change.

Beyond cohort differences, we also examined variability in milestone timing among subgroups defined by sex, sexual identity, and race/ethnicity. Bisexual participants reported older ages of milestones than lesbian/gay participants. Further, females reported older ages of first self-identification and same-sex sexual behavior than males. No racial/ethnic differences emerged in mean ages of milestones. Bisexual and female participants’ later milestone ages are consistent with prior studies (e.g., Bishop et al., 2020; Hall et al., 2021; Martos et al., 2015). Bisexual youth often report higher levels of minority stressors such as internalized stigma, victimization, and discrimination which arise from both heterosexual and sexual minority people (Arbeit et al., 2016; Feinstein et al., 2021; Wandrey et al., 2015). As such, bisexual youth may delay acknowledging, acting on, and disclosing an identity that may expose them to additional victimization. What’s more, bisexual youth’s experiences of both heteronormative attraction and sexual minority attraction in the context of compulsory heterosexuality may cause them to attend to heterosexual attraction and question the legitimacy of their sexual minority attractions (Rich, 1980).

Concerning sex differences, prior studies have found that gender nonconformity is more strongly policed for boys than girls, including sexual minority boys (Lisdonk et al., 2015; Thoma et al., 2021). Given that gender atypicality is common in sexual minority youth (D’Augelli, et al., 2008; Ploder & Fartacek, 2009), boys’ elevated gender scrutiny may facilitate earlier acknowledgement of sexual minority identities. Finally, although some recent studies have found racial/ethnic differences in milestone timing, this finding is inconsistent (Martos et al., 2015; Grov et al., 2006). Therefore, null findings with respect to racial/ethnic variability are not surprising.

To examine the extent to which demographic differences in milestone timing were cohort-specific, the second aim of the study was to examine how milestone age and pacing varies among subgroups of sexual minority youth at the intersection of cohort with sexual identity, gender, and race/ethnicity. Few differences emerged across subgroups with respect to milestone ages, and several subgroup differences emerged regarding milestone pacing. For example, while gay/lesbian youth from both the 1990s and the 2000s spent longer between self-identification with a sexual minority identity and first disclosing an identity than gay/lesbian youth from the 2010s, sexual minority youth’s pacing did not differ by sexual identity within cohort. More broadly, differences emerged almost exclusively across cohorts rather than within cohorts. These results support the importance of sociohistorical context for understanding differences among cohorts in milestone timing and pacing and the need to integrate life course perspectives that account for sociohistorical context into studies of sexual minority identity development.

Limitations and Future Directions

Several limitations of the current study should be noted. First, given that the 2000s dataset was limited to 15–19-year-olds in wave 1, we limited our analyses to youth ages 15–19. As such, we captured sexual minority adolescents’ sexual identity developmental processes, specifically. Although most sexual minority people acknowledge sexual identities during adolescence, others develop sexual minority identities in adulthood (Berona et al., 2018; Diamond, 2016). As such, the cohort differences captured in the current study should be interpreted as specific to sexual minority adolescents rather than the broader sexual minority population. To account for cohort differences among sexual minority people who develop sexual identities across the life course, future studies should integrate a broader age range of participants. Second, we defined cohorts based on the study in which youth participated, the historical timing of which coincided with critical sexual minority-related sociohistorical events thought to represent distinct cultural contexts for sexual diversity (Weststrate & McLean, 2022). That said, other researchers and sexual minority people may consider alternative events to be equally salient or meaningful and may opt to delineate cohorts differently, particularly given that data from the 1990s data set were collected between 1995–1996 and from the 2000s dataset were collected between 1999–2001, and therefore it is possible that youth across these datasets could be considered of the same generation. Future studies could examine the reliability of findings across distinct constructions of generations, and longitudinal data are needed to best capture age-period-cohort effects without the bias of researchers’ definitions of a generation. Third, in line with prior studies that have found variability in the sequence and ordering of milestones across cohorts and social identities (e.g., Floyd & Stein, 2002), we created a pacing variable to represent the time between and order of milestones. Given that the construct represents both time between and sequence, it is complex to interpret when negative values are included. These negative values indicate important differences in the sequence of milestones, however, future studies might operationalize time between and order as distinct constructs. Relatedly, we harmonized items from three datasets that contained similar, but not identical, measures. Additional measurement work is crucial for understanding how to operationalize milestones with high reliability and validity. Fourth, differences in racial/ethnic and sexual identity response options across datasets required that we collapse several distinct demographic identities for analyses. Yet, sexual minority populations are not homogenous and future research should highlight, rather than generalize across, sexual minority youth at the intersections of social identities. In the same vein, although we examined milestone timing and pacing by subgroups at the intersections of cohort with sex, sexual identity, and race/ethnicity, respectively, intersectionality theory suggests that these social identities are mutually constitutive. As such, future studies could employ advanced statistical approaches such as MAIHDA (Merlo, 2018) within an intersectional framework to model milestone timing and pacing along multiple intersecting axes of identity, including cohort, gender, sexual identity, race/ethnicity. Fifth, to model continuous milestone outcomes, we excluded non-numeric participant responses such as “never” or “don’t know”. However, these responses are not missing at random, but rather hold specific meaning for participants. Future studies could consider accounting for such data in modeling sexual identity development processes. To address this issue, in the current study we conducted several sensitivity analyses to account for some participants’ responses with unlikely values (e.g., awareness of attraction at age 1). In these additional analyses, the 2000s cohort reported later ages of first awareness of same-sex attraction than the 2010s generation. The discrepancy between sensitivity and main analyses with respect to this finding reflects a wider inconsistency in the literature regarding whether ages of awareness are shifting across generations and varying measures of this milestone across studies (Hall et al., 2021). Further research, particularly measurement work, is needed to understand this phenomenon. Lastly, participants were largely recruited from relatively more urban and liberal areas, and as such, their milestone ages may not be typical in more rural and/or unsupportive areas. Additional studies that capture milestone timing among such understudied populations is needed moving forward.

Conclusion

Using an age-matched sample of sexual minority adolescents, the current study extended the literature examining sociohistorical variability in the maturational timing and pacing of sexual minority developmental milestones. Results suggested that sexual minority youth’s self-identification and disclosure of a sexual minority identity occur increasingly earlier in the life course across sociohistorical time. In contrast, the age at which youth are first aware of same-sex attraction and act on such attraction through sexual behaviors is less contingent on socio-historical context. As young people self-identify and disclose sexual minority identities at increasingly younger ages, they have the potential to live longer and healthier lives in their authentic identities. Conversely, in discriminatory contexts, sexual minority youth’s earlier milestones may expose them to more health-threatening minority stress during an already sensitive developmental period. Thus, it is increasingly important to provide safe and supportive environments in which sexual minority youth can develop. Policies and practices that protect, include, and celebrate sexual minority young people, such as enumerated anti-bullying and non-discrimination policies, inclusive school curricula, affirming healthcare settings, and accepting families are more important than ever.

Public Significance Statement:

In this study, we found that sexual minority adolescents were self-identifying with and disclosing minoritized sexual identities at earlier ages over sociohistorical time. These findings highlight the importance of measuring sexual identity development processes in ways that attend to potential influences of both maturational age and sociohistorical change.