Sexual Identity Group Differences in Child Abuse and Neglect

Abstract

Research suggests that sexual minority women are more likely than heterosexual women to report childhood abuse, but little is known about potential within-group variations in experiences of abuse among sexual minority women. We investigated rates and characteristics of childhood sexual and physical abuse among women from five sexual identity groups. Our analyses used a pooled sample of women from a national probability study and a large community-based study of sexual minority women designed to replicate the national study’s methodology (pooled n = 953). As predicted, heterosexual women reported significantly less childhood abuse than did women who identified as mostly heterosexual, bisexual, mostly lesbian, or lesbian. There was also considerable variability in abuse rates and characteristics, including severity of abuse, among sexual minority subgroups. To the extent that differences in reports reflect the actual prevalence and severity of abuse experiences, sexual identity subgroup differences in childhood abuse have important clinical and public health implications.

Characteristics of Childhood Sexual and Physical Abuse Among Women of Diverse Sexual Identities

Rates of childhood sexual abuse (CSA) in sexual minority women are alarmingly high. Most community-based studies of sexual minority women report CSA rates from 30% to 60% (Hughes, Haas, Razzano, Cassidy, & Matthews, 2000; Matthews, Hughes, Johnson, Razzano, & Cassidy, 2002; Morris & Balsam, 2003; Wilsnack et al., 2008). Studies that include heterosexual comparison groups almost uniformly have found that sexual minority women are more likely than heterosexual women to report CSA (Austin et al., 2008; Balsam, Rothblum, & Beauchaine, 2005; Hughes, et al., 2000; Hughes, Johnson, & Wilsnack, 2001; Matthews, et al., 2002). For instance, in a study with a community sample of women, Hughes and colleagues (2001) found that lesbian women were more likely to report self-perceived CSA than were a demographically matched group of heterosexual women and were more likely to report childhood sexual experiences that met Wyatt’s (1985) criteria for CSA. These criteria include sexual activity before age 18 that was unwanted, that was with a family member five or more years older than the participant, or that occurred before age 13 with a non-family member five or more years older than the participant. More recently, with a pooled national and community-based sample, our research team found that women who identified as exclusively lesbian were significantly more likely to report CSA (59%) than were women who identified as exclusively heterosexual (29%) (reference blinded).

Researchers using large national samples have also found higher rates of CSA among sexual minority women than among heterosexual women. Austin and colleagues (2008) found that lesbian and bisexual women in the Nurses’ Health Study II were more than two times as likely as heterosexual women to report CSA. Similarly, using data from the National Epidemiologic Survey of Alcohol and Related Conditions (NESARC), Hughes, McCabe, Wilsnack, West, and Boyd (2010) found that lesbian and bisexual women were three times as likely as heterosexual women to report CSA.

A smaller yet growing body of research on childhood physical abuse (CPA) among sexual minority women also suggests elevated rates of CPA compared to heterosexual women (Austin, Jun, et al., 2008; Balsam, et al., 2005; Corliss, Cochran, & Mays, 2002; Saewyc et al., 2006; Stoddard, Dibble, & Fineman, 2009; Tjaden, Thoennes, & Allison, 1999). Using the National Survey of Midlife Development in the United States (MIDUS; 1996), Corliss, Cochran, and Mays (2002) found that a combined sample of lesbian and bisexual women were more likely to report major physical maltreatment by their parents or guardians than were heterosexual women. Major maltreatment included such incidents as being “kicked, bit, or hit … with a fist,” in contrast to minor physical maltreatment, such as being “pushed grabbed, or shoved….” Similarly, Tjaden, Thoennes, and Allison (1999) found that adults who lived with a same-sex partner, a proxy measure for sexual minority status, were more likely than participants who were opposite-sex married or who lived with opposite-sex partners to report being physically assaulted as a child by an adult caretaker.

Balsam, Rothblum, and Beauchaine (2005) examined childhood abuse in lesbian, gay, and bisexual (LGB) participants by comparing them with their heterosexual siblings. Balsam and colleagues found that sexual minority status was significantly associated with each of the childhood abuse measures, including sexual and physical abuse by parents or caretakers, with LGB participants reporting greater abuse. Notably, sexual orientation accounted for significant variance in victimization beyond the effect of family, suggesting that even within the same household, LGB children were at greater risk of abuse than were their heterosexual siblings. Stoddard, Dibble, and Fineman (2009) used a similar sibling-matched design and found discrepancies in CPA and CSA between lesbians and their heterosexual sisters. Whereas 20.4% of lesbians reported that they had been physically abused as children (before age 16), only 10.0% of their heterosexual sisters did. Similarly, 26.6% of lesbian participants reported sexual abuse as children, compared to 15.7% of heterosexual participants.

CSA and CPA Severity and Characteristics

Not only do prevalence rates of childhood abuse differ, but research findings are beginning to suggest that sexual minority women experience more severe forms of CSA and CPA (Austin, Jun, et al., 2008; Balsam, et al., 2005; reference blinded). In Austin and colleagues’ (2008) study of women from the Nurses’ Health Study II, bisexual and lesbian women reported greater chronicity of physical abuse in both childhood (before age 11) and adolescence (ages 11–17) than did heterosexual women. Lesbian women who reported physical abuse in childhood were also more likely than heterosexual women to experience physical revictimization in adolescence. Comparing CSA experiences, bisexual and lesbian women were more likely than heterosexual women to report both unwanted sexual touching and forced sexual activity during childhood, but the ratio of sexual minority women’s rates to heterosexual women’s rates was larger for forced, and presumably more severe, sexual experiences. Of the women who reported sexual abuse as children, lesbian and bisexual participants were more likely than heterosexual women to report sexual revictimization.

In our research team’s analysis of pooled community-based and national data, lesbian participants were more likely than heterosexual participants to report perpetrator use of physical force or coercion; genital CSA; vaginal or anal intercourse; younger age at first CSA; greater frequency and duration of CSA; and greater number of perpetrators (reference blinded). These findings are of concern, especially given that greater CSA severity is related to a greater likelihood of negative mental health outcomes (Bulik, Prescott, & Kendler, 2001; Kendler et al., 2000; Zink, Klesges, Stevens, & Decker, 2009).

Variations Across Sexual Minority Subgroups

There is growing recognition among researchers that sexual minority women are a heterogeneous group and that researchers should examine similarities and differences within sexual minority populations. In support of this, research has found significant differences in health risk behaviors among subgroups of sexual minority women (Hughes, Szalacha, & McNair, 2010; McCabe, Hughes, Bostwick, West, & Boyd, 2009; McNair, Kavanagh, Agius, & Tong, 2005). For example, in a study of substance use in a national probability sample of Australian women, different profiles of substance use emerged for the four sexual identity subgroups (Hughes, Szalacha, & McNair, 2010). Although all three sexual minority subgroups reported greater substance use compared to exclusively heterosexual women, women who identified as mainly heterosexual and bisexual were at greater risk of substance use on more outcome measures than were lesbian women. Similarly, in an earlier study of women’s mental health, also using the Australian Longitudinal Women’s Health Survey, McNair and colleagues (2005) found significant variation in mental health outcomes among sexual minority subgroups. Compared to exclusively heterosexual women, bisexual women were more likely to report self-harm, and mainly heterosexual women had poorer mental health outcomes on all dimensions assessed except medically diagnosed anxiety.

Examining the relationship between sexual victimization and hazardous drinking, Hughes, Szalacha, Johnson, and colleagues (2010) found several differences among sexual minority subgroups. Bisexual women had higher rates of both lifetime sexual victimization and hazardous drinking than exclusively lesbian, mostly lesbian, mostly heterosexual, and exclusively heterosexual women. Relationships between sexual revictimization and hazardous drinking were strongest among mostly lesbian and mostly heterosexual women, suggesting subgroup differences not only in outcome rates but also in the extent to which variables are related.

Apart from Hughes, Szalacha and colleagues (2010) study, relatively few studies have compared rates of CSA across subgroups of sexual minority women, and in these studies findings have been mixed. Saewyc and colleagues (2006) analyzed data from seven high school-based surveys and found some sexual minority subgroup differences in sexual abuse. In all but one of the surveys analyzed, lesbian girls had a higher prevalence of sexual abuse than heterosexual girls. Most of the seven surveys analyzed did not demonstrate significant differences between lesbian and bisexual girls, although in one survey lesbian girls had higher odds of sexual abuse. Two surveys included assessments that permitted comparisons of bisexual and mostly heterosexual-identified girls; in one of these there were no differences and in the other bisexual girls were more likely to report childhood sexual abuse. Wilsnack and colleagues (2008) found that exclusively heterosexual women were significantly less likely to report CSA than any of the four sexual minority groups. A higher percentage of bisexual women reported CSA (73.8%) than the other sexual minority groups (only lesbian = 59.1%; mostly lesbian = 57.9%; mostly heterosexual = 41.9%), but due to the small number of bisexual women in the study (wt n = 16), differences among sexual minority groups were not statistically significant. Examining differences in victimization between mostly heterosexual and exclusively heterosexual predominately Black and Latina young women, Austin and colleagues (2008) found that young women who identified as mostly heterosexual were substantially more likely that those who identified as exclusively heterosexual to report childhood sexual abuse (45% vs. 15%). In Balsam and colleagues (2005) study of LGB adults and their heterosexual siblings, no significant differences in rates of CSA were found for lesbian and bisexual women.

Even less research has explored subgroup differences in childhood physical abuse. Saewyc and colleagues (2006) found that in all four high school surveys that included information on physical abuse, bisexual and lesbian girls reported greater physical abuse than heterosexual girls. In one survey, bisexual and lesbian girls reported more physical abuse than mostly heterosexual girls. Additionally, although two of the four surveys found lower odds of physical abuse among bisexual girls compared with their lesbian peers, there were no differences between bisexual and lesbian youth in the other two.

Currently, there is not enough information about sexual identity subgroup differences in CSA and CPA prevalence or characteristics to draw firm conclusions. Most studies of CSA and CPA have not assessed sexual subgroup identity or have needed to combine subgroups to increase statistical power. Given the dearth of information and the growing acknowledgment of the importance of subgroup analyses (Austin, Jun, et al., 2008; Hughes, Szalacha, & McNair, 2010), further research on sexual minority subgroup differences is indicated. Such information will permit more accurate descriptions of the childhood abuse experiences of sexual minority women, and may increase understanding of the factors contributing to sexual orientation group differences in abuse. Information on sexual minority subgroup differences would also be useful in planning targeted intervention strategies.

The present study extends our earlier research comparing rates and severity of CSA in exclusively lesbian and exclusively heterosexual women (reference blinded). Like this previous research, the analyses presented here used a pooled data set from two large studies: the 2001 National Study of Health and Life Experiences of Women (NSHLEW) and the 2001 Chicago Life Experiences of Women (CHLEW) study. The NSHLEW is 20-year longitudinal study of drinking among women in the U.S. general population. The CHLEW is a 10-year longitudinal study that replicated and extended the NSHLEW with a diverse sample of self-identified sexual minority women in the greater Chicago metropolitan area. Combining data from these studies provides a unique opportunity to compare rates and characteristics of CSA and CPA across sexual identity subgroups and to identify subgroups at greatest risk.

Going beyond the previous comparisons of exclusively lesbian and exclusively heterosexual women, we compared CSA and CPA prevalence in women of five sexual identity groups: exclusively heterosexual, mostly heterosexual, bisexual, mostly lesbian, and exclusively lesbian. We also examined and compared characteristics of CSA and CPA across the five sexual identity groups. Based on previous research we expected that, compared with exclusively heterosexual women, rates of CSA and CPA would be higher among all four sexual minority subgroups. Given the relative dearth of information about subgroup differences in prevalence and characteristics of childhood abuse, our remaining analyses were exploratory.

Method

Study Samples

National Study of Health and Life Experiences of Women (NSHLEW)

The NSHLEW was designed to assess predictors and consequences of hazardous drinking in a nationally representative sample of more than 1200 adult women. Data were collected every five years in 1981, 1986, 1991, 1996, and 2001. Additional subsamples of women aged 21–30 were added in 1991 and 2001 (references blinded). The data were statistically weighted to adjust for variations in non-response rates by sampling unit and by age, ethnicity, education, and marital status.

In the NSHLEW, 75-minute face-to-face interviews were conducted by female interviewers in participants’ homes or other private settings. For consistency with methods used in previous NSHLEW surveys, participants of the 2001 survey provided verbal consent to participate. All procedures were reviewed and approved by the University of North Dakota Institutional Review Board.

Chicago Health and Life Experiences of Women (CHLEW)

The CHLEW replicated and extended the NSHLEW by examining risk and protective factors for hazardous drinking among a sample of 447 sexual minority women, aged 18 or older. Other eligibility criteria included in the initial telephone screening included self-identity as lesbian and residence in the greater Chicago metropolitan area. Participants were recruited via local newspapers advertisements, Internet listservs, and flyers posted in churches and bookstores. Recruitment staff distributed flyers and encouraged participation among individuals and organizations via formal and informal social events and social networks. Study participants were encouraged to pass along information about the study to recruit their personal contacts. Efforts were made to maximize subgroups typically underrepresented in studies of lesbian health, including women under age 25 and over 50, those with a high school education or less, and women from racial/ethnic minority backgrounds. Interested women were invited to call the project office to determine if they were eligible to participate in the study.

The CHLEW study closely replicated the interview protocol used in the NSHLEW. Face-to-face interviews were conducted in the participants’ homes or other private settings by trained female interviewers. The interviews averaged 90 minutes in length. As in the national study, questions about potentially sensitive topics such as physical and sexual abuse were asked toward the end of the interview, when rapport was well-established. Women in the CHLEW provided written consent to participate. All procedures were reviewed and approved by the University of Illinois at Chicago Institutional Review Board.

Pooled NSHLEW and CHLEW sample

To maximize the comparability of CHLEW and NSHLEW participants we selected cases for the pooled sample based on age and residence. Because nearly all CHLEW participants were from Chicago or the surrounding suburbs, we limited the pooled sample to participants who reported living in large or medium-sized cities or suburbs near a large city. As a result, 14 CHLEW participants and 544 NSHLEW participants were excluded from the pooled sample. To match the samples on age, we excluded 25 CHLEW participants aged 18–20, one CHLEW participant aged 83, and 30 NSHLEW participants aged 71 and older, resulting in age ranges of 21–70 for both samples.

Sexual identity subgroups

In both the CHLEW and the NSHLEW, sexual identity was defined using responses to the question: “Recognizing that sexuality is only one part of your identity, how do you define your sexual identity?” Response options were exclusively lesbian, mostly lesbian, bisexual, mostly heterosexual, and exclusively heterosexual. Although women who identified as bisexual (or heterosexual or transgender) during screening for eligibility were excluded from the CHLEW study, in the actual interviews 11 women identified as bisexual and were retained for our analyses.

The final pooled sample included 953 participants. Of these, 482 (50.6%) identified as exclusively heterosexual; 42 (4.4%) identified as mostly heterosexual; 22 (2.3%) as bisexual; 111 (11.6%) as mostly lesbian; and 296 (31.1%) as exclusively lesbian. The weighted n’s and percentages for the analytic sample were exclusively heterosexual: n = 502 (52.7%); mostly heterosexual: n = 32 (3.4%); bisexual: n = 16 (1.7%); mostly lesbian: n = 100 (10.5%); and exclusively lesbian: n = 303 (31.8%).

Measures

The NSHLEW and CHLEW interview questionnaires included detailed questions about drinking patterns, drinking contexts, and drinking-related problems. Other questions asked about a wide range of possible antecedents, correlates, and consequences of drinking, such as childhood experiences (including CSA and CPA), adult employment and social roles, sexual experience, characteristics of close personal relationships, experiences of sexual and physical victimization in adulthood, and physical and mental health. Questions, indexes, and scales were originally selected, wherever possible, from well-validated instruments. The interview questionnaire was pretested prior to each wave of data collection and refined over time to retain variables with the greatest predictive value. For additional information about the NSHLEW questionnaire see [references blinded].

Some of the questions in the NSHLEW survey were slightly modified for use in the CHLEW study to be more inclusive of lesbians’ experiences (Skrocki, 1996). For example, the term ‘husband’ or ‘spouse’ was changed to ‘partner.’ In addition, sexual-minority-specific questions, such as those related to sexual identity development and disclosure, were added to the CHLEW questionnaire. These questions were developed in two focus groups with Chicago-area lesbians. In 1996, the sexual orientation questions and wording changes were added to the NSHLEW.

CSA prevalence and type

We assessed rates and types of CSA using questions about experiences of eight sexual activities before age 18 (ranging from exposure and exhibitionism to anal or vaginal penetration). To avoid suggesting sexual abuse, terms such as “abuse” or “sexual abuse” were not used in any of these questions. Using Wyatt’s (1985) criteria, CSA was defined as (1) any intrafamilial sexual activity before age 18 that was unwanted by the participant, or that involved a family member 5 or more years older than the participant; or (2) any extrafamilial sexual activity that occurred before age 18 and was unwanted, or that occurred before age 13 and involved another person five or more years older than the participant. Additional information about CSA measures, prevalence rates, and correlates can be found in [reference blinded] and [reference blinded].

CSA self-perception and characteristics

All participants were asked whether they felt that they had been sexually abused in childhood (‘yes’ or ‘no’). We analyzed data from participants who met Wyatt’s criteria for CSA to create a measure of perceived CSA. Participants were asked the age of onset of all sexual experiences prior to age 18. We used the youngest age of onset of experiences that met Wyatt’s criteria for CSA as our age of onset measure. Participants who met Wyatt’s criteria for CSA were also asked follow-up questions to determine whether they viewed the CSA experience(s) as distressing or difficult to deal with; if so, they were asked additional questions about the most distressing experience (number of perpetrators, frequency, duration, use of physical force, and physical consequences such as pain, cuts/bruises or broken bones).

CPA prevalence

Participants were asked whether, when they were growing up, they were physically hurt by their parents or other family members, including aunts, uncles, grandparents, siblings, and cousins. Response options were “never”, “rarely”, “sometimes”, “often”, and “very often”. All participants who said that they had ever been physically hurt (‘rarely’ to ‘very often’) on this filter question were asked the follow-up question, “Do you feel that you were physically abused by your parents or other family members when you were growing up?” Participants’ yes/no responses to this question served as our measure of self-perceived CPA. We do not report statistical comparisons of responses to the filter question itself because the phrasing of this question may have unintentionally captured accidental injury by family members.

Characteristics of CPA

To further assess potential CPA, all participants (not just those who reported that they perceived themselves to have been physically abused) were asked how their parents disciplined them when they were growing up. The question stem stated “People have different views about ways of disciplining children. Thinking back to when you were about 10 years old, what were your parents’ usual methods of disciplining you?” Participants were asked about eight different forms of discipline: (1) explained why something was wrong; (2) put in ‘time-out’ or sent to room; (3) took away privileges or grounded; (4) neglected my basic needs (food, clothing, shelter, love); (5) shouted, yelled or screamed; (6) spanked with bare hand; (7) spanked or hit with belt, switch, or other object; and (8) beat up, punched, choked, or threw me down. For the present study, we restricted our analyses to the four forms of discipline (items 4 and 6–8) that have the greatest potential to result in physical or psychological harm.

Data Analysis

The NSHLEW sample had been previously weighted to reflect the selection probabilities (e.g., the over-sampling of heavier drinking women; see [reference blinded], for details). For the analyses reported here, we weighted the CHLEW sample to reflect the age and education structure of the NSHLEW sample. We created dichotomous variables for all of the CSA and CPA measures. Cross-tabulation tables were produced using SAS 9.2 (2008) to assess differences in the proportions of CSA and CPA characteristics across the five sexual identity groups. We used Rao-Scott Chi-Square analyses to adjust for cluster sampling.

We computed post hoc comparisons by conducting Chi-square analyses on pairwise comparisons for all five sexual minority groups on each of the CSA and CPA characteristics. Given the large number of post-hoc analyses, we used a modified Bonferroni correction (Holm, 1979) to adjust significance levels so as not to violate an overall alpha-level of .05 for the family of post hoc tests (ten pairwise comparisons for each CSA and CPA characteristic).

Results

In Table 1 we report the rates and characteristics of CSA for the five sexual identity subgroups. We identify subgroup similarities and differences with Arabic superscripts. Exclusively heterosexual women were less likely to report CSA than were women in the four sexual minority groups. Post-hoc analyses revealed that exclusively heterosexual and mostly heterosexual women did not differ significantly in reported rates of CSA, but exclusively heterosexual women were significantly less likely to report CSA than bisexual (χ² = 13.79, df = 1, p = .0002), mostly lesbian (χ² = 28.28, df = 1, p < .0001), and exclusively lesbian ( χ² = 66.05, df = 1, p <.0001) women.

Table 1.

Self-Reported Prevalence and Characteristics of Childhood Sexual Abuse (CSA) among Women of Five Sexual Identity Subgroups

Overall CSA Prevalence and Type (unweighted N=953)	Exclusively Heterosexual Wt. n = 470.89 % (SE)	Mostly Heterosexual Wt. n = 29.69 % (SE)	Bisexual Wt. n = 15.45 % (SE)	Mostly Lesbian Wt. n = 100.30 % (SE)	Exclusively Lesbian Wt. n = 303.30 % (SE)	Chi-Square
Any CSA	28.80a (2.68)	41.94a,b (8.62)	73.76b (10.73)	57.86b (4.26)	59.13b (2.99)	91.09***
Genital contact	13.68a (2.15)	22.20a,b (7.54)	36.48a,b (7.11)	37.29b (6.15)	38.34b (1.79)	75.60***
Vaginal and/or anal penetration	5.40a (1.29)	9.03a,b (5.01)	11.86a,b (7.29)	16.62b (6.22)	14.69b (0.98)	17.70***
CSA Self-Perception and Characteristics (unweighted N=392)	Exclusively Heterosexual Wt. n = 135.55 % (SE)	Mostly Heterosexual Wt. n = 12.20 % (SE)	Bisexual Wt. n = 10.94 % (SE)	Mostly Lesbian Wt. n = 51.20 % (SE)	Exclusively Lesbian Wt. n = 160.59 % (SE)	Chi-Square
Perceived being sexually abused as child	36.96a (6.19)	46.05a,b (19.34)	67.93a,b (17.06)	53.07a,b (1.01)	68.37b (7.87)	14.50**
Onset 8 years old or younger	34.12a (4.97)	27.60a,b (10.01)	34.54a,b (12.88)	54.87a,b (4.97)	57.32b (3.96)	33.60***
Most Distressing CSA Experience
Two or more perpetrators	5.84a (2.03)	4.91a,b (4.94)	31.95b (6.06)	33.21b (0.46)	27.22b (3.17)	54.06***
Occurred more than once	50.23a (6.21)	87.35a,b (7.66)	74.20a,b (18.94)	76.25b (7.30)	75.38b (2.45)	18.75***
Duration of 1 year or more	35.19a (5.55)	47.50a,b (13.75)	67.65a,b (19.28)	63.49b (5.85)	57.07b (0.93)	19.61***
Physical force used	44.65 (5.66)	39.41 (13.67)	60.06 (21.44)	45.95 (17.64)	50.87 (0.88)	1.02
Physical consequences (pain, cuts, bruises, or broken bones)	28.71 (4.53)	24.77 (10.88)	29.11 (19.21)	42.13 (8.71)	36.03 (4.02)	2.60

Notes: *p<.05;

^**p<.01;

^***p<.001.

Arabic superscripts indicate that the percentages with different superscripts by row are significantly different from one another.

Regarding types of CSA, sexual minority women as a group were more likely to report genital contact than were their exclusively heterosexual counterparts. Post-hoc analyses showed that exclusively heterosexual, mostly heterosexual, and bisexual women reported similar rates of genital CSA and vaginal/anal penetration, but mostly lesbian and exclusively lesbian women reported significantly higher rates of genital CSA (mostly lesbian: χ² = 28.29, df = 1, p < .0001; exclusively lesbian: χ² = 60.06, df = 1, p <.0001) and vaginal/anal penetration (mostly lesbian: χ² = 14.42, df = 1, p = .0001; exclusively lesbian: χ² = 18.93, df = 1, p <.0001) than did exclusively heterosexual women.

Sexual minority women were also more likely than exclusively heterosexual women to report that they perceive themselves to have been sexually abused in childhood. Post hoc analyses revealed that exclusively lesbian women were more likely to report self-perceived CSA than were exclusively heterosexual women (χ² = 26.07, df = 1, p < .0001).

Sexual minority women reported higher rates of indicators of more severe CSA than did exclusively heterosexual women. There were significant differences in age of onset of CSA by sexual identity group; exclusively lesbian women were more likely to report early age of onset of CSA than were exclusively heterosexual women (χ² = 15.82, df = 1, p < .0001). For the most distressing event, we found statistically significant group differences for number of perpetrators (two or more vs. one), frequency (more than once vs. once), and duration (one year or more vs. less than one year). Post-hoc analyses of number of perpetrators, frequency, and duration of CSA showed that mostly lesbian (χ² = 20.50, df = 1, p < .0001; χ² = 8.12, df = 1, p = .004; χ² = 9.62, df = 1, p = .002, respectively) and exclusively lesbian (χ² = 21.00, df = 1, p < .0001; χ² = 17.52, df = 1, p < .0001; χ² = 12.16, df = 1, p = .0005, respectively) women reported higher rates of these indicators of more severe CSA than did exclusively heterosexual women. Bisexual women were also more likely than heterosexual women to report that their most distressing CSA experience involved two or more perpetrators (χ² = 9.20, df = 1, p < .005). Differences in experiences of physical force or physical injury associated with the CSA event did not differ significantly across the sexual identity groups.

Table 2 summarizes comparisons of the rates and characteristics of CPA across the five sexual identity groups. We found significant sexual identity group differences in reports of self-perception of being physically abused by a family member, having one’s basic needs neglected, being spanked/hit with a belt/object, and being beat/punched/choked. We did not find significant group differences in reports of being spanked with a bare hand.

Table 2.

Self-Reported Prevalence and Characteristics of Childhood Physical Abuse (CPA) among Women of Five Sexual Identity Subgroups

	Exclusively Heterosexual Wt. n =469.79 % (SE)	Mostly Heterosexual Wt. n = 29.69 % (SE)	Bisexual Wt. n = 15.45 % (SE)	Mostly Lesbian Wt. n = 100.30 % (SE)	Exclusively Lesbian Wt. n = 303.30 % (SE)	Chi- Square
Perceived being physically abused as child	8.65a (1.88)	16.81a,b (6.08)	44.98b (16.52)	24.48b (3.74)	26.40b (2.19)	84.42***
CPA Characteristics
Neglected basic needs	1.73a (0.69)	3.38a,b (2.42)	35.74c (12.45)	8.11b (5.59)	14.79b,c (1.32)	44.02***
Spanked with bare hand	37.23 (3.39)	35.29 (9.06)	39.07 (13.36)	51.55 (3.75)	41.97 (3.24)	9.58
Spanked or hit with belt/object	34.69a (3.17)	20.93a (7.03)	58.28a,b (14.34)	54.64b (5.24)	61.91b (2.31)	119.16***
Beat, punched, choked	2.49a (0.92)	3.39a,b (2.33)	23.97b (11.68)	11.23b (6.95)	10.91b (0.33)	17.59*

Notes: *p<.05;

^**p<.01;

^***p<.001.

Arabic superscripts indicate that the percentages with different superscripts by row are significantly different from one another.

Post-hoc results indicated that exclusively heterosexual and mostly heterosexual women were not statistically different on any of the CPA measures. Bisexual, mostly lesbian, and exclusively lesbian women were more likely to perceive being physically abused as children than were exclusively heterosexual women (bisexual: χ² = 22.33, df = 1, p < .0001; mostly lesbian: χ² = 20.46, df = 1, p = .0001; exclusively lesbian: χ² = 44.02, df = 1, p <.0001). Bisexual, mostly lesbian, and exclusively lesbian women were also more likely than exclusively heterosexual women to report that their basic needs were neglected during childhood (bisexual: χ² = 62.29, df = 1, p < .0001; mostly lesbian: χ² = 12.16, df = 1, p = .0005; exclusively lesbian: χ² = 49.18, df = 1, p <.0001), and were more likely to report that they were beaten/choked/punched as children (bisexual: χ² = 22.03, df = 1, p < .0001; mostly lesbian: χ² = 16.36, df = 1, p < .0001; exclusively lesbian: χ² = 23.92, df = 1, p <.0001). Mostly lesbian and exclusively lesbian women were more likely to report being spanked/hit with a belt than were mostly heterosexual women (χ² = 10.45, df = 1, p = .001; χ² = 18.65, df = 1, p < .0001, respectively) and exclusively heterosexual women (χ² = 13.93, df = 1, p = .0002; χ² = 54.75, df = 1, p < .0001, respectively). Bisexual women were more likely to report that their basic needs were neglected than were mostly heterosexual (χ² = 8.55, df = 1, p = .003) and mostly lesbian (χ² = 9.65, df = 1, p = .002) women.

Discussion

This paper contributes to a small but growing literature on sexual identity disparities in violence and victimization. Consistent with previous studies, we found that exclusively heterosexual women were less likely than women who identified as bisexual, mostly lesbian, and exclusively lesbian to report childhood sexual and childhood physical abuse. Additionally, these three groups of women reported more severe abuse experiences. There was, however, considerable variability with regard to CSA and CPA rates and characteristics among the sexual minority subgroups.

Regarding CSA, exclusively lesbian and mostly lesbian women reported significantly higher rates of CSA and more severe CSA experiences than did exclusively heterosexual women. Mostly heterosexual women’s CSA rates and severity generally fell between rates for exclusively heterosexual women and those of the bisexual, mostly lesbian and exclusively lesbian subgroups, and did not differ significantly from either exclusively heterosexual women or from other sexual minority women. Bisexual women had higher rates of self-perceived CSA and were more likely to report a greater number of perpetrators during their most distressing event compared to exclusively heterosexual women. However, they did not have significantly higher rates of other CSA severity indicators. Based on the raw percentages, and considering the small bisexual sample size, low statistical power may account for some of the non-significant findings. In terms of overall CSA prevalence, duration of CSA, and use of physical force, bisexual women reported the highest rates of any of the five sexual identity groups, but again, these differences were not statistically significant.

Similarly, exclusively lesbian, mostly lesbian, and bisexual women were more likely to report that they were physically abused as children and to report more severe CPA characteristics than were exclusively heterosexual women. As with CSA, rates and characteristics of CPA were similar for mostly heterosexual and exclusively heterosexual women. Findings regarding bisexual women’s experiences of CPA were more consistent with those of exclusively lesbian and mostly lesbian women than on CSA variables, and bisexual women reported the highest rate of childhood neglect of all sexual identity groups.

Several explanations have been posited to explain sexual identity differences in childhood abuse. Some researchers have suggested that gender atypical appearance and behaviors may contribute to higher rates of abuse in childhood (Balsam, et al., 2005). Adults who negatively perceive gender atypicality may use physical violence as a means of punishing girls who do not conform to societal norms and standards. Consistent with this explanation, D’Augelli, Grossman, and Starks (2006) found that among LGB youth, those who expressed gender atypicality as children were more likely to be verbally and physically victimized than those who did not. It is possible that among sexual minority subgroups, girls who later identify as exclusively and mostly lesbian are more likely to express gender atypicality than girls who later identify as bisexual and mostly heterosexual, partially explaining differences in childhood abuse rates.

“Acting out” behaviors, such as substance use use, have been proposed as another explanation for higher rates of victimization among sexual minority youth (Balsam, et al., 2005). Sexual minority adolescents have higher rates of alcohol and other drug use than heterosexual adolescents (Marshal et al., 2008), and preliminary evidence suggests that bisexual and mostly heterosexual self-identified youth have different substance use trajectories than youth who identify as exclusively gay/lesbian or exclusively heterosexual (Marshal, Friedman, Stall, & Thompson, 2009). Substance use increases vulnerability to victimization as youth become involved in more risky situations and/or are physically and psychologically impaired from escaping those situations. Parents and caregivers may also violently punish substance-using or delinquent youth for their acting-out behaviors. Through these mechanisms higher rates of substance use among sexual minority youth, and differences in alcohol- and drug-use patterns among sexual minority subgroups, could contribute to sexual identity group differences in childhood abuse.

Among sexual minority subgroups, differences in sexual identity formation and development may moderate abuse risk, as well. D’Augelli (2003) found that lesbian and bisexual youth who self-identified earlier in life reported more lifetime sexual orientation-related victimization than those who self-identified later. Although he did not examine bisexual/lesbian differences, women from different sexual identity subgroups are believed to have distinct sexual identity formation trajectories (Diamond, 2000; Rosario, Schrimshaw, Hunter, & Levy-Warren, 2009) that may confer differing levels of risk or protection from childhood abuse.

Future Research Directions

Variability of childhood abuse experiences by sexual orientation subgroups highlights the importance of research that examines subgroup differences. This paper represents an important step in this direction, but additional research is needed that examines potential explanations for subgroup differences. This could be accomplished with theory-driven prospective designs and, when prospective data are unavailable, retrospective cross-sectional studies that are designed to tease out these explanations. In addition to future studies of sexual identity subgroups, research that examines other demographic differences in childhood abuse experiences among sexual minority women is also needed. For example, Balsam and colleagues (2010) found that African American and Latino/a LGB individuals were more at risk for childhood abuse than Caucasian LGBs, suggesting that race/ethnicity may also play an important role in the childhood abuse experiences of sexual minority women. Consistent with the growing emphasis on intersectionality of gender, sexual, and racial/ethnic identity (Balsam, Molina, Beadnell, Simoni, & Walters, 2011; Meyer, 2010; Stirratt, Meyer, Ouellette, & Gara, 2008), future studies are needed that examine the intersection of racial/ethnic and sexual minority identities in childhood abuse outcomes.

Study Limitations

One limitation of the current study is the relatively small number of bisexual women in the sample. With greater power to detect subgroup differences, the findings may have pointed toward significantly higher rates of childhood abuse and more severe abuse experiences among bisexual women than among other sexual minority women. Our study also would have benefited from a larger number of mostly heterosexual participants. Mostly heterosexual women had rates of CSA prevalence and characteristics that were intermediate between, and generally not significantly different from, exclusively heterosexual women’s rates on the one hand and exclusively lesbian, mostly lesbian, and bisexual women’s rates on the other. Trends in the data suggest that with a larger subsample size many of the differences observed between the mostly heterosexual and exclusively heterosexual women might have been significantly different. Given findings from several studies indicating that women who identify as mostly heterosexual are at significantly greater risk for negative mental health outcomes (Austin, Jun, et al., 2008; Hughes, Szalacha, Johnson, et al., 2010; Hughes, Szalacha, & McNair, 2010), additional research is needed to determine the factors that lead women to self-label as mostly heterosexual and how these factors may contribute to elevated health risks.

Our study was also limited in that information about CSA and CPA is based entirely on self-reports. Although efforts were made to minimize potential response bias by placing these questions near the end of the survey, after rapport was developed, it is possible that participants may have under-reported their childhood abuse experiences. It is also possible that heterosexual and sexual minority women may differ in their willingness to report childhood abuse (Hughes, et al., 2001) or in their interpretation of childhood experiences as abuse (Austin, Jun, et al., 2008), which may have biased prevalence estimates. Finally, given that the majority of sexual minority women in this sample were recruited from the community, this study is subject to the inevitable limitations of community sampling of so-called “hidden” populations. In particular, the sample may not be representative of the total population of sexual minority women, in that it likely excludes women who are have not disclosed their minority sexual identity or who are not comfortable participating in a study of sexual minority women (Meyer & Wilson, 2009). Although national probability studies increasingly include questions that assess sexual identity, even these studies likely exclude those sexual minority women (and men) who are most closeted; and these national probability samples typically include relatively small numbers of individuals who identify as lesbian, mostly lesbian, bisexual, and mostly heterosexual.

Despite its limitations, this study provides new and important information about potential sexual identity subgroup differences in childhood abuse. Given the alarmingly high rates of CSA and CPA among sexual minority women (Austin, Jun, et al., 2008; Balsam, et al., 2005; Corliss, et al., 2002; Hughes, et al., 2000; Matthews, et al., 2002) and the lifelong consequences of childhood abuse experiences (Bulik, et al., 2001; Kendler, et al., 2000; Zink, et al., 2009), understanding the factors that contribute to the elevated risks of childhood abuse among sexual minority women—as well as how these risks may differ between groups within sexual minority populations—is critically important. Such information could be used to develop targeted prevention and intervention strategies aimed at reducing rates of childhood abuse overall, as well as eliminating sexual orientation-related disparities in experiences of childhood abuse.