The Association Between Sexual Orientation Identity and Behavior Across Race/Ethnicity,Sex, and Age in a Probability Sample of High School Students

Objectives. We examined the prevalence and associations between behavioral and identity dimensions of sexual orientation among adolescents in the United States, with consideration of differences associated with race/ethnicity, sex, and age.Methods. We used pooled data from 2005 and 2007 Youth Risk Behavior Surveys to estimate prevalence of sexual orientation variables within demographic sub-groups. We used multilevel logistic regression models to test differences in the association between sexual orientation identity and sexual behavior across groups.Results. There was substantial incongruence between behavioral and identity dimensions of sexual orientation, which varied across sex and race/ethnicity. Whereas girls were more likely to identify as bisexual, boys showed a stronger association between same-sex behavior and a bisexual identity. The pattern of association of age with sexual orientation differed between boys and girls.Conclusions. Our results highlight demographic differences between 2 sexual orientation dimensions, and their congruence, among 13- to 18-year-old adolescents. Future research is needed to better understand the implications of such differences, particularly in the realm of health and health disparities.The Institute of Medicine (IOM) recently called for increased data collection on the lesbian, gay, and bisexual (LGB) community in population health studies.1 Although numerous studies have identified sexual orientation health disparities among youths,2–9 researchers have only more recently begun to question if there is variation in prevalence of health outcomes among LGB populations depending on which dimension of sexual orientation is used (i.e., identity, behavior, or attraction).10 Sexual orientation is a multidimensional construct,11,12 and how it is measured changes the prevalence of nonheterosexual orientations.13,14 Large studies of probability samples of youths typically only include 1 item assessing a single dimension of sexual orientation,15 and, therefore, it is important to understand the level of congruency across these items.The IOM report also called for the use of an intersectional perspective, which is one that recognizes an individual’s co-occurring social identities and how they interact with sexual orientation.1 Along these lines, there is much interest in the basic question of how different dimensions of sexual orientation are interrelated and if these relationships vary across development, sex, and race/ethnicity.11,12,16 Understanding how various dimensions of sexual orientation are interrelated, and if this varies across demographic groups, has important implications for our basic understanding of the development of sexual orientation and its measurement in future adolescent health research.There is a growing body of research on incongruence between sexual orientation identity and sexual behaviors.13,17,18 A Canadian study found that 12% of 1878 adolescents endorsed at least 1 dimension of nonheterosexuality.13 Of these students, discrepancies were evident in their reports of sexual identity, attraction, and behavior; the majority only selected a single item (62%) whereas only 15% endorsed all items. The second-largest group (35%) was youths who reported same-sex attractions but no same-sex behavior or minority identity labels. From a developmental perspective it is important to recognize that sexual attractions typically emerge during early adolescence, whereas sexual behavior and internal adoption of identity labels occur during middle-to-late adolescence.11 These patterns reflect the complexity in researching youths’ sexual orientation as adolescents’ identity may be in flux and opportunities for sexual expression may not exist.11,12A number of studies have found that girls are more likely to adopt both-sex–oriented identities (i.e., bisexual or mostly heterosexual) and to report same-sex attraction and same-sex sexual behavior.13,19,20 Studies have also found lower levels of congruence between identity and behavior among adult women compared with men,21,22 leading some to hypothesize that female sexual orientation is more fluid or plastic.23,24 Studies of sex differences in sexual orientation stability across development have not been fully supportive of this hypothesis as results have been inconsistent, with some studies finding greater stability of identity among sexual-minority girls,25,26 no sex differences among sexual minorities,27 or that it depends on sexual orientation.20There has been limited research on how the experience of being LGB varies across racial or ethnic groups. Lack of significant differences in the timing of psychosexual milestones or in sexual identity formation among Black, Asian, Latino, and White youths has been reported,28–30 whereas others have found differences.31,32 Factors such as internalized homophobia, perceptions of rejection, and limited availability of support resources have been hypothesized to delay the timing of identity labeling and disclosure among racial/ethnic minority youths and thereby potentially produce less concordance among dimensions of sexual orientation.33,34 However, little research has examined how associations among different dimensions of sexual orientation may differ by race/ethnicity among adolescents.There are crucial gaps in our knowledge of the congruence among the dimensions of sexual orientation among youths, which have an impact on our understanding of measurement of sexual orientation and, by extension, characterization of health disparities. This is primarily attributable to the small number of sexual minority individuals in most probability samples, which typically do not permit exploring variation within the LGB sample. The objective of this study was to examine the congruency of sexual orientation identity and behavior across sex, race/ethnicity, and development in a probability sample of adolescents achieved through pooling Youth Risk Behavior Survey (YRBS) data.35