Introduction

Resilience is defined as adaptation under specific risks or stress (Dohrenwend 2000; Masten 2001). The presence of a prominent stressor distinguishes resilience from typical development (Luthar 2006; Luthar et al. 2000). Rejection by family due to sexual orientation is a risk factor unique to sexual minority women (SMW) as compared to heterosexual women, but only some SMW experience family rejection (Ryan et al. 2009). During adolescence and young adulthood (YA) the majority of SMW develop a lesbian or bisexual identity, disclose that identity to family, form connections with sexual minority communities, and many develop a sense of collective, affiliation-based lesbian or bisexual self-esteem or group pride (Calzo et al. 2011; Savin-Williams and Ream 2003). As a healthy sexual minority identity is formed, identity risk factors, such as self-stigma, internalized homophobia and concealment motivations decrease. SMW who are not rejected by family experience typical sexual minority development with family support as a protective developmental asset. In contrast, SMW who are rejected experience a significant developmental risk within the family, including the loss of family to buffer against societal discrimination risks. Resilience cannot be directly observed, rather resilience processes are inferred by observations of adaptation in the context of risk. Protective–enhancing resilience is defined by enhanced resilience under conditions with greater risk, such as when families reject YA SMW daughters (Luthar et al. 2000). We infer that resilience occurs when rejected SMW develop greater community connectedness and collective self-esteem than is typical among non-rejected SMW. We conceptualize the resilience process of adaptation to family rejection within a social ecological development model.

Developmental and community psychologists identify the importance of accounting for social risks and resources at individual, family and community levels (Bronfenbrenner 1977; Sameroff and Rosenblum 2006; Sandler 2001). Overlapping microsystems of SMW and family, and SMW and community, interact to increase or decrease the influence of any single risk or protective factor. During young adulthood (ages 18-25) family influences are still strong, but new social roles in community are explored (Arnett 2000). Meeting developmental milestones during this period, such as identity development and disclosure, occur largely through social interactions and can protect against health risks (Balsam and Mohr 2007; McCarn and Fassinger 1996; Meyer 2010; Rosario et al. 2011). Developing a connection to sexual minority communities and developing a sense of esteem based on sexual minority affiliation also protect against societal discrimination and minority stress (Detrie and Lease 2007; Frost and Meyer 2012). We examined family rejection as a stressor that interacts with identity development and outness to family to predict community connectedness and collective self-esteem over time. We sought to understand interactions between individual, family and community risk and protective factors in terms of SMW strengths (Fergus and Zimmerman 2005).

Family Rejection in Young Adulthood

Minority stress theory guides research to understand the higher prevalence of mental disorder and substance use among sexual minorities as compared to heterosexual peers (Lehavot and Simoni 2011; Meyer 2003). The general focus of minority stress theory is on individual experiences and processing of societal pressures, such as stigma, prejudice and discrimination, in predicting health outcomes (Mays and Cochran 2001). Yet, most SMW do not exhibit disorder or ill health, indicating that despite discrimination, resilience is normative for the vast majority of SMW (Savin-Williams 2001). In the minority stress framework, greater stress is expected to predict greater distress and worse health and mental health outcomes. Yet, researchers are also observing patterns of resilience, in which reports of higher stress and fewer resources, are associated with lower rates of disorder (Meyer 2010). These findings are difficult to explain within the existing minority stress model, which has focused primarily on the development of psychopathology. Perhaps this is because seminal work on sexual minority stress does not operationalize the role of families during development (Meyer 2003). Because interdependent family and community social processes are not examined within a developmental framework, minority stress theory may not adequately explain normative sexual minority development and resilience processes. We believe focusing on families among YA SMW is warranted due to the role of families in healthy development, the potential of families to help combat societal risk, and empirical links between family rejection or support and the health and well-being of SMW.

Families are one of the most important social resources for support and coping during development. Family support is frequently associated longitudinally with resilience despite other risks (Luthar 2006). For example, research indicates that ethnic minority youth and young adults may learn to cope with societal discrimination through ethnic socialization practices in their families (i.e., ethnic identity information, ethnic values and behaviors) (Neblett et al. 2012; Umaña-Taylor et al. 2013). In contrast, heterosexual parents raise the majority of SMW and do not share a sexual minority identity with their daughters. SMW must also disclose sexual orientation to family, which confers the unique risk of family rejection.

On average SMW come out to family between 17 and 25 years old (Morris et al. 2001; Parks and Hughes 2007; Savin-Williams and Ream 2003). Supportive parental response to coming out has been found to predict health and well-being (Legate et al. 2012). Researchers have found that family support reduces depression and increases self-esteem (Detrie and Lease 2007; Rosario et al. 2011). Yet, many YA SMW also experience family rejection and family rejection is associated with poorer health (Bregman et al. 2012; D’augelli et al. 1998; Ryan et al. 2009). Family rejection is associated with increases in suicide, depression, substance abuse and sexual risk behavior (Ryan et al. 2009). For a minority of SMW, family rejection experiences can be severe, including verbal and/or physical attacks from family members because of sexual orientation (D’augelli et al. 1998). Therefore, family rejection may exacerbate societal risk. Researchers have found that although perceived discrimination and hate crime victimization partially accounted for psychiatric morbidity among sexual minorities, family factors explained nearly all of the additional risk (Frisell et al. 2009). In the present study, we asked participants about rejection by mothers, fathers, siblings or other family members because of lesbian or bisexual self-identification. We also asked whether families refused to talk about participants’ sexual orientation or recognize their same sex partners.

Measurement of Sexual Minority Identity Risk Factors and Outness to Family

Although in some contexts disclosure can make SMW unsafe, research also indicates that it can be healthy for SMW to develop and acknowledge their sexual minority identity with supportive others (Moradi et al. 2010; Rosario et al. 2011). Identity development includes a reduction in intrapersonal risk factors that are theoretically and empirically linked with health and well being among SMW (Balsam and Mohr 2007). In the present study, we examined six identity risk factors derived from previous SMW research: (1) concern with societal stigma, (2) motivation to conceal sexual identity, (3) identity uncertainty or confusion, (4) internalized homonegativity, (5) perceived difficulty with identity development, and (6) identity superiority, i.e., denigration of heterosexuals (Mohr and Fassinger 2000). These six risk factors comprise our operational definition of “identity risk,” due to associations between these six subscales of our identity risk measure with reduced well-being and psychological adjustment among SMW (Balsam and Mohr 2007; Rosario et al. 2011).

Outness is an aspect of sexual minority development that may or may not be related to identity formation (Mohr and Fassinger 2000; Parks and Hughes 2007). Lack of disclosure may be due to realistic expectations regarding social responses and not due to the strength of an individual’s self-identification as lesbian or bisexual. Disclosure can be adaptive or maladaptive depending on context. Researchers have found that being out in supportive social contexts was associated with less anger, depression and higher self-esteem, but not in controlling social contexts (Legate et al. 2012). In these cases, concealment of one’s sexual identity, is a form of coping with costs for the individual, such as limiting access to social support and reducing self esteem (Frable et al. 1998; Meyer 2003). High levels of identity risk factors, such as internalized negativity and concealment, likely interfere with the developmental process of forming connections with other SMW in lesbian, gay, bisexual, transgender or queer (LGBTQ) communities. For these SMW, family rejection may interact with unresolved identity risk and render inaccessible the protective benefits of community connectedness and self-esteem derived from collective identification with other sexual minorities. Conversely, among SMW who reduce self-stigma, identity confusion and other risks, family rejection may promote more social exploration outside the family and within sexual minorities communities. As internal barriers to developing connections are reduced, family rejection may prompt greater adaptive coping through community relationships among rejected SMW, than among SMW who do not experience family rejection.

Outness to family likely increases the potential for family rejection to occur. We measured outness in four family relationships (mother, father, siblings and extended family). In addition to asking about each relationship, our operational definition captured a range of outness, from open and direct communication self-identifying as lesbian or bisexual, to family knowledge about sexual orientation without conversation, to no knowledge or discussion (Mohr and Fassinger 2000). Outness to family is associated with higher levels of affiliation with other SMW (Mohr and Fassinger 2000). Therefore, we expected that as outness to family increased, community connections would also increase, particularly for SMW rejected by their families. This is expectation is consistent with a protective–enhancing resilience process.

Conceptual Framework: Protective–Enhancing Resilience in Community

Protective–enhancing resilience processes are defined by an interaction between a high-risk stressor and adaptation, such that as risk increases, resilience is amplified (Luthar et al. 2000). Using this conception, for resilient YA SMW, as sexual identity is strengthened and outness to family increases, higher family rejection should spur even greater use of community resources than is typical for SMW who are not rejected. These rejected SMW demonstrate resilience by adapting to meet their social needs in community when family support is unavailable. During YA many SMW come to identify with the LGBTQ community to experience acceptance, belonging and pride (Frost and Meyer 2012). Developing these community connections can be considered an aspect of normal development for YA SMW. But, rejected SMW may demonstrate resilience by using community resources more than SMW with family support. This process of enhanced resilience under greater family stress can be distinguished from typical SMW social development, since most SMW are not rejected by family (Ryan et al. 2009).

Minority stress theorists posit that participation in minority communities and culture confers protective benefits against societal stigma and prejudice (Detrie and Lease 2007; Frost and Meyer 2012; Meyer 2003). In fact, community connectedness predicts psychological well being among sexual minorities even when controlling for social support (Detrie and Lease 2007). In the present study, we defined the sexual minority community broadly, asking participants about their connectedness to the “LGBTQ community.” We defined community connectedness by a sense of belonging or feeling a part of the LGBTQ community through participation, bonds, pride and a sense of shared problems (Frost and Meyer 2012). We defined collective self-esteem, as deriving a personal sense of worth and value based on identification with lesbians and bisexuals as a collective group (Detrie and Lease 2007). We expected that SMW experiencing more family rejection would cope by seeking community connections to a greater extent than SMW with low or no family rejection. Through participation and bonds within the LGBTQ community, we expected that rejected YA SMW would derive greater collective self-esteem over time.

Other factors may influence adaptation to family rejection through community. YA SMW may vary in their experiences accessing the resources of sexual minority communities when needed. For example, the trajectory from family rejection experiences to community relationships may differ between lesbian- and bisexual-identified women (Kertzner et al. 2009). Some researchers have identified lower levels of disclosure and community connection among bisexuals as compared to lesbians and gays (Balsam and Mohr 2007), whereas others have found no significant differences between lesbians and bisexuals in community connectedness (Frost and Meyer 2012). These divergent research findings do not clarify whether bisexual women experience less inclusion from sexual minority communities, or whether community connectedness is valued or perceived similarly or differently by YA lesbians and bisexuals. There may be differences in timing or experience of disclosure to family between lesbians and bisexuals, or among SMW of color, as compared to White SMW. There may also be reduced in-group acceptance due to racism within sexual minority communities (Balsam et al. 2011; Moradi et al. 2010). If racism reduces community supports, research is needed to explain why SMW of color do not report more disorders than do White SMW (Meyer 2010). One possible explanation is that family rejection or support may vary based on race or ethnicity and family responses may moderate relationships between identity or outness and community connectedness. Longitudinal examinations of family rejection when YA SMW are coming out to family are relatively rare and there may also be unknown differences by age. We examined these potential variations within our large, nationally recruited SMW sample accounting for lesbian/bisexual identification, racial identity, ethnic identity and age. Determining whether or not risk and protective factors exert universal effects on SMW or whether there are conditional effects based on family reaction to identity formation and coming out will help to bridge the current gap between the minority stress and resilience/developmental literatures (Meyer 2010). Identification of any of these within-group differences also informs the SMW health disparities literature by alerting researchers to discrepancies in coping and resources likely to be associated with health and mental health outcomes (Schwartz and Meyer 2010). Given that family rejection is associated with increases in suicide, depression, substance abuse and sexual risk behavior, research is needed to help reduce these risks (Ryan et al. 2009).

The Present Study

The present investigation sought to understand SMW strengths adapting to minority stress within the family through sexual minority community resources. At present, the minority stress model underspecifies the developmental interplay of individual, family and community risk and protective factors unique to SMW. By studying developmental milestones during young adulthood we sought to yield theoretical insights regarding when and for whom connections to sexual minority communities may be particularly protective. We sought to determine whether identity formation and outness alone predict sense of community and collective self-esteem over time for all SMW. Or conversely, whether associations between individual and community protective factors are moderated by family rejection. Our hypotheses were that as SMW reduce identity risk factors and increase outness to family, SMW experiencing higher family rejection would demonstrate resilience and report greater community connectedness and collective self-esteem over and above that of non-rejected peers.

Methods

Participants

Lesbian and bisexual women (N = 843) between the ages of 18–25 (M = 21.4; SD = 2.1) completed baseline and 12-month online surveys. Thirty-one percent were from large urban areas, 26 % from medium sized cities, 29 % from smaller cities and towns, 9 % from suburban areas, and 5 % from rural areas. Participants were from New England (8.9 %), Mid-Atlantic (14.4 %), East North Central (18.7 %), West North Central (6.1 %), South Atlantic (18.4 %), East South Central (4 %), West South Central (9 %), Mountain (5.4 %) and Pacific (14.6 %) regions of the United States. In terms of racial background, the majority of the sample identified as White (54.2 %), followed by multiple backgrounds (16.6 %), African American (9.6 %) and Asian (3.1 %); 10.2 % endorsed a Hispanic or Latina ethnicity. Fifty-seven percent (n = 483) of the sample identified as bisexual.

Procedures

Self-identified lesbian and bisexual women between the ages of 18–25 were recruited primarily using online advertisements on the social networking site Facebook. Ads were shown to 1,028,700 lesbian and bisexual female Facebook users at any given time. In addition, online advertisements were placed on Craigslist in twelve cities selected for geographic diversity: Atlanta, Austin, Boston, Chicago, Houston, Los Angeles, New York, Philadelphia, San Francisco, Seattle, South Florida, and Washington DC. Recruitment occurred between September 2010 and May 2011. 4,150 SMW who endorsed their consent to participate via an online consent form were routed to a 5-min eligibility screening. Eligibility criteria included women who: (1) lived in the U.S., (2) had a valid e-mail address, (3) were between the ages of 18–25, and (4) self-identified as lesbian or bisexual. Identification as lesbian or bisexual was determined by participant responses to one self-report item: “Understanding that sexual identity can be complex, which one category best describes your sexual identity now?” Eligible participants (n = 2,109) were emailed the URL for the baseline assessment and a personal identification number (PIN) for study participation. 1,089 participants completed informed consent via an online form and completing the 45-min online survey. Participants were compensated $25 for the survey. Procedures were approved by the University of Washington.

Attrition

One year after their baseline participation, participants were contacted by email to complete a second survey, 843 participants (77.4 %) used their PIN to log-on to the 12-month survey. One-way ANOVAs were used to test for differences based on retention for the second survey. Results indicated that there were no differences between participants who were retained as compared to those who did not respond in terms of age, race, ethnicity, identity, outness to family, family rejection, community connectedness or collective self esteem. Therefore, we do not believe that our 22.6 % attrition rate introduced systematic bias.

Measures

Demographic measures included as covariates in our models are age and sexual orientation (lesbian or bisexual), and self-reported racial and ethnic background. We used the Age of Coming Out questionnaire to assess age of coming out (Parks and Hughes 2007; Rosario et al. 2011). Table 1 displays reliability coefficients (Cronbach’s alpha) for time 1 and 2 for each variable. Scale reliability in the present study was consistent with validation samples (validation alphas are listed below). All measures were selected based on validation and use with diverse samples of lesbian and bisexual women.

Table 1 Pairwise correlation matrix, means, standard deviations, and scale reliability at time 1 and time 2
Full size table

Lesbian, Gay, and Bisexual Identity Scale (validation α = .65–.81 range) (Mohr and Fassinger 2000). This 27-item self-report measure validated for use with SMW is designed to assess six identity risk factors through six subscales [stigma concerns, concealment motivation, identity uncertainty, internalized homonegativity, difficulty with the identity development and identity superiority (reverse scored)] associated with poorer health. Item response scales range from strongly disagree (1) to strongly agree (7). Responses were averaged to create the scale (scale range 1–7). Sample items include “I prefer to keep my same-sex romantic relationships rather private” and “I wish I were heterosexual.” To assess change in identity risk, time 1 scores were subtracted from time 2 so that 0 indicates relatively consistent identity. Positive values indicate an increase in overall identity risk across the six factors (change scale range −3 to 4).

Outness Inventory (validation α = .79) (Mohr and Fassinger 2000). This 11-item self-report measure validated for use with SMW is designed to measure the degree to which individuals have disclosed a sexual orientation identity to others. We used four items that ask about outness to mother, father, siblings and extended family to derive an “outness to family” measure. Item response scales ranged from “person definitely does not know about your sexual orientation status” (1) to “person definitely knows about your sexual orientation status, and it is openly talked about” (7) and responses were averaged to create the scale (scale range 1–7). Difference scores were calculated to assess change in outness to family. Time 1 scores were subtracted from time 2 so that 0 indicates relatively stable outness in relationships and positive values indicate increases in overall outness across relationships (change scale range −6 to 6).

Family rejection was measured using 6 family rejection items from the Daily Heterosexist Experiences Questionnaire (DHEQ; validation α = .79) (Balsam et al. 2012). The full DHEQ scale assesses several potential sources of sexual minority stress. We used the six items from this scale that assess family. These items include family members not accepting your partner as part of the family, your family avoiding talking about your identity, and being rejected by your mother, father, siblings or extended family for being lesbian or bisexual. Response scales to these items range from never (0) to almost every day (5). Items were summed and higher scores indicate more rejection (scale range 0–30).

The Connectedness to the LGBTQ Community Scale (eight items; validation α = .81) (Frost and Meyer 2012) has been validated with SMW and is a measure of community belonging. Example items include “I feel I am a part of the LGBTQ community” and “I feel a bond with lesbians or bisexual women.” Item response scales range from disagree strongly (1) to agree strongly (5). Items were averaged and higher scores indicate higher connectedness (range 1–5).

Collective Self-Esteem Scale (16 items; validation α = .85) (Luhtanen and Crocker 1992). This measure is used for assessing collective self-esteem based on group membership and has been used with SMW samples to assess the internalized positive salience of community across four subscales: group membership self-esteem, private collective self-esteem, public collective self-esteem and importance to identity (Detrie and Lease 2007). Items referred to the “lesbian or bisexual community” in the present study. Sample items include “In general, others respect the lesbian/bisexual community that I am a member of” and “I often feel I’m a useless member of the lesbian/bisexual community” (reversed). Item response scales range from strongly disagree (1) to strongly agree (7). Items were reversed scored as necessary and averaged. Higher scores indicate higher collective self-esteem (range 1–7).

Data Analysis

Descriptive statistics were completed for the sample and scale distributions were examined for each model construct. Pearson correlations were estimated between measures and within measures across the two time points. Data analyses were conducted using SPSS (version 21). One-way analysis of variance were conducted to assess for differences in each study variable by lesbian or bisexual identification, and based on racial or ethnic background. We used multiple linear regression for our primary analyses. We considered use of structural equation modeling (SEM), however, we had one measure for each construct of interest and felt that the use of manifest variables was reasonable since measures were validated with other diverse samples of SMW and had adequate internal consistency in the current sample. We tested regression models including interaction terms instead of a multi-group design because our family rejection moderator was a continuous variable that could not be decomposed into conceptually meaningful dichotomous groups and we did not want to underestimate the variability in our community outcomes (MacCallum et al. 2002). We tested the following hypotheses: (1) Decreased identity risk over 1 year will predict higher community connection and higher collective self-esteem. (2) Increased outness to family over 1 year will predict higher community connection and collective self-esteem. (3) Decreased identity risk over 1 year will interact with family rejection to predict higher connection to community and collective self-esteem. (4) Increased outness to family will interact with family rejection over 1 year to predict increased connection to the community and collective self-esteem. For all models, we included the baseline level of community connectedness or collective self-esteem as a covariate so that coefficients describe the change in the outcome for a 1-unit increase in the covariate. For hypotheses 3 and 4, we included interaction terms between 1-year change in identity risk or change in outness to family and baseline family rejection. We also included baseline age, sexual orientation (0 = lesbian, 1 = bisexual), race (0 = White, 1 = all other races), and ethnicity (0 = all other ethnicities, 1 = Hispanic or Latina) as covariates.

Results

Descriptive Statistics

Means of the age of coming out measure, indicated participants first wondered about their sexual identity at age 12. Participants reported a mean age of 14 or 15 years old for their first sexual experiences with men or women. On average, participants decided they identified as lesbian or bisexual at age 15 and first told someone else about their identity around age 16. Scale descriptive statistics and correlations are presented in Table 1. Scales were normally distributed and distributions were similar across racial and ethnic groups and across lesbian and bisexual participants. Analysis of variance indicated mean change in identity and change in outness to family, as well as mean family rejection and collective self-esteem were not significantly different between lesbian and bisexual participants or between racial or ethnic minority and White participants. Racial minorities reported significantly lower community connectedness than did White participants F(1,827) = 4.419, p = .04). Family rejection responses ranged from never (59–82 % across items) to everyday (10–24 % across items).

Hypothesis Tests

Table 2 displays the standardized β coefficients and t and p values from the regression models to test hypotheses 1 and 2. Demographic covariate codes and overall model statistics are displayed in the notes of Table 2. The models accounted for 38–39 % of the variance in community connectedness and 36 % of the variance in collective self-esteem; f 2 values indicate a large standardized effect size for the predictors in the models (Cohen 1988). Much of the variance in the models was accounted for by baseline levels of the outcomes; however, changes in identity risk and outness to family also showed statistically significant associations with the outcomes. Participants who reported an increase in identity risk factors reported significantly less community connectedness (β = −.09; p = .001) and lower collective self-esteem 1 year later (β = −.13; p = .001). Participants who reported an increase in outness to family reported more community connectedness (β = .06; p = .049) and higher collective self-esteem (β = .07; p = .022). In the identity risk and outness to family models, participants who endorsed a non-White racial background reported lower community connectedness, but there were no racial differences in collective self-esteem. No other covariates were significant. Residuals were normally distributed at each level of our outcomes.

Table 2 Change in identity risk and outness to family predicting community connectedness and collective self-esteem
Full size table

Table 3 displays results from regression models to test hypotheses 3 and 4. Demographic codes and overall model statistics are displayed in the notes of Table 3. Identity risk was not moderated by family rejection in predicting community connectedness. Participants who endorsed a non-White racial background reported lower community connectedness, but not collective self-esteem. No other covariates were statistically significant. However, there were significant moderating effects of family rejection in the other three models. Figure 1 shows model-predicted levels of the outcomes in measure response scales according to change in identity risk or outness depicted in standard deviation units for individuals with high and low family rejection; high refers to one SD above the mean and low refers to one SD below the mean. High family rejection moderated associations between increase in identity risk and lower collective self-esteem such that effects of change in identity risk were restricted to those with higher levels of family rejection (top figure; interaction-p = .034). Similarly, effects of change in outness to family on community connectedness (middle figure; interaction-p < .001) and collective self-esteem (bottom figure; interaction-p = .006) were restricted to women reporting higher levels of family rejection.

Table 3 Change in identity risk or outness to family interacting with family rejection to predict community connectedness and collective self-esteem
Full size table
Fig. 1
figure 1

Change in identity risk or outness to family interacting with family rejection to predict community connectedness and collective self-esteem. Notes In each figure, the dependent variable (DV) is listed on the y-axis. The independent variable (IV) is listed on the x-axis. The level of the moderator variable (MV) is identified in the legend. The high/low designations for the IV and MV represent ± 1 SD. The DV is in the Likert response scale of the measure

Full size image

Discussion

Resilience is defined by overcoming substantive adversity or unique risks (Masten 2001). Family responses to lesbian and bisexual identity development vary among YA SMW. Some YA SMW experience affirmation and support from family, while others characterize the experience of coming out to family as very difficult (Bregman et al. 2012). Reporting “never” experiencing family rejection was modal across family rejection items (range from 59–82 %). In light of previous findings that family acceptance is related to good health (e.g., self-esteem, general health) and family rejection is related to poor health (e.g., depression, suicide, substance abuse) (Ryan et al. 2009), our family findings provide one explanation for why the majority of SMW do not develop disorders despite potential risks (Savin-Williams 2001). Since many YA SMW never experience family rejection, typical sexual minority development includes family resources to cope with societal adversity rather than additional family risks. This is consistent with a large body of research that identifies family support as a protective factor (Luthar 2006), and with research on the role of families in ethnic minority resilience despite societal discrimination (Umaña-Taylor et al. 2013).

However, a significant minority of participants reported experiencing family rejection everyday (range 10–24 % across items). For these SMW, family rejection experiences likely exacerbate minority stress. Despite high levels of family rejection, coming out to family was associated with increasing community protective factors 12 months later. We infer that a protective–enhancing resilience process occurred because rejected SMW reported community connectedness that exceeded non-rejected peers (Luthar et al. 2000). This resilience process is defined by adversity that enhances resilience, which was evidenced by increasing outness and increasing community connections under higher levels of family risk. However, family rejection enhanced self-stigma, confusion, internalized negativity and other identity risk factors to lower collective self-esteem. It is not surprising that SMW with internalized negative self-evaluations for being lesbian or bisexual would also report lower esteem from identification with sexual minority communities. The minority stress model describes expectations of rejection as an intrapersonal risk that develops in response to prejudicial experiences and proposes that community connectedness buffers against these risks (Meyer 2003). Our findings suggest that there are SMW who may be more at risk for internalizing stigma when family is rejecting. Other researchers have found that a well-integrated identity was associated with family and friend support and with better psychological adjustment over time (Rosario et al. 2011). When identity confusion and family rejection combine, however, SMW may be less likely to benefit from the protective potential that sexual minority communities provide. However, family rejection did not moderate the relationship between identity risk and community connectedness. Therefore, among SMW with high identity risk factors family rejection may or may not contribute additional interference with community involvement and bonding. Instead, intrapersonal confusion, self-stigma, concealment motivation, and other identity risks may preclude participation in sexual minority communities altogether.

Supportive families should attend to YA SMW who report identity confusion or internalized negativity and encourage community participation. Experiences in sexual minority communities may help to resolve identity risks and promote healthy development (Frost and Meyer 2012). Sexual minority communities engage in a number of outreach activities, such as the GLBT National Help Center, which provides YA peer-support via instant messaging or phone and online links to local LGBTQ centers and resources. These ongoing efforts by sexual minority communities exemplify the active role communities play for many YA lesbians and bisexuals without family support who may need peer relationships, sense of belonging, political engagement and access to healthy role models to reduce internalization of prejudice and counter negative stereotypes about sexual minorities. Moreover, these resources are critical when rejected YA are experiencing acute risks such as suicide, homelessness or interpersonal violence.

As researchers continue to identify the role of family in the health and well-being of sexual minorities, more resources are being developed to aid families and practitioners. Parents and friends of lesbians and gays (PFLAG) is one long-standing community organization that can be a resource to families when daughters come out as lesbian or bisexual. In addition, the Substance Abuse and Mental Health Services Administration (SAMSHA) recently released “A Practitioner’s Resource Guide: Helping Families Support their LGBT Children,” which includes educational information for families regarding the impact of rejecting behaviors on substance abuse and other risks (Substance Abuse and Mental Health Services Administration 2014). Guidebooks for families who want to support YA SMW are also available from the Family Acceptance Project, which are available in English, Spanish and Chinese and are being tailored to specific faiths (Ryan 2009).

Although Hispanic or Latina ethnicity was not significant in any models, endorsing multiple racial backgrounds, African American, or Asian/Asian American race was associated with lower community connectedness than endorsing a white race. Developing connections with sexual minority communities may be of relatively lower importance for racial minority YA SMW. Or, racial differences in sexual minority community connectedness may be due to the relative salience of sexual minority identity as compared to racial minority identity. Finally, it is also possible that racial minority YA SMW do not find similar levels of belonging in LGBTQ communities. Each of these explanations are possible based on research conducted to date (Balsam et al. 2011; Meyer 2010; Moradi et al. 2010). However, since race was not a significant covariate in any of the collective self-esteem models, racial minority SMW may also generalize strengths learned to combat racial minority discrimination and apply them to their sexual minority identity (Neblett et al. 2012). This interpretation is somewhat consistent with studies that have examined internalized homonegativity and found no racial/ethnic differences (Rosario et al. 2004).

Limitations

Limitations of the present study include the potential for unknown sampling biases or errors associated with recruitment via social network websites and online data collection. Researchers are finding that YA SMW enlist online communities to make virtual and in person connections (DeHaan et al. 2013). Therefore, these findings will likely generalize to many SMW. However, since many participants self-identified as interested in women on their Facebook profiles our findings may be less likely to generalize to YA SMW who do not self-identify in online communities. Online communities are an additional place where women, especially those in rural or more isolated areas, may be able to engage with the community and receive social support. Among our participants, 9 % were from suburban areas, and 5 % from rural areas. These data are also limited to self-reports from individual SMW. Reports from both supportive and rejecting family members could improve understanding of the barriers and facilitators that impact effective family supports. That said, recruitment of rejecting family members could prove practically challenging and may be undesirable for some SMW. In addition, more precise geographic location measures would be useful to assess for differences in access to sexual minority communities when needed. For example, SMW in rural areas may not have the same density of sexual minority community members with whom to bond (Poon and Saewyc 2009; Swank et al. 2013).

More work is needed to determine whether these associations continue after young adulthood or abate over time. Our findings indicate that 1 year was an adequate interval for identifying prospective developmental relationships. Use of change scores and including baseline covariates of our dependent variables helped to identify changes that likely would have been obscured using a cross-sectional design and retrospective reports. However, prospective research like the present study cannot establish causal inferences. Future longitudinal research should seek to establish developmental intervals with greater specificity. Many of these participants began coming out prior to age 18 indicating that examinations of family responses at earlier ages would also be valuable. Alternative analytic approaches to estimate directional relationships, such as path analysis, or to assess associations without measurement error, such as SEM may identify greater developmental specificity in the observed longitudinal relationships. Replication is needed to clarify whether these developmental risks and resources during young adulthood impact health behaviors, social relationships and health disparities later in life.

Conclusions

The present investigation has several strengths, including a large sample of lesbian and bisexual women largely representative of US geographic regions and racial/ethnic backgrounds. Social ecological development studies accounting for YA SMW in their families and communities should continue, since family factors can alter relationships between individual and community protective factors and may confound relationships between societal discrimination and health risks (Frisell et al. 2009). We found that most families do not reject their lesbian or bisexual daughters. In addition, rejected SMW who increased outness to family also reported more community connectedness and collective self-esteem. Identifying this resilient response to family risk is a valuable addition to the relatively small research literature on SMW strengths (Savin-Williams 2001). YA SMW who do not resolve internalized negativity and who conceal their sexual orientation may be at the greatest risk of poor mental and physical health, particularly in highly rejecting families. The present investigation improves understanding of interdependent developmental factors related to health and well-being among YA SMW, an understudied and at-risk population (Boehmer 2002; Coulter et al. 2014).