Arch Sex Behav DOI 10.1007/s10508-017-1099-x ORIGINAL PAPER The Relationship Between Indicators of Depression and Anxiety and Sexual Orientation in Canadian Women Lanna J. Petterson1 • Doug P. VanderLaan2,3 • Tonje J. Persson4 • Paul L. Vasey1 Received: 3 December 2015 / Revised: 3 October 2017 / Accepted: 10 October 2017 Springer Science+Business Media, LLC 2017 Abstract Previous studies examining the associations between women’s sexual orientation and mental health have obtained inconsistent results. Whereas some studies have suggested that statusasalesbianorbisexualwomanmaybeassociatedwithgreater mental health risk, others have suggested that bisexual women may be specifically vulnerable to mental health problems. The current study examined two competing models in a non-clinical sample of Canadian women (N = 278). The first model predicted that women who reported bisexual attraction would endorse more indicators of depression and anxiety compared to women who reported monosexuality(eithersame-oropposite-sexattraction).Thesecondmodel predicted that women who reported relatively greater same-sex attraction would exhibit elevated indicators of depression and anxiety comparedto women who reported opposite-sex attraction. Consistent with Model 1, greater bisexual attraction predicted greaterendorsement ofindicatorsofdepressionandanxietycompared to greater same-sex or opposite-sex attraction. These findings suggest that, in women, bisexuality may be associated with higher risk of depression and anxiety than monosexuality. Future researchmaybenefitfromexploringriskfactorspotentiallyuniqueto the mental health of bisexual women. & Lanna J. Petterson firstname.lastname@example.org 1 Department of Psychology, University of Lethbridge, Lethbridge, AB T1K 3M4, Canada 2 Department of Psychology, University of Toronto Mississauga, Mississauga, ON, Canada 3 Underserved Populations Research Program, Child, Youth and Family Division, Centre for Addiction and Mental Health, Toronto, ON, Canada 4 Department of Psychology, Concordia University, Montreal, QC, Canada Keywords Sexual orientation Depression Anxiety Mental health Bisexuality Introduction Research examining the association between mental health and sexualorientationminoritystatus,irrespectiveofgender,hassuggested that lesbian, gay, and bisexual (LGB) individuals may be at greater risk of poor mental wellness and mental health problems, including greater depression and anxiety symptoms, than heterosexual individuals (Abelson, Lambevski, Crawford, Bartos, &Kippax,2006;Balsam,Beauchaine,Mickey,&Rothblum,2005; Booker, Rieger, & Unger, 2017; Burton, Marshal, Chisolm, Sucato, & Friedman, 2013; Cochran & Mays, 2000a, b; Cochran, Sullivan, & Mays, 2003; Conron, Mimiaga, & Landers, 2010; Fergusson, Horwood, & Beautrais, 1999; Frisell, Lichtenstien, Rahman, & Långerström, 2010; Gilman et al., 2001; Jorm, Korten, Rodgers, Jacomb, & Christensen, 2002; King et al., 2003, 2008; Sandfort, Bakker, Schellevis, & Vanwesenbeeck, 2006; Zietsch, Verweij, Bailey, Wright, & Martin, 2011; Zietsch et al., 2012). However, results have been less consistent when women’s mental health has been studied independently (discussed below). Given that the associationbetweenwomen’ssexualorientationandmentalhealth has varied across studies, additional research is needed to examine the relationship between sexual orientation and mental health outcomes. Somestudiesindicatethatlesbianwomenexhibitworsemental health outcomes than heterosexual women (e.g., Bolton & Sareen, 2011; Kerr, Santurri, & Peters, 2013). Yet, this only appears to be the case for certain aspects of mental health and may vary based on measures of sexual orientation. For example, Sandfort, de Graaf, Bijl,andSchnabel(2001)foundthatlesbianwomenwereatgreater risk of depression, but not anxiety, compared to heterosexual women. Bostwick, Boyd, Hughes, and McCabe(2010)found that 123 Arch Sex Behav although women who identified as lesbian had a higher prevalence of mood and anxiety disorders than those who identified as heterosexual, women who reported only same-sex sexual partners had a lower prevalence of mood and anxiety disorders. A number of studies have found that both bisexual and lesbian women experience greater mental health risk compared to heterosexualwomen,butbisexualwomenexperiencegreatermental health risk compared to both heterosexual and lesbian women (Bolton & Sareen, 2011; Bostwick et al., 2010; Conron et al., 2010; Fredriksen-Goldsen, Kim, Barkan, Balsam, & Mincer, 2010; Kerr et al., 2013; Udry & Chantala, 2002). These negative outcomes include more anxiety, anger, depressive symptoms, self-injury, suicidal ideation, and suicide attempts. However, other studies have found that bisexual and lesbian women report similar mental health outcomes, but ones that are worse when compared to heterosexual women (Koh & Ross, 2006; Warner et al., 2004). Itisessentialtohaveaclearunderstandingofhowmentalhealth disparities are patterned within the population to inform the development of evidenced-based prevention and intervention efforts. In the absence of replicated studies, there is a risk of developing strategies for prevention and intervention that are ill-informed and may target groups inappropriately. At best, a failure to garner a clear understanding of mental health patterning within the population may lead to ineffective allocation of resources (such aspersonnelandfunding).Atworst,suchafailuremaylessenattention to those groups who are most in need of mental health support. With this in mind, the current study investigated the association between women’s sexual orientation and indicators of major depressive disorder (MDD) and generalized anxiety disorder (GAD). Specifically, the focus of the study was to examine the relationship between these mental health domains and the continuous Kinsey rating scale of sexual orientation in women to garnerinsightsintohowmentalhealthindicatorsaredistributedacross this continuum without artificially creating sexual orientation groupings. Two competing predictions were tested based on the existing mental health and sexual orientation literature. Prediction 1 Women with bisexual attraction will report more indicators of depression and anxiety than women with monosexual attraction (attraction to either same- or opposite-sex individuals). The first prediction was derived from previous findings indicating that bisexual women experience the highest risk of depression and anxiety (e.g., Bolton & Sareen, 2011; Bostwick et al., 2010; Conron et al., 2010; Fredriksen-Goldsen et al., 2010; Kerr et al., 2013; Udry & Chantala, 2002). If correct, the association between women’s sexual orientation, along the heterosexual– bisexual–lesbiancontinuum,andindicatorsofdepressionandanxiety would show a negative quadratic relationship (i.e., an inverted U-shape). Prediction 2 Variation away from exclusive opposite-sex attraction is associated with greater endorsement of indicators of depression and anxiety. The second prediction was derived 123 frompreviousfindingsindicatingthatlesbianandbisexualwomen may both experience higher risk of depression and anxiety than heterosexual women (e.g., Cochran et al., 2003; Fergusson et al., 1999; Frisell et al., 2010; Gilman et al., 2001; King et al., 2008; Koh & Ross, 2006; Sandfort et al., 2006; Warner et al., 2004; Zeitsch et al., 2011, 2012). If correct, the association between women’s sexual orientation, along the heterosexual–bisexual– lesbiancontinuum,andindicatorsofdepressionandanxietyshould show a positive linear relationship. Method Participants Atotalof278womenwereincludedinthepresentstudy.AKinseytypescalewasusedtoassessparticipantsexualorientation(Kinsey, Pomeroy, & Martin, 1948). This measure asked participants to select which response option most accurately described their sexual feelings over the last year. Participants were given the response range from 0 (‘‘sexual feelings toward the opposite sex only’’) to 6 (‘‘sexual feelings toward the same sex only’’), as well as the option of‘‘no sexual feelings.’’Of the total sample, 19.8%(n = 55)reportedaKinsey0rating(‘‘sexualfeelingstoward the opposite sex only’’); 18.0% (n = 50) reported Kinsey 1 rating (‘‘most sexual feelings toward the opposite sex, but an occasional fantasy about the same sex’’); 16.9% (n = 47) reported a Kinsey 2 rating (‘‘most sexual feelings toward the opposite sex, but some definite fantasy about the same sex’’); 15.1% (n = 42) reportedaKinsey3rating(‘‘sexualfeelingsequallydividedbetween males and females; no strong preference foroneor theother’’); 6.8% (n = 19)reportedaKinsey4rating(‘‘mostsexualfeelingstowardthe same sex, but some definite sexual feelings toward the opposite sex’’);8.6%(n = 24)reportedaKinsey5rating(‘‘mostsexualfeelings toward the same sex, but an occasional fantasy about the oppositesex’’); and14.7% (n = 41)reportedaKinsey 6rating(‘‘sexual feelings toward the same sex only’’). This study was conducted via an online survey that was advertised to 757 university distribution lists and community organizations throughout Canada, and through the social networking site, Facebook. Individuals were informed that the survey would containquestionspertainingtosexuality,personality,andfamilyrelations. Facebook uses an algorithm that determines to whom the advertisement will appear. The survey’s advertisement was targeted to English-speaking Canadian men and women who were 18 years of age and older. Additionally, to ensure that adequate numbersofnon-heterosexualparticipantswereobtained,weselected the option to target women who were interested in women or in men and men who were interested in men or in women. Because participantrecruitment fromuniversitydistributionlistsandcommunity organizations was limited (2.2%), the sample largely reflects the responses of participants who were collected through Facebook (97.8%). Arch Sex Behav Measures In addition to information on sexual orientation, biographic information was collected pertaining to the participant’s age (in years), socioeconomic status during childhood, annual income, level of education completed, religious affiliation, and religiosity. Indicators of depression and anxiety in childhood and adulthood were assessed through measures that we created based on the diagnostic criteria in the DSM-IV-TR (American Psychiatric Association [APA], 2000). Participants were provided with items that were based on the diagnostic criteria for MDD, GAD, agoraphobia, SAD,PD,OCD,andspecificphobiasaspartofalargerstudy.However, because our analysis pertained to 278 participants, which is relatively small for the intended analysis, we limited our focus to examiningonlytheMDD(6items)andGAD(9items)items.Additionally, we did not have any predictions for the present study regarding childhood expression of mental health. As such, only the adulthood MDD and GAD items were included in the present analysis. The adulthood DSM items asked participants how accurately the descriptions reflected their experiences since the age of 18 years. Participants used a 5-point Likert-type scale that ranged from 1 (‘‘strongly disagree’’) to 5 (‘‘strongly agree’’). The standardized inter-item reliability (alpha)forthe 13 itemsthat wereused in the final analysis was .91 for the sample (Note: two items were dropped during principal component analysis; see below). Statistical Analysis Analyses were conducted using SPSS Statistics version 23. Principal component analysis (PCA) was used to reduce the indicators of depression and generalized anxiety (based on the MDD and GADitems)toacorecomponent.Componentscoreswerederived using the regression method. The standardized component scores wereusedasacompositemeasureoftheitemsforfurtheranalyses. A curvilinear estimation was conducted (with Kinsey rating included as a continuous variable) to examine the relationship between participants’ Kinsey ratings and indicators of depression/generalized anxiety component scores. Primarily, this analysis was used to test (1) whether greater non-monosexual sexual attraction was predictive of greater endorsement of indicators of depression and generalized anxiety or (2) whether greater attraction to the same sex was predictive of greater endorsement of indicators of depression and generalized anxiety. A nonsignificant linear relationship, but a significant quadratic relationship, would furnish support for Prediction 1, whereas a significant linear relationship would furnish support for Prediction 2. Thecurvilinearestimationwaslimited,however,inwhatinformation it could yield regarding the strength of the relationship between the Kinsey ratings and indicators of depression/generalized anxiety. As such, two hierarchical linear regressions were conducted. The first examined the relative predictive contribution of the linear term and the quadratic term by first including the linear term and then adding the quadratic term to the model. This analysis provided insight into which term (i.e., either the linear term described in Prediction 2 or the quadratic term described in Prediction 1) was better fit to the data. The second examined how well the quadratic Kinsey rating variable predicted indicators of depression and generalized anxiety, to test the goodness of fit of the negative quadratic model proposed in Prediction 2. Biographic variables were entered first for both hierarchical linear regressions. Results Principal Component Analysis PCA was conducted to assess whether core components underlay items related to depression and generalized anxiety. Two MDD items (‘‘I experienced rapid weight loss or weight gain’’and‘‘I feel as though my regular movements are slow’’) were excluded from subsequentanalysisbecausepreliminarybivariatecorrelationanalysis revealed several correlation coefficients that were below r = .30 for these items (low correlation coefficients may suggest that the items are relatively distinct from the other items). The remaining 13 items were included in the PCA. The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy was .89, and Bartlett’s test of sphericity was significant at p\.001, indicating that PCA was appropriate (Bartlett, 1954; Kaiser, 1974). All KMO values for the individual items were .83 or higher. A scree plot using PCA extraction indicated that a onecomponent solution was most appropriate. One component was extracted using the maximum likelihood procedure. The loading coefficients, displayed in Table 1, yielded the interpretable component: indicators of depression/generalized anxiety. The indicators of depression/generalized anxiety component accounted for 44.17%ofthevariance.Theextractedcomponentscoresweresaved as a variable using the regression method. Component scores are standardized. Participant scores for the indicators of depression/ generalized anxiety component scores ranged from - 2.18 to 2.02 (M = 0; SD = .96). Regression Analyses Curvilinear Estimation The curvilinear estimation of the indicators of depression/generalized anxiety component scores by Kinsey score is presented in Fig. 1. Inconsistent with Prediction 2, statistical significance was not obtained for the linear equation, F(1, 276) = 2.93, p = .088, R2 = .01. However, consistent with Prediction 1, statistical significance was achieved when the final equation included a quadratic term, F(2, 275) = 6.58, p = .002, R2 = .05. The curve reached its peak between Kinsey ratings of 3 and 4. 123 Arch Sex Behav Table 1 Component loadings for indicators of depression and generalized anxiety Component 1 Generalized Anxiety Disorder I often feel restless or on edge .72 I tire easily by daily activity .69 I frequently forget what I have been previously been thinking about or am unable to concentrate .72 I am frustrated or bothered by things to a greater extent .60 than others seemed to be I frequently have tense or tight muscles that are not caused by physical activity .59 I am unable to fall asleep easily or, if I do fall asleep, I have difficulty remaining asleep .58 Major Depression I feel sad or empty frequently .72 I lost interest in things that I had previously been interested in .59 I experience rapid weight loss or weight gain – I have uncommon sleep patterns (slept excessively or infrequently) .64 I am often fatigued by daily activities I feel as though my regular movements are slow .68 – I have a lowered sense of self worth .66 Covariate Analysis I have difficulty making decisions or concentrating .78 I frequently have morbid thoughts .63 Fig. 1 Indicators of depression/generalized anxiety component scores as a function of Kinsey ratings 123 Consequently, the curvilinear estimate suggested that, as Kinsey ratings increased from Kinsey 0 (exclusive heterosexual attraction),participantsscoredhigherontheindicatorsofdepression/ generalized anxiety component. However, this positive relation peaked around the Kinsey 3–4 range (approximately equal attraction to men and women), after which point the inverse was found. As Kinsey ratings increased from Kinsey 4, participants scored lowerontheindicatorsofdepression/generalizedanxietycomponent. Thus, subsequent analysis primarily focused on examining whether a shift toward bisexual attraction significantly predicted elevated indicatorsofdepression/generalizedanxietycomponent scores,whilecontrollingforseveralpotentiallyconfoundingvariables. Because the subsequent analysis was conducted using linear regression, the Kinsey rating variable required modification to evaluate adherence to a negative quadratic model. To do so, the Kinsey rating variable was centered around 3.5 and then squared, basedonthecurveestimations(Fig. 1).Doingsomodifiedthescale to range from bisexual attraction(low) to exclusive attraction to either men or women (high). Covariate analysis was conducted to determine whether the biographic variables varied within the sample in a patterned manner, which could influence the variables of interest. Sexual orientation groups were formed to facilitate biodemographic comparisons. Descriptive statistics pertaining to biographic variables are shown Arch Sex Behav Table 2 Descriptive statistics regarding participants’ biographic information by sexual orientation group Heterosexual women n = 105 Bisexual women n = 108 Lesbian women n = 65 Religiosity: % (n) Not religious 52.38 (55) 70.37 (76) 66.15 (43) Some degree of religiosity 47.62 (50) 29.63 (32) 33.85 (22) Christianity 57.14 (60) 39.81 (43) 63.08 (41) Other 42.86 (45) 60.19 (65) 36.92 (24) 18.10 (19) 24.07 (26) 24.62 (16) 40 (42) 39.81 (43) 41.54 (27) 41.90 (44) 36.11 (39) 33.85 (22) 37.01 (17.31) 23.57 (9.70) 28.06 (12.95) Religious affiliation: % (n) Childhood socioeconomic status: % (n) Upper/upper middle Middle Lower/lower middle Age (in years): M (SD) Education: (starting at Grade 1) M (SD) Annual income (in CAD): M (SD) 15.07 (3.38) 13.61 (2.41) 14.03 (2.21) 25,101.56 (23,401.47) 15,211.10 (15,787.11) 17,294.23 (16,107.66) Table 3 Mean (SD) indicators of depression/generalized anxiety component scores (standardized) by Kinsey rating Kinsey 0 rating n = 55 Kinsey 1 rating n = 50 Kinsey 2 rating n = 47 Kinsey 3 rating n = 42 Kinsey 4 rating n = 19 Kinsey 5 rating n = 24 Kinsey 6 rating n = 41 - .33 (1.03) - .03 (.93) .03 (.95) .33 (.88) .27 (.72) .11 (.87) - .08 (1.02) inTable 2bysexual orientationgroup: heterosexualwomen (Kinsey0–1),bisexualwomen(Kinsey2–4),andlesbianwomen(Kinsey 5–6).1 Participants differed significantly regarding their response to the biographic variables: religiosity (dummy coded as not religious and some degree of religiosity), v2(2) = 7.82, p = .020; religious affiliation (dummy coded as Christian vs.‘‘other’’), v2(2) = 10.73, p = .005; age, Brown–Forsythe statistic F(2, 222.65) = 26.41, p\ .001; education, Brown–Forsythe statistic F(2, 252.23) = 8.25, p\.001;andannualincome,Brown–ForsythestatisticF(2,245.41)= 8.09, p\.001. Participants did not differ significantly in terms of socioeconomic status during childhood, v2(4) = 2.00, p = .737. Pearson correlations were conducted to determine whether thebiographicvariablesthatdifferedbetweengroupscorrelatedwith the focal variables (df = 276 for all correlations). The biographic variablesthatdifferedbetweengroupsandcorrelatedwiththemodified Kinsey score were religious affiliation (negatively correlated: r = - .20, p = .001), religiosity (positively correlated: r = .22, p\ .001), education (positively correlated: r = .19, p = .001), annual income (positively correlated: r = .22, p\.001), and age (positively correlated: r = .44, p\.001). The biographic variables that differedbetweengroupsandcorrelatedwiththeindicatorsofdepres1 Groups were formed only for covariate analysis. Kinsey scores were combined to ensure that groups had adequate numbers of participants to make comparisons. sion/generalized anxiety component were education (negatively correlated:r = - .22,p\.001)andannualincome(negativelycorrelated: r = - .18, p = .002). No other correlations achieved statistical significance (all other p-values ranged from p = .278 to p = .64). Thus, because religiosity, religious affiliation, age, education, and annual income may have a confounding influence on associationsbetweensexualorientationandmentalhealth,theywerecontrolled for in subsequent analysis. Multiple Regression Analysis Mean indicators of depression/generalized anxiety component scores by Kinsey rating are shown in Table 3. Two multiple hierarchicalregressionswereconductedtoexaminewhetherthedegree to which participants reported bisexual attraction predicted indicatorsofdepression/generalizedanxiety.Forbothregressions,the biographic variables—education, annual income, age, religiosity (dummy coded as not religious and somewhat religious), and religious affiliation (dummy coded as Christianity or other)—were entered into the first block of the regression model using forced entry. For the first regression, the linear Kinsey score term wasadded in thesecond step, and the quadratic Kinsey score term wasadded in the third step. Table 4 shows the models of the biographic variables, the linearterm, and thequadratic term predicting indicators of depression/generalized anxiety component scores. When all 123 Arch Sex Behav Table 4 Regression analyses predicting indicators of depression/generalized anxiety component scores B (95% CI) SE of B b p sr2i Step 1a Constant .84 (.23, 1.45) .31 Education - .07 (- .11, - .03) .02 - .20 .007 .001 .04 Income .02 \.01 (\.01,\.01) \.01 - .16 .017 Age .01 (\- .01, .01) \.01 .09 .195 .01 Religiosity .09 (- .15, .33) .12 .05 .459 \.01 Religious affiliation .13 (- .12, .37) .12 .07 .307 \.01 .78 (.17, 1.40) .31 Step 2b Constant .013 Education - .07 (- .11, - .03) .02 - .19 .002 .03 Income Age \.01 (\.01,\.01) .01 (\- .01, .02) \.01 \.01 - .15 .11 .023 .118 .02 .01 Religiosity .11 (- .13, .35) .12 .06 .371 \.01 Religious affiliation .15 (- .09, .40) .12 .08 .226 \.01 Kinsey score .04 (- .02, .10) .03 .09 .159 .01 Constant .81 (.21, 1.41) .31 Education - .06 (- .10, - .02) .02 - .18 .004 .03 Income \.01 (\.01,\.01) \.01 - .15 .022 .02 Age .01 (\- .01, .02) .01 .17 .016 .02 Religiosity .15 (- .09, .39) .12 .08 .208 .01 Religious affiliation .11 (- .13, .36) .12 .06 .358 \.01 Kinsey score - .01 (- .07, .06) .03 - .02 .83 \.01 Kinsey score squared - .05 (- .09, - .02) .02 - .24 .002 Step 3c .009 .03 Bold terms are those that have been entered at each step. Regression coefficients (B), 95% confidence intervals (CI) for B, standardized regression coefficients (b), t-test significances (p), and semi-partial correlations (sr2i ) for biographic variables and the linear and quadratic Kinsey score terms predicting indicators of depression/general anxiety component scores are shown in the table. Model summary statistics for the models are indicated by superscripts a b c R2 = .08; adjusted R2 = .06; F-ratio (5, 272) = 4.45, p = .001 R2 = .08; adjusted R2 = .06; F-ratio (6, 271) = 4.05, p = .001; DR2 = .01; DF (1, 271) = 1.99, p = .159 R2 = .12; adjusted R2 = .09; F-ratio (7, 270) = 5.06, p\.001; DR2 = .03; DF (1, 270) = 10.28, p = .002 seven variables were included, four were shown to predict indicators of depression/generalized anxiety component scores: education(negativerelationship;b = - .18,sr2i = .03,p = .004),annual income (negative relationship; b = - .15, sr2i = .02, p = .022), age (positive relationship; b = .17, sr2i = .02, p = .016), and Kinsey score squared (negative relationship [with low scores indicating greaterbisexual attraction andhighscores indicating greatermonosexual attraction]; b = - .24, sr2i = .03, p = .002). This analysis indicated that 11.6% of the variability in indicators of depression/ generalized anxiety component scores was predicted by the five biographicvariables,thelinearterm,andthequadraticterm.Whereas the addition of the linear term did not significantly improve the fit of the model (the inclusion of the term increased the amount of variance accounted for by less than 1%), the addition of the quadratic term did significantly improve the fit of model (the inclusion of the term increased the amount of variance accounted for by 3.37%). 123 Forthesecondregression,thequadraticKinseyscoreterm was added in the secondstep. This analysis indicated that 11.59% of the variability in indicators of depression and generalized anxiety component scores was predicted by the five biographic variables and the quadratic term (adjusted R2 = .10; F-ratio for Step 2 [6, 271] = 5.92, p\.001). The inclusion of the quadratic term in Step 2 improved the model’s goodness of fit (DR2 = .04; DF [1, 271] = 12.34, p = .001). Discussion The current study evaluated the association between women’s sexual orientation and indicators of depression and anxiety. One component was shown to underlie the items related to depression andgeneralizedanxiety.Componentscoresservedasacomposite measure of the MDD and GAD items for subsequent analysis. Arch Sex Behav Consistent with the first prediction, the extent to which participants endorsed indicators of depression and generalized anxiety was predicted by the degree to which they reported attraction to both sexes, as opposedtoonesex or the other. Inconsistent with thesecondprediction,theextenttowhichparticipantsendorsedindicators of depression and anxiety was not predicted by the degree to whichtheyreportedgreatersame-sexsexualattraction,asopposed to opposite-sex sexual attraction. Hence, the current study is more in line with the prediction that bisexuality, compared to monosexuality, predicts greater depression and anxiety (e.g., Bolton & Sareen, 2011; Bostwick et al., 2010; Conron et al., 2010; Fredriksen-Goldsen et al., 2010; Kerr et al., 2013; Persson, Pfaus, & Ryder, 2015; Udry & Chantala, 2002) than the prediction that non-heterosexual attraction predicts depression and anxiety (Cochran et al., 2003; Fergusson et al., 1999; Frisell et al., 2010; Gilman et al., 2001; King et al., 2008; Koh & Ross, 2006; Sandfort et al., 2006; Warner et al., 2004; Zeitsch et al., 2011, 2012). Importantly, however, our analysis indicated that much of the varianceinindicatorsofdepressionandgeneralizedanxietyremained unaccounted for by participants’ endorsement of (non)monosexual sexualfeelingsandthebiographicvariables.Assuch,itappearsthat additional factors are tied to women’s depression and generalized anxiety, which may be either related to sexual orientation or independent from it. That said, it appears the degree to which women experience non-monosexual sexual feelings is related to this aspect of mental health. It is possible that bisexual women scored higher than heterosexual women on indicators of depression and anxiety because theyexperiencegreaterstressassociatedwiththeirminoritysexual orientation status. The minority stress hypothesis has been suggestedasapossibleexplanationforLGBindividuals’mentalhealth risk (Meyer, 1995, 2003). In line with this idea, features of minority stress(e.g.,perceivedorexpecteddiscriminationandvictimization, internalized homonegativity, anticipated social rejection, and concealment of one’s sexual orientation identity) have been associated withnegativementalhealthoutcomes(e.g.,Balsam&Mohr,2007; Hatzenbuehler, Nolen-Hoeksema, & Erickson, 2008; Lea, de Wit, & Reynolds, 2014; Lewis, Derlega, Griffin, & Krowinski, 2003; Mays&Cochran,2001;Pachankis,Cochran,&Mays,2015;Vanden Berghe, Dewaele, Cox, & Vincke, 2010). Consistent with minority stress hypothesis, LGB adults and youth living in regions of the USA that are more supportive of the LGB community and that grant protection against LGB discrimination are mentally healthier than those living in other regions (Hatzenbuehler, 2011; Hatzenbuehler, Keyes, & Hasin, 2009). Additionally, mental healthcare visits and costs for gay and bisexual men decreased following the legalization of gay marriage in the USA, suggesting that mental health improved when institutional discrimination was abated (Hatzenbuehler, O’Cleirigh, Grasso, Mayer, Safren, & Bradford, 2012). Hatzenbuehler (2009) expanded on this hypothesis by proposing thatincreased stigma-related stress(i.e., stress that can result from experiencing stigma based on one’s minority status) can lead to increased problems associated with emotional regulation, social andinterpersonal relationships,and cognitiveprocesses— psychological processes that are associated with elevated mental health risk, regardless of an individual’s sexual orientation. These generalpsychologicalprocessesaresuggestedtomediatetherelationship between minority stress and LGB individual’s preponderance of adverse mental health outcomes. For instance, deficits in emotional regulation mediated the relationship between sexual minority statusand bothdepressionand anxiety among adolescents (Hatzenbuehler et al., 2008). The finding that bisexual women scored higher than lesbian womenonindicatorsofdepressionandanxietycouldbeexplained by an elevated frequency of experiencing stigma or discrimination by the heterosexual majority and other monosexual sexual minorityindividuals,knownas‘‘doublediscrimination.’’Ithasbeenfound, for example, that heterosexual individuals, gay men, and lesbian women report more negative attitudes toward bisexual individuals than toward both gay men and lesbian women (Friedman et al., 2014;Herek,2002;Zivony&Lobel,2014).Thus,bisexualwomen maybeatahigherriskofstigmatizationthanlesbianwomen,which in turn may account for their elevated scores for indicators of depression and anxiety. Additionally, bisexual women with oppositesex partners experience greater depression symptoms than lesbians,andthisrelationshipismediatedbyexperiencesofrejection and social exclusion by gay men and lesbians (Dyar, Feinstein, & London, 2014). Alternatively, although not mutually exclusive, research has suggested that greater exposure to adverse life events may partially account for poorer mental health among lesbian, gay, and bisexual individuals (Frisell et al., 2010; Mays & Cochran, 2001; Wamala,Boström,& Nyqvist,2007).Forexample,Balsam,Rothblum, and Beauchaine (2005) found that gay, lesbian, and bisexual men and women were more likely to have experienced victimization, such as childhood abuse, domestic abuse, psychological maltreatment,andsexual coercion thanheterosexual men and women. Additionally,however,bisexualmenandwomenweremorelikely to have experienced non-intercourse sexual coercion and rape than gay men and lesbian women. Other research has found that lesbian and bisexual women report more abuse, chronic abuse, and sexual abuse during adolescence than heterosexual women (Austin et al., 2008). Persson et al. (2015)demonstrated that non-monosexual women reported moresymptoms ofanxietyand depression,moreriskysexual behavior, more childhood abuse, and lower levels of sexual orientation disclosure than monosexual women. Further, they found that elevated risky sexual behavior and decreased sexual orientation disclosure mediated the relationship between sexual orientation and depression and anxiety. Childhood abuse did not moderate the relationship between sexual orientation and depression or anxiety. In sum,these findingssuggest that poorermental health among bisexual women may be partly related to risky sexual behavior and sexual orientation concealment. Other research has also suggested 123 Arch Sex Behav that sexual orientation disclosure may be associated with mental health.Forexample,Juster,Smith,Ouellet,Sindi,andLupien(2013) foundthatdisclosedsexualorientationminorityindividualsreported fewer psychiatric symptoms compared to those who were non-disclosed. The findings discussed above may all be relevant to explainingloweredmentalhealthamongbisexualindividualsandshouldbe taken into account when potential interventions are considered. Our findings further highlight the increased risk of depression and anxiety among bisexual women. Research has found that bisexual men and women are less likely than gay men and lesbian women to disclose their sexual orientation to their healthcare providers or to use available healthcare services (Durso & Meyer, 2013; Kerr et al., 2013). Therefore, bisexual men and women may not receive health careappropriatetotheirspecificneeds.Additionally,theirhealthcare providers may not target them as an at-risk group. The present findings, based on a community sample, underline the importance of implementing prevention and intervention strategies specifically targeted toward bisexual individuals. Limitations and Future Directions The present study assessed sexual orientation based on the participant’s response to a single measure of sexual orientation. Specifically,sexualorientationwasmeasuredviaparticipant’sself-reported sexual feelings, as opposed to sexual orientation identity or sexual behavior.This may havebeen problematic for two reasons. First,the participants might have varied in their interpretation of what may be considered ‘‘sexual feelings.’’ Although potentially imprecise, this measure is arguably beneficial because it permits the consideration ofwomenwho experience sexual attractions toorfantasies/thoughts about one sex (or both sexes) without acting or identifying in a manner to express these sexual feelings. Additionally, measures such as sexual behavior and sexual identity are limited to an extent because behavior and identity can be constrained by cultural factors and individual circumstances. The relationship between mental health and sexual orientation may vary based on which facets of individuals’ sexual orientation aredirectlyexaminedandthecompositionofeachsexual orientation group (e.g., Bostwick et al., 2010). For instance, women who engage in bisexual sexual behavior may face different degrees of social acceptance than women who have bisexual sexual feelings butdonotactonthem.Namely,behaviorallybisexualwomenmay experience social rejection, stigmatization, and discrimination from the heterosexual majority and the lesbian community, whereas women who experience sexual feelings for both men and women butdonotactonthemmaynotexperiencethesamenegativesocial reaction. Women who experience bisexual attraction but who do not act on these feelings may be differently affected by their sexual orientation(e.g.,theymayengageinvigilantself-monitoringtoconceal their feelings; they may experience sexual dissatisfaction; or they may feel that they are unable to express part of themselves). Future studies should opt to employ several measures of sexual 123 orientation to further disentangle the relationship between sexual orientation and mental health. Additionally, future studies should ask participants whether their friends, family, and broader social groups are aware of their sexual orientation status, whether they have felt supported, and whether they have experienced negative responses (or have had negative thoughts) related to their sexual orientation. It is important to note that the present sample was drawn from a Canadian population. The social and political climate of Canada tends to be relatively tolerant and accepting of minority sexual orientation individuals, relative to many other cultural groups (Anderson & Fetner, 2008; Widmer, Treas, & Newcombe, 1998). Lesbian, gay, and bisexual individuals may be protected from negativementalhealthoutcomeswhenpoliciesareimplementedto protect these individuals (e.g., institutional policies that protect sexual orientation minority category individualsandban employment discriminationbased on sexual orientation)(Hatzenbuehler, Keyes,&Deborah,2009).Consequently,sexualorientationminority individuals in Canada may experience less stigmatization and discrimination as a result of their sexual orientation status and, thus, less negative mental health outcomes compared to those living in places less protective of sexual minority rights. It is possible that, because Canada is relatively accepting of sexual minority rights, our results may represent a conservative estimate of negative mental health outcomes of lesbian and bisexual women. Hence, caution must be exercised when extrapolating the present findings to less accepting populations. It would be desirable if future research replicated the current study using samples drawn from other sociocultural settings. Furthermore, because our sample was self-selected there may have been a recruitment bias. Such a recruitment bias could, arguably,haveintroducedconfoundsthatmayhaveartificiallyimpacted the results. In other words, those who elected to participate may differ from those who declined to participate and, thus, those who participated may not be fully representative of Canadian women. If so, this bias might have contributed to biographic differences that could have impacted the variables of interest (i.e., indicators of depression and anxiety and Kinsey scale ratings). Nevertheless, it is important tonotethatanumberofbiographicvariableswereincludedascovariates, thus mitigating the potential influence of such additional factors on the variables of interest. Of additional note, the study advertisement was sent to 757 universitydistribution listsandcommunityorganizationsthroughout Canada, but this endeavor was not fruitful. As such, we would encourage future researchers who are seeking to recruit community samples for mental health research to consider alternative participant recruitment methods to prevent low response rates. However, it is important to note that no incentive was provided for participation (e.g., monetary compensation, entry into a gift card draw) in the present study. Participant response rates may have been higher if participants were incentivized to participate. Overall, thecurrent findingsshouldbe interpreted with caution. Our analyses pertained to a non-clinical sample and employed Arch Sex Behav measurement items that were based on diagnostic criteria but were not employedtodeterminewhetherparticipantsmetthediagnostic thresholdforclinicaldiagnosis.Instead,theseitemswereemployed to assay participants’ emotional or physiological states that are associatedwiththeDSMdisorders.Thus,thecurrentfindingsshouldnot beextrapolatedtoclinicalpopulationsandcannotspeaktotherateof diagnosable conditions within the sample. Furthermore, the current assessmentdidnotaccountforspecificcircumstances(e.g.,physical ailments, life events) that could influence participants’ responses. Such measures may limit the clinical implications of the findings; however, the current study provides direction to clinical research by examining the existence of trends that may be present among the general population that may be otherwise unaccounted for. Acknowledgements We thank all of the individuals who agreed to participateinourstudy.VariousstagesofthisresearchweresupportedbytheUniversity of Lethbridge; by a Social Sciences and Humanities Research Council of Canada (SSHRC) Masters Scholarship and a Lethbridge Public Interest Research Group Research Grant to LJP; by a Canadian Institutes of Health Research (CIHR) Postdoctoral Fellowship to DPV; by a CIHR Doctoral Award to TJP; and by a CIHR Catalyst Grant (Methods and Measures for Gender, Sex, and Health; Grant No. 45546) and an Alberta Innovates Health Solutions Sustainability Fund Grant (Grant No. 43528) to PLV. Compliance with Ethical Standards Conflict of interest Theauthorsdeclarethattheyhavenoconflictofinterest. Ethical Statement TheUniversityofLethbridgeHumanSubjectsResearch Committee approved this research. Participants were required to provide informed consent prior to taking part in the study. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the Canadian Tri-Council Policy Statement 2 (2014): Ethical Conduct for Research Involving Humans. References Abelson, J., Lambevski, S., Crawford, J., Bartos, M., & Kippax, S. (2006). Factors associated with ‘feeling suicidal’: The role of sexual identity. Journal of Homosexuality, 51, 59–80. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. Anderson, R., & Fetner, T. (2008). Cohort differences in tolerance of homosexuality: Attitudinal changes in Canada and the United States 1981– 2000. Public Opinion Quarterly, 72, 311–330. Austin, S. B., Jun, H.-J., Jackson, B., Spiegelman, D., Rich-Edwards, J., Corliss, H. L., & Wright, R. J. (2008). Disparities in child abuse victimization in lesbian, bisexual, and heterosexual women in the Nurses’ Health Study II. Journal of Women’s Health, 17, 597–606. Balsam, K. F., Beauchaine, T. P., Mickey, R. M., & Rothblum, E. D. (2005). Mentalhealthoflesbian,gay,bisexual,andheterosexualsiblings:Effects of gender, sexual orientation, and family. Journal of Abnormal Psychology, 114, 471–476. Balsam, K. F., & Mohr, J. J. (2007). Adaptation to sexual orientation stigma: A comparison of bisexual and lesbian/gay adults. Journal of Counseling Psychology, 54, 306–319. Balsam, K. F., Rothblum, E. D., & Beauchaine, T. P. (2005). Victimization over the life span: A comparison of lesbian, bisexual, and heterosexual siblings. Journal of Consulting and Clinical Psychology, 73, 477–487. Bartlett, M. S. (1954). A note on the multiplying factors for various chi square approximations. Journal of the Royal Statistical Society: Series B, 16, 296–298. Bolton, S. L., & Sareen, J. (2011). Sexual orientation and its relation to mental disorders and suicide attempts: Findings from a nationally representative sample. Canadian Journal of Psychiatry, 56, 35–43. Booker, C. L., Rieger, G., & Unger, J. B. (2017). Sexual orientation health inequality: Evidence fromUnderstandingSociety,the UKLongitudinal Household Study. Preventative Medicine. doi:10.1016/j.ypmed.2017. 06.010. Bostwick, W. B., Boyd, C. J., Hughes, T. L., & McCabe, S. E. (2010). Dimensions of sexual orientation and the prevalence of mood and anxiety disorders in the United States. American Journal of Public Health, 100, 468–475. Burton, C. M., Marshal, M. P., Chisolm, D. J., Sucato, G. S., & Friedman, M. S. (2013). Sexual minority-related victimization as a mediator of mental health disparities insexual minority youth:A longitudinal analysis.Journal of Youth and Adolescence, 42, 394–402. Cochran, S. D., & Mays, V. M. (2000a). Lifetime prevalence of suicide symptoms and affective disorders among men reporting same-sex sexual partners: Resultsfrom NHANESIII.American Journal of Public Health, 90, 573–578. Cochran, S. D., & Mays, V. M. (2000b). Relation between psychiatric syndromes and behaviorally defined sexual orientation in a sample of the US population. American Journal of Epidemiology, 151, 516–523. Cochran, S. D., Sullivan, J. G., & Mays, V. M. (2003). Prevalence of mental disorders, psychological distress, and mental health services use among lesbian, gay, and bisexual adults in the United States. Journal of Consulting and Clinical Psychology, 71, 53–61. Conron, K. J., Mimiaga, M. J., & Landers, S. J. (2010). A population-based study of sexual orientation identity and gender differences in adult health. American Journal of Public Health, 100, 1953–1960. Durso, L. E., & Meyer, I. H. (2013). Patterns and predictors of discloser of sexual orientation to healthcare providers among lesbians, gay men, and bisexuals. Sexuality Research and Social Policy, 10, 35–42. Dyar, C., Feinstein, B. A., & London, B. (2014). Dimensions of sexual identity and minority stress among bisexual women: The role of partner gender. Psychology of Sexual Orientation and Gender Diversity, 1, 441– 451. Fergusson, D. M., Horwood, L. J., & Beautrais, A. L. (1999). Is sexual orientation related to mental health problems and suicidality in young people? Archives of General Psychiatry, 56, 876–880. Fredriksen-Goldsen, K. I., Kim, H. J., Barkan, S. E., Balsam, K. F., & Mincer, S. L. (2010). Disparities in health-related quality of life: A comparison of lesbians and bisexual women. American Journal of Public Health, 100, 2255–2261. Friedman, M. R., Dodge, B., Schick, V., Herbenick, D., Hubach, R. D., Bowling, J., … Reece, M. (2014). From bias to bisexual health disparities: Attitudes toward bisexual men and women in the United States. LGBT Health, 1, 309–318. Frisell, T., Lichtenstien, P., Rahman, Q., & Långerström, N. (2010). Psychiatric morbidity associated with same-sex sexual behaviour: Influence of minority stress and familial factors. Psychological Medicine, 40, 315–324. Gilman, S. E., Cochran, S. D., Mays, V. M., Hughes, M., Ostrow, D., & Kessler, R. C. (2001). Risk of psychiatric disorders among individuals reporting same-sex sexual partners in the national comorbidity survey. American Journal of Public Health, 9, 933–939. Hatzenbuehler,M.L.(2009).Howdoessexualminoritystigma‘‘getunderthe skin’’? A psychological medication framework. Psychological Bulletin, 135, 707–730. Hatzenbuehler, M. L. (2011). The social environment and suicide attempts in lesbian, gay, and bisexual youth. Pediatrics, 1, 896–903. Hatzenbuehler, M. L., Keyes, K. M., & Hasin, D. S. (2009). State-level policies and psychiatric morbidity in lesbian, gay, and bisexual populations. American Journal of Public Health, 99, 2275–2281. 123 Arch Sex Behav Hatzenbuehler, M. L., Nolen-Hoeksema, S., & Erickson, S. J. (2008). Minority stress predictors of HIV risk behavior, substance use, and depressive symptoms: Results from a prospective study of bereaved gay men. Health Psychology, 27, 455–462. Hatzenbuehler, M. L., O’Cleirigh, C., Grasso, C., Mayer, K., Safren, S., & Bradford, J. (2012). Effects of same-sex marriage laws on health care use and expenditures in sexual minority men: A quasi-natural experiment. American Journal of Public Health, 102, 285–291. Herek, G. M. (2002). Heterosexuals’ attitudes toward bisexual men and women in the United States. Journal of Sex Research, 39, 264–274. Jorm, A. F., Korten, A. E., Rodgers, B., Jacomb, P. A., & Christensen, H. (2002). Sexual orientation and mental health: Results from a community survey of young and middle-aged adults. British Journal of Psychiatry, 180, 423–427. Juster, R. P., Smith, N. G., Ouellet, É., Sindi, S., & Lupien, S. J. (2013). Sexual orientation and disclosure in relation to psychiatric symptoms, diurnal cortisol, and allostatic load. Psychosomatic Medicine, 75, 103–116. Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika, 39, 31–36. Kerr, D. L., Santurri, L., & Peters, P. (2013). A comparison of lesbian, bisexual, and heterosexual college undergraduate women on selected mental health issues. Journal of American College Health,61, 185–194. King, M., McKeown, E., Warner, J., Ramsay, A., Johnson, K., Cort, C., … Davidson, O. (2003). Mental health and quality of life of gay men and lesbians in England and Wales: Controlled, cross-sectional study. British Journal of Psychiatry, 183, 552–558. King, M., Semlyen, J., See Tai, S., Killaspy, H., Osborn, D., Popelyuk, D., & Nazareth,I.(2008).Asystematicreviewofmental disorder,suicide,and deliberate self harm in lesbian, gay, and bisexual people. BMC Psychiatry, 8, 70. doi:10.1186/1471-244X-8-70. Kinsey, A.C.,Pomeroy,W.B., &Martin,C. E.(1948).Sexualbehaviorinthe human male. Philadelphia, PA: W. B. Saunders. Koh, A. S., & Ross, L. K. (2006). Mental health issues: A comparison of lesbian, bisexual and heterosexual women. Journal of Homosexuality, 51, 33–57. Lea, T., de Wit, J., & Reynolds, R. (2014). Minority stress in lesbian, gay, and bisexual: Associations with psychological distress, suicidality, and substance abuse. Archives of Sexual Behavior, 43, 1571–1578. Lewis, R. J., Derlega, V. J., Griffin, J. L., & Krowinski, A. C. (2003). Stressors for gay men and lesbians: Life stress, gay-related stress, stigma consciousness,anddepressivesymptoms.JournalofSocialandClinicalPsychology, 22, 716–729. Mays, V. M., & Cochran, S. D. (2001). Mental health correlates of perceived discrimination among lesbian, gay, and bisexual adults in the United States. American Journal of Public Health, 91, 1869–1876. 123 Meyer, I. H. (1995). Minority stress and mental health in gay men. Journal of Health and Social Behavior, 36, 38–56. Meyer, I. H. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129, 674–697. Pachankis, J. E., Cochran, S. D., & Mays, V. M. (2015). The mental health of sexual minority adults in and out of the closet: A population-based study. Journal of Consulting and Clinical Psychology, 83, 890–901. Persson, T. J., Pfaus, J. G., & Ryder, A. G. (2015). Explaining mental health disparities for non-monosexual women: Abuse history and risky sex, or the burdens of non-disclosure? Social Science and Medicine, 128, 366– 373. Sandfort,T.G.M.,Bakker,F.,Schellevis,F.G.,&Vanwesenbeeck,I.(2006). Sexual orientation and mental and physical health status: Findings from a Dutch population survey. American Journal of Public Health, 96, 1119–1125. Sandfort, T. G. M., de Graaf, R., Bijl, R. V., & Schnabel, P. (2001). Same-sex sexual behavior and psychiatric disorders: Findings from the NetherlandsMentalHealthSurveyandIncidenceStudy(NEMESIS).Archives of General Psychiatry, 58, 85–91. Udry, J. R., & Chantala, K. (2002). Risk Assessment of adolescents with same-sex relationships. Journal of Adolescent Health, 31, 84–92. Vanden Berghe, W. V., Dewaele, A., Cox, N., & Vincke, J. (2010). Minority specific determinants of mental well-being among lesbian, gay, and bisexual youth. Journal of Applied Social Psychology, 40, 153–166. Wamala, S., Boström, G., & Nyqvist, K. (2007). Perceived discrimination and psychological distress in Sweden. British Journal of Psychiatry, 190, 75–76. Warner, J., McKeown, É., Griffin, M., Johnson, K., Ramsay, A., Cort, C., & King, M. (2004). Rates and predictors of mental illness in gay men, lesbians and bisexual men and women. British Journal of Psychiatry, 185, 479–485. Widmer, E. D., Treas, J., & Newcombe, R. (1998). Attitudes toward nonmarital sex in 24 countries. Journal of Sex Research, 35, 349–358. Zietsch, B. P., Verweij, K. J. H., Bailey, J. M., Wright, M. J., & Martin, N. G. (2011). Sexual orientation and psychiatric vulnerability: A twin study of neuroticism and psychoticism. Archives of Sexual Behavior, 40, 133– 142. Zietsch, B. P., Verweij, K. J. H., Heath, A. C., Madden, P. A. F., Martin, N. G., Nelson, E. C., & Lynskey, M. T. (2012). Do shared etiological factors contribute to the relationship between sexual orientation and depression? Psychological Medicine, 42, 521–532. Zivony, A., & Lobel, T. (2014). The invisible stereotypes of bisexual men. Archives of Sexual Behavior, 43, 1165–1176.