RESEARCH AND PRACTICE Exploring Alcohol-Use Behaviors Among Heterosexual and Sexual Minority Adolescents: Intersections With Sex, Age, and Race/Ethnicity Amelia E. Talley, PhD, Tonda L. Hughes, RN, PhD, Frances Aranda, PhD, Michelle Birkett, PhD, and Michael P. Marshal, PhD Although alcohol consumption by youths has declined in recent years, it remains a major public health problem.1 Underage drinking is associated with a range of physical, academic, and social problems. Youths who drink, especially those who drink heavily, are more likely to engage in delinquent behavior, experience violence and victimization, and commit suicide.1,2 Of great concern is that alcohol consumption is a leading contributor to injury, the main cause of death for people younger than 21 years.1 Early onset drinking is also associated with increased risk for developing an alcoholuse disorder during the lifespan.3 Despite the fact that almost all US youths grow up in a culture permeated by alcohol, the prevalence of early and heavy drinking and its consequences vary across demographic groups. For example, considerable variation exists in alcohol consumption between White and racial/ethnic---minority youths; data consistently show the highest rates of drinking and drinking-related problems among White and American Indian or Alaska Native youths, followed by Hispanic/Latino, Black/African American, and Asian youths.4---6 Non-Hispanic White youths generally start drinking at younger ages than their racial/ethnic minority counterparts. Greater percentages of racial/ ethnic minority youths abstain or drink very little,7 and significant differences exist in levels of drinking between racial/ethnic minority boys and girls.7 Generally, across racial/ethnic groups, prevalence rates of drinking for boys and girls tend to be similar in younger age groups2; among older adolescents, however, more boys than girls engage in frequent and heavy drinking,2 and boys show higher rates of drinking-related consequences.8 Alcohol use and heavy drinking are more prevalent among lesbian, gay, and bisexual (LGB) youths and adults than among their Objectives. We examined sexual orientation status differences in alcohol use among youths aged 13 to 18 years or older, and whether differences were moderated by sex, age, or race/ethnicity. Methods. We pooled data from the 2005 and 2007 Youth Risk Behavior Surveys and conducted weighted analyses, adjusting for complex design effects. We operationalized sexual orientation status with items assessing sexual orientation identity, sexual behavior, sexual attraction, or combinations of these. Results. Compared with exclusively heterosexual youths, sexual-minority youths were more likely to report each of the primary study outcomes (i.e., lifetime and past-month alcohol use, past-month heavy episodic drinking, earlier onset of drinking, and more frequent past-month drinking). Alcohol-use disparities were larger and more robust for (1) bisexual youths than lesbian or gay youths, (2) girls than boys, and (3) younger than older youths. Few differences in outcomes were moderated by race/ethnicity. Conclusions. Bisexual youths, sexual-minority girls, and younger sexual-minority youths showed the largest alcohol-use disparities. Research is needed that focuses on identifying explanatory or mediating mechanisms, psychiatric or mental health comorbidities, and long-term consequences of early onset alcohol use, particularly frequent or heavy use, among sexual-minority youths. (Am J Public Health. 2014;104: 295–303. doi:10.2105/AJPH.2013.301627) heterosexual counterparts, and this is especially true for LGB girls and women.9---12 Because most research on alcohol use among youths has focused on heterosexual youths or has not assessed sexual orientation, little is known about how sexual orientation interacts with other demographic characteristics to influence drinking patterns. Studies comparing sexual-minority and heterosexual adults suggest that drinking patterns of sexual minorities differ in substantial ways from those of the general population. For example, differences in alcohol-use patterns between lesbian or bisexual women and gay or bisexual men are much smaller than those between heterosexual women and men.12---18 Also, rates and patterns of drinking may be more similar among sexual minorities across racial/ethnic minority statuses, compared with their heterosexual counterparts.19 Few studies have examined alcohol-use disparities among subgroups of sexual-minority youths (SMYs; < 18 years). In a meta-analysis, February 2014, Vol 104, No. 2 | American Journal of Public Health Marshal et al.20 found that SMYs had more than twice the odds of ever drinking alcohol, and 3 of the 4 studies that examined alcoholrelated effects for boys and girls separately reported larger alcohol-use disparities among girls than among boys.21---23 Moreover, although studies analyzed by Marshal et al.20 included racially and ethnically diverse samples, none examined the intersecting influences of sexual-minority and racial/ethnic status on substance-use outcomes. The current analyses build on existing literature in 2 ways. First, previous studies with SMYs have examined a limited number of alcohol-use outcomes (e.g., any lifetime drinking).21,22,24 Second, researchers have typically combined subgroups of SMYs25---31 in analyses— often because of small subgroup sample sizes. Because alcohol-use patterns appear to differ on the basis of how sexual orientation is defined—that is, whether data are analyzed on the basis of sexual identity, behavior, attraction, Talley et al. | Peer Reviewed | Research and Practice | 295 RESEARCH AND PRACTICE or some combination of these17,32---34—we extended previous work25,35 by examining a variety of sexual-orientation subgroup differences. In addition to our primary goal of examining sexual-orientation differences in alcohol-use patterns, we conducted moderation analyses to determine whether relations between sexual orientation and drinking patterns varied on the basis of sex, age, or race/ ethnicity. METHODS We used pooled data from 2005 and 2007 Youth Risk Behavior Survey (YRBS) samples. The YRBS study design details and data set characteristics are described elsewhere in this issue.36 The YRBS data provide an opportunity to compare alcohol-use patterns among SMYs and their heterosexual peers in a populationbased school analytic sample. In the current analyses, we used data from 14 US jurisdictions that assessed sexual-orientation identity, sex of sexual partners, or sexual attraction. We excluded youths who did not respond to the sexual-orientation questions or who were 12 years old or in seventh grade. The final sample size (n  37 543) varies across statistical models depending on which sexual-orientation variable is included. Measures Sexual orientation. We used a binary variable, constructed from questions about sexual identity, behavior, and attraction, to categorize participants as SMYs or exclusively heterosexual.36 We considered participants who reported a nonheterosexual identity, any same-sex behavior, or any same-sex attraction to be SMYs. We coded participants who indicated that they were unsure of their sexual identity as missing unless they also reported any same-sex sexual behavior (n = 116; 0.30%), in which case they were coded as SMYs. Otherwise, we considered participants to be exclusively heterosexual. Sexual attraction was only assessed in Hawaii; thus, sexualminority status is the only sexual-orientation variable that distinguishes participants on the basis of attraction. We coded participants as having a sexualmajority identity if they identified as heterosexual or a sexual-minority identity if they identified as lesbian, gay, or bisexual. We coded participants who were unsure of their sexual identity and were not sexually active (n = 273; 1.0%) or did not answer a question about sexual partners (n = 187; 0.7%) as missing. We considered participants who were sexually experienced (n = 294; 1.0%) yet unsure of their sexual identity to be SMYs. We based this decision on preliminary data analyses and previous studies showing heightened risk for substance-use outcomes among individuals who are unsure of their sexual identity.14,17 Because alcohol-use patterns for individuals who report both same-sex and opposite-sex sexual partners may be distinct from those who report only same- or only opposite-sex sexual partners,17,28,37 we constructed 2 dummy variables to distinguish participants who reported only opposite-sex sexual partners from those who reported only same-sex sexual partners and those who reported both male and female sexual partners. We used responses to questions about sexual identity and sex of sexual partners to create 6 nominal sexual-orientation subgroups: we coded 2 groups as lesbian or gay, or bisexual, respectively, on the basis of self-identity. Those who identified as heterosexual or were unsure of their sexual identity yet also reported any same-sex sexual partners constituted the third group. Participants who were unsure of their identity and reported only opposite-sex sexual partners were the fourth group, and those unsure of their identity and not sexually active constituted the fifth group. The sixth, exclusively heterosexual group, included youths who identified as heterosexual and reported that they were not sexually active or reported only opposite-sex sexual partners. We created dummy variables for the first 5 subgroups, with the last serving as the reference group.38 Alcohol outcomes. Participants were asked, “How old were you when you had your first drink of alcohol, other than a few sips?” Response options were (1) I have never had a drink of alcohol other than a few sips, (2) 8 years old or younger, (3) 9---10 years old, (4) 11---12 years old, (5) 13---14 years old, (6) 15---16 years old, and (7) 17 years old or older. We recoded responses to create a variable ranging from 8 to 17, corresponding to the approximate age that alcohol was first consumed. 296 | Research and Practice | Peer Reviewed | Talley et al. We coded participants who never had a drink of alcohol (n = 10 825) as missing. We used responses to the age of drinking onset question to construct a dichotomous variable indicating any versus no lifetime drinking. To create a variable reflecting number of past-month drinking days, we used a question in which participants were asked: “During the past 30 days, on how many days did you have at least one drink of alcohol?” Response options were no days, 1---2 days, 3---5 days, 6---9 days, 10---19 days, 20---29 days, and all 30 days. We used responses to the past-month drinking days question to create a dichotomous indicator of any versus no past-month drinking. To assess heavy episodic drinking, participants were asked on how many days in the past month they drank 5 or more drinks of alcohol within a couple of hours. We recoded response options (0 days, 1 day, 2 days, 3---5 days, 6---9 days, 10---19 days, 20 or more days) to construct a dichotomous variable indicating any versus no past-month heavy episodic drinking. Demographics. We created a dichotomous variable that differentiated younger participants (aged 13---15 years) from older participants (aged 16 to 18 years or older). This variable represented roughly equal-sized age groups and allowed for examination of moderated relationships relevant to an adolescent milestone (obtaining a driver’s license) associated with greater independence and access to alcohol. We used the nominal race/ethnicity variable constructed by Mustanski et al.36 that recategorized 8 racial/ethnic groups into 6 (Asian, Black/African American, Hispanic/Latino, American Indian/Alaskan Native/Native Hawaiian/Pacific Islander, White, and multiple or other ethnicity). We created 5 dummy variables; White youths served as the reference group. Covariates. Our models included covariates shown to be strong correlates of problematic drinking among adolescents in the general population39---42 as well as among sexual minorities.14,43---45 Age of first sexual intercourse was assessed by asking, “How old were you when you had sexual intercourse for the first time?” We created a grand-mean-centered variable with response options ranging from (1) aged 11 years or younger to (8) aged 17 years or American Journal of Public Health | February 2014, Vol 104, No. 2 RESEARCH AND PRACTICE older. We coded sexual assault (0 = no; 1 = yes) on the basis of the question “Have you ever been physically forced to have sexual intercourse when you did not want to?” Data Analysis We conducted descriptive analyses with SPSS version 21 (IBM, Somers, NY); we fit multiple regression models with HLM, version 7 (Scientific Software International Inc, Skokie, IL). We used the complex sample module in SPSS, which takes into account weight, stratum, and primary sampling unit variables, in analyses to adjust for the complex sampling design of the YRBS.36 We used hierarchical linear modeling to account for clustering of the data, with jurisdiction entered at level 2 in each model. We used full-information maximum likelihood estimation. We modeled dichotomous outcomes with a Bernoulli distribution; we modeled number of past-month drinking days with a Poisson distribution, accounting for overdispersion. We estimated separate models for each of the sexual-orientation indices; all models included sex, age, race/ethnicity, and covariates. We conducted directed tests of interaction effects for hierarchical linear models based on results from preliminary ordinary least squares regression analyses. Coefficients associated with the covariates were similar across models examining each of the primary outcomes. Older age of first sexual intercourse was associated with lower likelihood of lifetime drinking (odds ratio [OR] = 0.71; 95% confidence interval [CI] = 0.69, 0.73; P < .001), any past-month drinking (OR = 0.72; 95% CI = 0.70, 0.73; P < .001), and any past-month heavy episodic drinking (OR = 0.72; 95% CI = 0.70, 0.73; P < .001). It was also associated with older age of drinking onset (b = 0.38; SE = 0.01; 95% CI =0.36, 0.40) and fewer past-month drinking days (b = –0.28; SE = 0.002; 95% CI = -0.28, ---0.28). History of sexual assault was associated with greater likelihood of lifetime drinking (OR = 1.82; 95% CI = 1.61, 2.06; P < .001), any past-month drinking (OR = 1.72; 95% CI = 1.55, 1.91; P < .001), and any past-month heavy episodic drinking (OR = 1.74; 95% CI = 1.57, 1.93; P < .001), as well as a younger drinking onset (b = –0.37; SE = 0.06; 95% CI = ---0.49, ---0.25) and more past-month drinking days (b = 0.52; SE = 0.01; 95% CI = 0.50, 0.54; data not shown). Because coefficients were similar across models and because of space limitations, we do not present associations between covariates and study outcomes in other models. Sexual Orientation RESULTS Table 1 summarizes sample characteristics and provides P values for bivariate comparisons among moderator and primary study variables. Results of cross-tabulations between select sexual orientation variables and alcohol outcomes are presented in Table 2. In Table 3, we show main effects between sexual orientation variables and alcohol-use outcomes, as well as interactions with sex and age. Main effects are discussed first, followed by interaction effects. Sexual Minority Status As shown in Table 3, SMYs reported higher rates of each of the outcomes than did exclusively heterosexual youths. Sexual-minority youths were more likely to report lifetime drinking and earlier drinking onset. They were also more likely to report past-month drinking and heavy episodic drinking, as well as more drinking days in the past month. Main effects between sexual-orientation identity and alcohol use replicated those reported previously. Youths who identified as gay or lesbian, or bisexual or who were unsure of their sexual identity yet reported same-sex sexual behavior, were at higher risk than their exclusively heterosexual counterparts for all outcomes. Youths who reported both male and female sexual partners differed from those who reported only opposite-sex sexual partners on all outcomes. It was notable that youths who reported only same-sex sexual partners differed significantly from those with only opposite-sex sexual partners on just 1 outcome (number of past-month drinking days). Gay- or lesbian-identified youths differed from exclusively heterosexual youths only in that they reported more past-month drinking days. By contrast, bisexual-identified youths, and youths who identified as heterosexual or unsure but reported same-sex sexual behaviors had greater odds of lifetime drinking, February 2014, Vol 104, No. 2 | American Journal of Public Health past-month drinking, and past-month heavy episodic drinking. These youths also began drinking at earlier ages and reported a greater number of past-month drinking days. Youths who were unsure of their sexual identity and not sexually active were less likely to report lifetime drinking, past-month drinking, or heavy episodic drinking; they also reported fewer past-month drinking days. Moderated Differences in Alcohol-Use Patterns There were 8 significant interaction effects between sexual orientation indicators and sex. The pattern of effects was robust across the majority of interactions, indicating larger alcohol-use disparities for girls than for boys, on the basis of sexual-minority status. Sexualminority girls reported higher rates of lifetime alcohol use and past-month heavy episodic drinking than did sexual-minority boys, heterosexual girls, or heterosexual boys. More than three quarters (81.3%) of sexualminority girls were lifetime drinkers, compared with 68.9% of sexual-minority boys, 66.9% of heterosexual girls, and 65.6% of heterosexual boys. We found a similar pattern in the interaction with sexual-orientation identity. Nearly one third (30%) of sexualminority girls reported past-month heavy episodic drinking compared with 25.4% of sexual-minority boys, 16.4% of heterosexual girls, and 19.3% of heterosexual boys. Bisexual-identified girls were more likely to be lifetime drinkers than were bisexual-identified boys or heterosexual girls or boys. Likewise, girls with both-sex sexual partners were more likely to be lifetime drinkers than were boys with both-sex sexual partners and boys or girls with only opposite-sex sexual partners. Girls who reported only same-sex sexual partners began drinking at earlier ages than did girls who reported only opposite-sex sexual partners. Finally, compared with heterosexual girls, bisexual-identified girls and girls who were unsure of their sexual identity and not sexually active reported younger drinking onset. For all outcomes except age of drinking onset, participant age interacted with sexualminority status and sexual-orientation identity, respectively. The pattern of interactions indicated that sexual orientation---related Talley et al. | Peer Reviewed | Research and Practice | 297 RESEARCH AND PRACTICE TABLE 1—Sample Characteristics in Analysis of Sexual Orientation Status Differences in Alcohol Use Among Youths Aged 13 to 18 Years or Older: Youth Risk Behavior Surveys, United States, 2005 and 2007 Sex a Race Age d White (n = 13 548), No. (%) Non-White (n = 23 080), No. (%) 13–15 y (n = 15 500), No. (%) 16–‡18 y (n = 21 843), No. (%) 5956 (32.1) 3738 (27.9) 7381 (33.7) 5727 (39.1) 5652 (25.3) 1021 (5.1) 1549 (8.0) 895 (4.6) 1111 (5.9) 698 (4.8) 1827 (8.4) 1162 (7.6) 1385 (5.9) 593 (4.1) 1376 (6.4) 996 (6.6) 3873 (11.0) 1844 (10.4) 1010 (4.4) 2002 (11.7) 1455 (10.8) 2329 (11.3) 1976 (14.0) 1880 (9.2) 13–14 y 8141 (24.6) 15–16 y ‡ 17 y 5918 (18.6) 1004 (3.2) 4429 (27.3) 3663 (22.0) 3608 (27.3) 4370 (21.9) 3577 (26.3) 4548 (23.6) 3172 (19.9) 501 (3.4) 2709 (17.3) 498 (3.1) 2611 (21.7) 414 (3.5) 3195 (15.4) 559 (2.9) 824 (6.5) 0 (0) 5077 (26.3) 1001 (5.3) Never had alcohol 11 433 (30.7) 5408 (29.3) 5956 (32.1) 3738 (27.9) 7381 (33.7) 5727 (39.1) 5652 (25.3) Reported lifetime use 23 553 (69.3) 11 863 (70.7) 11 532 (67.9) 9379 (72.1) 13 657 (66.3) 8536 (60.9) 14 902 (74.7) Total Sample (n = 37 543), No.,b (%) Female (n = 18 471), No. (%) Male (n = 18 822), No. (%) 11 433 (30.7) 5408 (29.3) £8 y 2595 (6.6) 9–10 y 2022 (5.2) 11–12 y Variables Age of drinking onset Never had drink < .001 Lifetime drinking Pc < .001 < .001 Past-mo drinking days 0d Pc < .001 < .001 < .001 < .001 < .001 21 605 (58.4) 10 722 (57.9) 10 782 (59.0) 6920 (51.9) 14 189 (65.5) 9900 (66.7) 11 626 (53.1) 1–2 d 6740 (20.7) 3598 (22.3) 3084 (19.1) 2760 (22.0) 3823 (19.3) 2506 (18.8) 4208 (22.0) 3–5 d 6–9 d 3063 (9.9) 1664 (5.7) 1528 (10.0) 782 (5.6) 1515 (10.0) 870 (5.9) 1546 (12.2) 936 (7.6) 1462 (7.5) 695 (3.7) 997 (7.5) 460 (3.6) 2058 (11.5) 1196 (7.1) 10–19 d 1028 (3.6) 462 (3.4) 560 (3.9) 557 (4.6) 448 (2.5) 292 (2.3) 732 (4.5) 20–29 d 224 (0.7) 83 (0.5) 138 (0.9) 112 (0.8) 102 (0.6) 67 (0.5) 153 (0.9) 30 d 308 (0.8) 76 (0.4) 224 (1.2) 121 (0.8) 172 (0.8) 91 (0.6) 190 (0.9) Past-mo. drinking NS No 21 605 (58.4) 10 722 (57.9) 10 782 (59.0) Yes 13 027 (41.6) 6530 (42.1) 6390 (41.0) Past-mo HED No Yes < .001 6920 (51.9) 14 189 (65.6) 6031 (48.1) 6703 (34.4) < .001 29 385 (77.5) 14 768 (78.6) 14 443 (76.3) 6991 (22.5) 3286 (21.4) 3649 (23.7) Sexual minority statuse 34 430 (91.6) Sexual minority 3113 (8.4) 16 598 (89.7) 17 610 (93.4) 1872 (10.3) 9566 (70.7) 19 111 (84.4) 3723 (29.3) 3121 (15.6) 962 (7.5) 21 030 (90.7) 2050 (9.3) < .001 4413 (33.3) 8537 (46.9) 13 040 (84.6) 16 234 (72.8) 2053 (15.4) 4874 (27.2) 14 366 (92.4) 19 909 (91.0) 1134 (7.6) 1934 (9.0) 10 629 (94.8) 14 684 (93.5) < .001 .003 < .001 25 427 (94.0) 12 411 (92.5) Sexual minority Sex of sexual partners 1597 (6.0) 1006 (7.5) Only opposite sex 14 972 (53.0) 6802 (49.4) 8170 (56.7) 5679 (51.7) 9002 (54.7) 4852 (42.2) Both sex partners 1015 (3.7) 718 (5.3) 284 (2.0) 397 (3.7) 601 (3.6) 354 (3.1) 614 (2.2) 287 (1.9) 327 (2.4) 180 (1.9) 409 (2.4) 181 (1.7) 423 (2.4) 6809 (43.4) 5707 (38.9) 5617 (42.7) 6683 (39.3) 6516 (53.0) 6026 (33.9) No sex partners 575 (4.5) 8875 (94.8) 15 961 (93.4) 489 (5.2) 1062 (6.6) .001 Sexual majority Only same sex 12 847 (95.5) 11 626 (53.1) < .001 12 586 (92.5) 1212 (6.6) Sexual orientation identity < .001 9900 (66.7) < .001 < .001 Sexual majority < .001 12 598 (41.2) Pc 578 (5.2) 995 (6.5) .001 < .001 10 084 (59.7) 649 (4.0) Note. HED = heavy episodic drinking; NS = not significant. a Totals may differ because of missing or excluded cases. b Unweighted frequencies (No.) are provided; percentages (%) reflect adjusted sampling weights. c P value for associated v2 test of independence using adjusted sampling weights. d Non-White defined as African American/Black, Hispanic/Latino, Asian, Native Hawaiian, Pacific Islander, American Indian, Alaska Native, multiple, or other. e Constructed from questions about sexual identity, behaviors, and attractions. 298 | Research and Practice | Peer Reviewed | Talley et al. American Journal of Public Health | February 2014, Vol 104, No. 2 RESEARCH AND PRACTICE TABLE 2—Cross-Tabulations for Sexual Orientation Variables by Alcohol Outcomes Among Youths Aged 13 to 18 Years or Older: Youth Risk Behavior Surveys, United States, 2005 and 2007 Sexual Minority Status,a No. (%) Sexual Orientation Identity, No. (%) Sexual Majority (n = 34 430) Sexual Minority (n = 3113) 10 752 (31.7) 2196 (6.0) 681 (19.4) 399 (12.8) 7829 (30.7) 1648 (6.2) 267 (17.2) 220 (13.5) 2507 (16.4) 1198 (7.3) 9–10 y 1785 (5.0) 237 (7.9) 1361 (5.2) 135 (9.2) 1022 (6.3) 100 (9.0) 53 (8.8) 11–12 y 3463 (10.7) 411 (14.9) 2613 (11.0) 215 (14.4) 1939 (13.2) 168 (17.8) 72 (14.4) 13–14 y 7435 (24.4) 706 (26.9) 5559 (25.3) 381 (26.4) 4040 (30.0) 264 (30.7) 138 (26.5) 15–16 y 5508 (18.9) 410 (15.5) 3986 (18.3) 238 (16.6) 3021 (22.9) 154 (17.1) 85 (16.2) ‡ 17 y 925 (3.3) 78 (2.6) 690 (3.3) 41 (2.6) 517 (3.9) 25 (1.9) 25 (4.5) Variable Sexual Majority (n = 25 427) Sex of Sexual Partners, No. (%) Sexual Minority (n = 1597) Only Opposite Sex (n = 14 972) Both Sexes (n = 1015) Only Same Sex (n = 614) 70 (7.6) 177 (15.8) 131 (18.1) 73 (11.4) Age of drinking onset Never had drink £8 y Lifetime drinking Never had alcohol Reported lifetime use 10 752 (31.7) 21 312 (68.3) 681 (19.4) 2241 (80.6) 7829 (30.7) 15 858 (69.3) 267 (17.2) 1231 (82.8) 2507 (16.4) 11 736 (83.6) 70 (7.6) 889 (92.4) 131 (18.1) 445 (81.9) Past-mo drinking days 0d 20 279 (59.7) 1326 (43.6) 15 168 (60.1) 622 (41.7) 6663 (44.2) 288 (30.7) 236 (41.2) 1–2 d 6143 (20.5) 596 (23.1) 4680 (21.0) 347 (25.1) 3498 (25.9) 217 (24.1) 119 (23.9) 3–5 d 2751 (9.7) 312 (13.0) 2054 (9.6) 168 (12.6) 1788 (14.0) 164 (17.6) 56 (14.1) 6–9 d 1491 (5.5) 174 (8.6) 1049 (5.1) 91 (8.5) 1023 (8.6) 86 (11.0) 35 (10.0) 10–19 d 884 (3.4) 144 (6.5) 613 (3.1) 80 (6.3) 625 (5.4) 68 (8.9) 30 (7.0) 20–29 d 30 d 188 (0.6) 204 (0.6) 36 (1.6) 103 (3.5) 118 (0.5) 143 (0.6) 16 (1.5) 66 (4.4) 124 (1.0) 144 (0.9) 16 (2.1) 65 (5.6) 9 (1.2) 17 (2.6) Past-mo drinking No 20 279 (59.7) 1326 (43.6) 15 168 (60.1) 622 (41.7) 6663 (44.2) 288 (30.7) 236 (41.2) Yes 11 661 (40.3) 1365 (56.4) 8658 (39.9) 769 (58.3) 7202 (55.8) 617 (69.3) 268 (58.8) No 27 294 (78.5) 2091 (66.0) 20 500 (79.6) 1033 (65.0) 10 499 (67.5) 553 (54.7) 395 (64.1) Yes 6135 (21.5) 856 (34.0) 4374 (20.4) 485 (35.0) 4088 (32.5) 411 (45.3) 167 (35.9) Past-mo HED Note. HED = heavy episodic drinking. Unweighted frequencies (No.) are provided; percentages (%) reflect adjusted sampling weights. Age of drinking onset = “How old were you when you had your first drink of alcohol other than a few sips?” (continuous); past-mo. drinking days = “During the past 30 days, on how many days did you have at least one drink of alcohol?” (Poisson with overdispersion factor); lifetime drinking = dichotomous indicator of whether participants had had at least 1 drink of alcohol in their lifetime (binary); past-mo. drinking = dichotomous indicator of whether participants had had at least 1 drink of alcohol in the past month (binary); past-mo. HED = “During the past 30 days, on how many days did you have 5 or more drinks of alcohol in a row, that is, within a couple of hours?” used to create a dichotomous indicator of whether participants had had at least 1 heavy drinking episode in the past month (binary). All v2 test for independence results were significant (P < .001) with adjusted sampling weights. Totals may differ because of missing or excluded cases. a Constructed from questions about sexual identity, behaviors, and attractions. alcohol-use disparities were larger for younger than for older participants. Two thirds (76%) of older SMYs reported lifetime alcohol use—a 4% higher rate than older heterosexual youths. In contrast, the rate of lifetime alcohol use among younger SMYs (74%) was 16% higher than among younger heterosexual youths. We found similar patterns for past-month drinking and heavy episodic drinking. We found only 2 significant interaction effects between sexual orientation and race/ ethnicity in predicting outcomes. The first interaction (OR = 0.65; 95% CI = 0.47, 0.89; P = .008) showed that, similar to White SMYs who reported higher lifetime drinking rates than White exclusively heterosexual youths (79.9% vs 69.1%), Asian SMYs were more likely than their exclusively heterosexual counterparts to report lifetime drinking (54.8% vs 46.2%). The second interaction (b = –0.48; SE = 0.19; 95% CI = -0.85, -0.11), showed that bisexual White and racial/ethnic minority youths initiated drinking at similar ages, whereas exclusively heterosexual racial/ ethnic minorities were significantly younger February 2014, Vol 104, No. 2 | American Journal of Public Health than their White counterparts when they first drank alcohol. DISCUSSION Consistent with findings from previous studies,20,25 SMYs—whether defined on the basis of any sexual minority identity, behavior, or attraction, or on sexual identity alone—were more likely than heterosexual youths to report lifetime drinking and to have initiated drinking at younger ages. Also consistent with previous work,26---28,34,46 SMYs in the current Talley et al. | Peer Reviewed | Research and Practice | 299 RESEARCH AND PRACTICE TABLE 3—Estimated Coefficients From Hierarchical Linear Models in Analysis of Sexual Orientation Status Differences in Alcohol Use Among Youths Aged 13 to 18 Years or Older: Youth Risk Behavior Surveys, United States, 2005 and 2007 Age of Drinking Onset Sexual Orientation Status Differences Lifetime Drinking b SE 95% CI b OR (95% CI) ... ... ... ... ... –0.33 0.06 –0.45, –0.21 0.44 1.55 (1.33, 1.82) –0.56 –0.38 0.57 (0.40, 0.82) 0.68 (0.53, 0.88) Past-Mo. Drinking Days b Past-Mo. Drinking OR (95% CI) Past-Mo. HED RR (95% CI) B B OR (95% CI) ... ... ... ... ... ... 0.37 1.44 (1.41, 1.48) 0.40 1.49 (1.34, 1.65) 0.36 1.44 (1.29, 1.60) –0.45 0.64 (0.60, 0.68) –0.38 0.69 (0.55, 0.86) –0.40 0.55 0.67 (0.54, 0.84) 0.58 (0.46, 0.73) Sexual minority status Sexual majority (Ref) Sexual minority · Sexa,b · Agea,c Sexual orientation identity ... ... ... ... ... ... ... ... ... ... ... –0.27 0.08 –0.43, –0.11 0.66 1.94 (1.63, 2.32) 0.39 1.48 (1.43, 1.53) 0.49 1.63 (1.41, 1.88) 0.41 1.51 (1.30, 1.75) · Sexa,b –0.93 0.39 (0.28, 0.57) · Agea,c –0.62 0.54 (0.37, 0.78) –0.57 0.57 (0.53, 0.61) –0.57 0.57 (0.42, 0.77) –0.60 0.55 (0.40, 0.75) Sexual majority (Ref) Sexual minority Sex of sexual partners OSPs (Ref) SSPs ... –0.05 ... 0.15 ... –0.34, 0.24 · Sexa,b 0.91 0.30 0.32, 1.50 –0.52 0.09 –0.70, –0.34 ... –0.13 ... 0.88 (0.65, 1.19) ... 0.15 ... 1.16 (1.10, 1.23) ... 0.05 ... 1.05 (0.83, 1.33) ... 0.05 ... 1.05 (0.83, 1.33) 0.89 2.45 (1.77, 3.38) 0.44 1.56 (1.50, 1.61) 0.48 1.62 (1.34, 1.95) 0.46 1.59 (1.34, 1.88) –0.80 0.45 (0.24, 0.84) –0.57 0.57 (0.37, 0.87) –0.62 0.54 (0.36, 0.79) · Agea,c Both-sex partners · Sexa,b · Agea,c Sexual orientation subgroups ... –0.13 ... 0.20 ... –0.52, 0.26 –0.39 0.10 –0.56 0.23 –0.32 Not sure, no SP · Sexa,b Heterosexual, no SP or OSP (Ref) Gay or lesbian ... –0.01 ... 0.99 (0.71, 1.39) ... 0.32 ... 1.38 (1.27, 1.51) ... 0.14 ... 1.15 (0.83, 1.58) ... 0.17 ... 1.19 (0.83, 1.70) –0.59, –0.19 1.00 2.73 (2.11, 3.52) 0.46 1.59 (1.52, 1.67) 0.66 1.93 (1.60, 2.33) 0.49 1.63 (1.35, 1.97) –0.01, –0.11 –0.80 0.45 (0.28, 0.73) 0.14 –0.59, –0.05 0.79 2.21 (1.55, 3.15) 0.50 1.66 (1.57, 1.75) 0.56 1.75 (1.36, 2.26) 0.55 1.74 (1.37, 2.21) –0.52 0.31 –1.13, 0.09 –0.93 0.39 (0.27, 0.56) –1.01 0.37 (0.29, 0.47) –1.08 0.34 (0.21, 0.55) –0.79 0.45 (0.25, 0.84) 1.45 0.58 0.31, 2.59 0.01 0.23 –0.44, 0.46 0.38 1.46 (0.89, 2.38) 0.31 1.36 (1.22, 1.52) 0.34 1.35 (0.90, 2.02) 0.22 1.25 (0.82, 1.90) · Sexa,b · Agea,c Bisexual · Sexa,b · Agea,c Heterosexual or not sure, SSP · Sexa,b · Agea,c · Agea,c Not sure, OSP · Sexa,b · Agea,c Note. CI = confidence interval; HED = heavy episodic drinking; OR = odds ratio; OSP = opposite-sex partner; RR = rate ratio; SP = sexual partner; SSP = same-sex partner. All models included sex, age, race/ethnicity, and covariates (age of first sexual intercourse; history of sexual assault). Age of drinking onset = “How old were you when you had your first drink of alcohol other than a few sips?” (continuous); past-mo. drinking days = “During the past 30 days, on how many days did you have at least one drink of alcohol?” (Poisson with overdispersion factor); lifetime drinking = dichotomous indicator of whether participants had had at least 1 drink of alcohol in their lifetime (binary); past-mo. drinking = dichotomous indicator of whether participants had had at least 1 drink of alcohol in the past-month (binary); past-mo. HED = “During the past 30 days, on how many days did you have 5 or more drinks of alcohol in a row, that is, within a couple of hours?” used to create a dichotomous indicator of whether participants had had at least 1 heavy drinking episode in the past-month (binary). Sexual orientation subgroups: gay or lesbian and bisexual subgroups were based on sexual identity, regardless of behavior. Heterosexual or not sure, SSP, were participants who identified as heterosexual or were unsure of their sexual identity yet also reported any same-sex sexual partners. Not sure, OSP, were participants who were unsure of their identity and reported only opposite-sex partners. Not sure, no SP, were those unsure of their identity and not sexually active. Heterosexual, no SP or OSP, were participants who identified as heterosexual and reported that they were not sexually active or only had opposite-sex sexual partners. a Only significant interaction model results are presented. Because of the paucity of interaction effects related to race/ethnicity, these results were not included in Table 3 and only appear in text. b Sex: female = 0; male = 1. c Age: 13–15 y = 0; 16–18 y or older = 1. 300 | Research and Practice | Peer Reviewed | Talley et al. American Journal of Public Health | February 2014, Vol 104, No. 2 RESEARCH AND PRACTICE study were more likely to report any drinking, more frequent drinking, and heavy episodic drinking in the past month. In contrast to other studies that combined youths reporting only same-sex sexual partners with those reporting both-sex sexual partners,25---31,47 we examined these groups separately. Although youths with same-sex sexual partners and those with both-sex sexual partners reported more frequent drinking in the previous month than youths with only opposite-sex sexual partners, youths with only same-sex sexual partners did not differ significantly from those with only opposite-sex sexual partners on any of the other 4 outcomes. However, youths who reported both male and female sexual partners showed consistently higher risk for all outcomes.22,28 By examining diverse subgroups of SMYs, we found insight about subgroups that may be at greater risk for problematic drinking. For example, consistent with Udry and Chantala,22 gay- or lesbian-identified youths in our study did not show heightened risk on most outcomes. That is, gay or lesbian youths did not differ from heterosexual youths on age of drinking onset, lifetime drinking, and pastmonth drinking or heavy episodic drinking, whereas bisexual-identified youths and heterosexual or unsure youths who reported same-sex sexual partners were more likely than exclusively heterosexual youths to report these drinking outcomes. Consistent with previous research findings, which have suggested that heterosexual-identified women who report discordant (i.e., same-sex) sexual behavior are at heightened risk for alcohol-use disorders compared with behaviorally concordant heterosexual women,48 our results provide evidence that discordant sexual identity and behavior may be a marker of risk for alcohol-related problems.49 This finding points to the value of interventions aimed at validating and supporting the development of minority sexual identities, with the ultimate goal of reducing sexual-orientation healthrelated disparities. Findings from the examination of interactions among sexual orientation indicators and sex, age, and race/ethnicity support previous research on alcohol use among SMYs and can be used to inform tailored alcohol-use interventions. For example, our findings are consistent with studies that have shown greater drinking-related disparities among sexualminority females than sexual-minority males.22,23,34 Moreover, current findings show that sexual orientation---related alcohol-use disparities are larger among younger than older adolescents. Although previous research has not explicitly examined how the relation between sexual orientation and alcohol use is moderated by age, findings from longitudinal studies have indicated that alcohol-use disparities may become less or more pronounced during emerging adulthood depending on which subgroup is examined.50,51 Unlike most previous studies in which sample sizes were too small to permit racial/ethnic---group comparisons, we examined differences in alcohol-use patterns across racial/ethnic groups. Our results are consistent with one study of sexualminority adult women19 that found fewer racial/ethnic differences in alcohol use among sexual-minority than among heterosexual women. Findings suggest that racial/ethnic minority status may not provide the same level of protection against alcohol-related problems among sexual minorities as among heterosexuals. Despite numerous strengths, this study has limitations. First, because of content variations in surveys across jurisdictions, we were unable to include data from all jurisdictions in all analyses. Second, we based variables included in the study on single questions, limiting understanding of alcohol-use history and consequences, such as symptoms of alcohol-use disorders. Third, we generally constructed nominal sexual-orientation subgroups from identity categories, combining sexually active and non---sexually active youths. Although we further distinguished some youths (e.g., those unsure of their sexual identity) on the basis of sexual behavior, this was not the case across all identity categories. We chose to present findings by using a variety of operationalizations of sexual orientation to allow for comparisons across models. Fourth, readers should be aware of the inherent potential for response bias in surveys that assess sensitive topics such as sexual orientation and underage alcohol use. Finally, unlike other studies that provide a broader range of sexual-identity response options,50,52,53 YRBS participants chose February 2014, Vol 104, No. 2 | American Journal of Public Health among heterosexual, lesbian or gay, bisexual, or unsure. Intermediate response options, such as mostly heterosexual or mostly lesbian or gay were not included despite assertions that these identities may be particularly salient for SMYs.29,54 A broader range of options could provide valuable, nuanced information about subgroups at greatest risk. Future research is needed to better understand at which points of the sexual-identity development process alcohol use or misuse is most likely to occur. More information is needed about predictors of early onset of alcohol use as well as predictors of frequent and heavy alcohol use among SMYs.51,54 More generally, the field would benefit from further empirical work that focuses on identifying explanatory or mediating mechanisms, such as psychiatric comorbidities, of known alcoholuse disparities. Additional information about long-term consequences of alcohol use, particularly frequent or heavy use, among SMYs is also needed. Findings from our study provide information that can be used in the development of more effective prevention and early intervention strategies to reduce sexual-orientation alcoholrelated disparities. In particular, strategies are needed that focus on delaying the initiation of alcohol use as well as preventing frequent or heavy use. In addition to understanding the overall greater risk of hazardous drinking among SMYs, it is important to know that sexual-minority girls, younger SMYs, and SMYs who are also racial/ethnic minorities may be at particularly high risk for early initiation of drinking and for hazardous drinking. In light of consistent findings that bisexual youths are at particularly high risk for hazardous drinking, additional research is needed to determine what aspects of bisexual orientation (e.g., identity, attraction, or behavior) are most predictive of risk. Ultimately, school- or community-based prevention or intervention programs may benefit from including messages that are inclusive of SMYs and that target these at-risk groups. j About the Authors Amelia E. Talley is with the Department of Psychology, Texas Tech University, Lubbock. Tonda L. Hughes and Frances Aranda are with the Department of Health Systems Science, the University of Illinois at Chicago. Michelle Birkett is with the Department of Medical Social Sciences, Northwestern University, Chicago. Michael P. Marshal is Talley et al. | Peer Reviewed | Research and Practice | 301 RESEARCH AND PRACTICE with the Department of Psychiatry and Pediatrics, the University of Pittsburgh, Pittsburgh, PA. Correspondence should be sent to Amelia E. Talley, PhD, Texas Tech University, Psychology Building, Room 217, Lubbock, TX 79409 (e-mail: amelia.talley@TTU.edu). Reprints can be ordered at http://www.ajph.org by clicking the “Reprints” link. This article was accepted August 13, 2013. Contributors A. E. Talley and T. L. Hughes contributed equally to this article as lead coauthors. All authors contributed extensively to the conceptualization of the research questions and analytic approach. A. E. Talley and F. Aranda culled descriptive statistics. M. Birkett assembled the input data file and conducted the primary inferential statistical tests. A. E. Talley, T. L. Hughes, F. Aranda, and M. P. Marshal aided in the literature review, provided conceptual advice, and contributed to the writing of the article. Acknowledgments This project was supported by a grant from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (award R21HD051178) and by the IMPACT LGBT Health and Development Program at Northwestern University. The authors would also like to acknowledge support to A. E. Talley (K99 AA019974), T. L. Hughes (R01 AA13328), and M. P. Marshal (DA030385; DA026312). Assistance from the Centers for Disease Control and Prevention, Division of Adolescent and School Health, and the work of the state and local health and education departments who conduct the Youth Risk Behavior Survey made the project possible. Note. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, the Centers for Disease Control and Prevention, or any agencies involved in collecting the data. Human Participant Protection Human participant protection approval was unnecessary because Youth Risk Behavior Survey data were de-identified and are publicly available. References 1. Faden VB, Goldman M. The scope of the problem. Alcohol Res Curr Rev. 2005;28(3):111---120. 2. Sommers AR, Sundararaman, R. Congressional Research Service Report to Congress, Alcohol Use Among Youth. New Providence, NJ: Library of Congress; 2007. 3. Grant BF, Dawson DA. Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence: results from the National Longitudinal Alcohol Epidemiologic Survey. J Subst Abuse. 1997;9 (1):103---110. 4. Ma GX, Shive S. A comparative analysis of perceived risks and substance abuse among ethnic groups. Addict Behav. 2000;25(3):361---371. 5. Newcomb MD, Maddahian E, Skager R, Bentler PM. Substance abuse and psychosocial risk factors among teenagers: associations with sex, age, ethnicity, and type of school. Am J Drug Alcohol Abuse. 1987;13(4):413---433. 6. Perkins DF, Borden LM. Positive behaviors, problem behaviors, and resiliency in adolescence. In: Lerner RM, Easterbrooks MA, Mistry J, eds. Developmental Psychology. Handbook of Psychology. 6th ed. Hoboken, NJ: John Wiley & Sons; 2003:373---394. 7. Vega WA, Zimmerman RS, Warheit GJ, Apospori E, Gil AG. Risk factors for early adolescent drug use in four ethnic and racial groups. Am J Public Health. 1993;83 (2):185---189. 8. Stewart C, Power TG. Ethnic, social class, and gender differences in adolescent drinking examining multiple aspects of consumption. J Adolesc Res. 2003;18(6):575--598. 9. Eisenberg M, Wechsler H. Substance use behaviors among college students with same-sex and opposite-sex experience: results from a national study. Addict Behav. 2003;28(5):899---913. 10. Institute of Medicine. The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: National Academies Press; 2011. 11. McCabe SE, Boyd C, Hughes TL, d’Arcy H. Sexual identity and substance use among undergraduate students. Subst Abus. 2003;24(2):77---91. 12. McCabe SE, Hughes TL, Bostwick W, Boyd CJ. Assessment of difference in dimensions of sexual orientation: implications for substance use research in a collegeage population. J Stud Alcohol. 2005;66(5):620---629. 13. Hughes TL. Alcohol use and alcohol-related problems among lesbians and gay men. Annu Rev Nurs Res. 2005;23:283---325. 14. Hughes T, McCabe SE, Wilsnack SC, West BT, Boyd CJ. Victimization and substance use disorders in a national sample of heterosexual and sexual minority women and men. Addiction. 2010;105(12):2130---2140. 15. Cochran SD, Mays VM. Relation between psychiatric syndromes and behaviorally defined sexual orientation in a sample of the US population. Am J Epidemiol. 2000;151(5):516---523. 16. Gilman SE, Cochran SD, Mays VM, Hughes M, Ostrow D, Kessler RC. Risk of psychiatric disorders among individuals reporting same-sex sexual partners in the National Comorbidity Survey. Am J Public Health. 2001;91(6):933---939. 17. McCabe SE, Hughes TL, Bostwick WB, West BT, Boyd CJ. Sexual orientation, substance use behaviors and substance dependence in the United States. Addiction. 2009;104(8):1333---1345. 23. Whitbeck LB, Chen X, Hoyt DR, Tyler KA, Johnson KD. Mental disorder, subsistence strategies, and victimization among gay, lesbian, and bisexual homeless and runaway adolescents. J Sex Res. 2004;41(4):329---342. 24. Ciro D, Surko M, Bhandarkar K, Helfgott N, Peake K, Epstein I. Lesbian, gay, bisexual, sexual-orientation questioning adolescents seeking mental health services. Soc Work Ment Health. 2005;3(3):213---234. 25. Garofalo R, Wolf RC, Kessel S, Palfrey J, DuRant RH. The association between health risk behaviors and sexual orientation among a school-based sample of adolescents. Pediatrics. 1998;101(5):895---902. 26. Bontempo DE, d’Augelli AR. Effects of at-school victimization and sexual orientation on lesbian, gay, or bisexual youths’ health risk behavior. J Adolesc Health. 2002;30(5):364---374. 27. Cochran BN, Stewart AJ, Ginzler JA, Cauce AM. Challenges faced by homeless sexual minorities: comparison of gay, lesbian, bisexual, and transgender homeless adolescents with their heterosexual counterparts. Am J Public Health. 2002;92(5):773---777. 28. Robin L, Brener ND, Donahue SF, Hack T, Hale K, Goodenow C. Associations between health risk behaviors and opposite-, same-, and both-sex sexual partners in representative samples of Vermont and Massachusetts high school students. Arch Pediatr Adolesc Med. 2002; 156(4):349. 29. Austin SB, Ziyadeh N, Fisher LB, Kahn JA, Colditz GA, Frazier AL. Sexual orientation and tobacco use in a cohort study of US adolescent girls and boys. Arch Pediatr Adolesc Med. 2004;158(4):317---322. 30. Caldwell LL, Kivel BD, Smith EA, Hayes D. The leisure context of adolescents who are lesbian, gay male, bisexual and questioning their sexual identities: an exploratory study. J Leis Res. 1998;30(3):341---355. 31. Rostosky SS, Owens GP, Zimmerman RS, Riggle EDB. Associations among sexual attraction status, school belonging, and alcohol and marijuana use in rural high school students. J Adolesc. 2003;26(6):741---751. 32. Hughes TL, Szalacha LA, Johnson TP, Kinnison KE, Wilsnack SC, Cho Y. Sexual victimization and hazardous drinking among heterosexual and sexual minority women. Addict Behav. 2010;35(12):1152---1156. 33. Talley AE, Sher KJ, Littlefield AK. Sexual orientation and substance use trajectories in emerging adulthood. Addiction. 2010;105(7):1235---1245. 18. Sandfort TG, Graaf R, Bijl RV, Schnabel P. Same-sex sexual behavior and psychiatric disorders: finding from the Netherlands Mental Health Survey and Incidence Study (NEMESIS). Arch Gen Psychiatry. 2001;58(1):85---91. 34. Ziyadeh NJ, Prokop LA, Fisher LB, et al. Sexual orientation, gender, and alcohol use in a cohort study of US adolescent girls and boys. Drug Alcohol Depend. 2007;87(2-3):119---130. 19. Hughes TL, Wilsnack SC, Szalacha LA, et al. Age and racial/ethnic differences in drinking and drinkingrelated problems in a community sample of lesbians. J Stud Alcohol. 2006;67(4):579---590. 35. Faulkner AH, Cranston K. Correlates of same-sex sexual behavior in a random sample of Massachusetts high school students. Am J Public Health. 1998;88 (2):262---266. 20. Marshal MP, Friedman MS, Stall R, et al. Sexual orientation and adolescent substance use: a meta-analysis and methodological review. Addiction. 2008;103(4): 546---556. 36. Mustanski B, Van Wagenen A, Birkett M, Eyster S, Corliss HL. Identifying sexual orientation health disparities in adolescents: analysis of pooled data from the Youth Risk Behavior Survey, 2005 and 2007. Am J Public Health. 2014;104(2):211---217. 21. Noell JW, Ochs LM. Relationship of sexual orientation to substance use, suicidal ideation, suicide attempts, and other factors in a population of homeless adolescents. J Adolesc Health. 2001;29(1):31---36. 22. Udry JR, Chantala K. Risk assessment of adolescents with same-sex relationships. J Adolesc Health. 2002;31 (1):84---92. 302 | Research and Practice | Peer Reviewed | Talley et al. 37. Midanik LT, Drabble L, Trocki K, Sell RL. Sexual orientation and alcohol use: identity versus behavior measures. J LGBT Health Res. 2007;3(1):25---35. 38. Cohen J, Cohen P, West SG, Aiken LS. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. 3rd ed. Hillsdale, NJ: Erlbaum; 2003. American Journal of Public Health | February 2014, Vol 104, No. 2 RESEARCH AND PRACTICE 39. Behnken MP, Le YCL, Temple JR, Berenson AB. Forced sexual intercourse, suicidality, and binge drinking among adolescent girls. Addict Behav. 2010;35(5):507-- 509. 40. Coker AL, Richter DL, Valois RF, McKeown RE, Garrison CZ, Vincent ML. Correlates and consequences of early initiation of sexual intercourse. J Sch Health. 1994;64(9):372---377. 41. Robertson JA, Plant MA. Alcohol, sex and risks of HIV infection. Drug Alcohol Depend. 1988;22(1-2):75---78. 42. Thompson NJ, McGee RE, Mays D. Race, ethnicity, substance use, and unwanted sexual intercourse among adolescent females in the United States. West J Emerg Med. 2012;13(3):283---288. 43. Drabble L, Trocki KF, Hughes TL, Korcha RA, Lown AE. Sexual orientation difference in the relationships between victimization and hazardous drinking among women in the National Alcohol Survey. Psychol Addict Behav. 2013; Epub ahead of print. 44. Hughes TL, Johnson TP, Wilsnack SC, Szalacha LA. Childhood risk factors for alcohol abuse and psychological distress among adult lesbians. Child Abuse Negl. 2007;31(7):769---789. 45. Johnson TP, Hughes TL, Cho YI, Wilsnack SC, Aranda F, Szalacha LA. Hazardous drinking, depression, and anxiety among sexual-minority women: selfmedication or impaired functioning? J Stud Alcohol Drugs. 2013;74(4):565. 46. Brewster KL, Tillman KH. Sexual orientation and substance use among adolescents and young adults. Am J Public Health. 2012;102(6):1168---1176. 47. Smith AMA, Lindsay J, Rosenthal DA. Same-sex attraction, drug injection and binge drinking among Australian adolescents. Aust N Z J Public Health. 1999;23 (6):643---646. 48. Gattis MN, Sacco P, Cunningham-Williams RM. Substance use and mental health disorders among heterosexual identified men and women who have same-sex partners or same-sex attraction: results from the National Epidemiological Survey on Alcohol and Related Conditions. Arch Sex Behav. 2012;41(5):1185---1197. 49. Smith A, Rissel CE, Richters J, Grulich AE, Visser RO. Sex in Australia: sexual identity, sexual attraction and sexual experience among a representative sample of adults. Aust N Z J Public Health. 2003;27(2):138---145. 50. Corliss HL, Rosario M, Wypij D, Fisher LB, Austin SB. Sexual orientation disparities in longitudinal alcohol use patterns among adolescents: findings from the Growing Up Today Study. Arch Pediatr Adolesc Med. 2008;162(11):1071---1078. 51. Hatzenbuehler ML, Corbin WR, Fromme K. Trajectories and determinants of alcohol use among LGB young adults and their heterosexual peers: results from a prospective study. Dev Psychol. 2008;44(1):81---90. 52. Hughes T, Szalacha LA, McNair R. Substance abuse and mental health disparities: comparisons across sexual identity groups in a national sample of young Australian women. Soc Sci Med. 2010;71(4):824---831. 53. McCabe SE, Hughes TL, Bostwick W, Morales M, Boyd CJ. Measurement of sexual identity in surveys: implications for substance abuse research. Arch Sex Behav. 2012;41(3):649---657. 54. Rosario M, Schrimshaw EW, Hunter J. Predictors of substance use over time among gay, lesbian, and bisexual youths: an examination of three hypotheses. Addict Behav. 2004;29(8):1623---1631. February 2014, Vol 104, No. 2 | American Journal of Public Health Talley et al. | Peer Reviewed | Research and Practice | 303 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Clin Child Fam Psychol Rev (2014) 17:1–18 DOI 10.1007/s10567-013-0135-1 Growing Up Wired: Social Networking Sites and Adolescent Psychosocial Development Lauren A. Spies Shapiro • Gayla Margolin Published online: 4 May 2013  Springer Science+Business Media New York 2013 Abstract Since the advent of social networking site (SNS) technologies, adolescents’ use of these technologies has expanded and is now a primary way of communicating with and acquiring information about others in their social network. Overall, adolescents and young adults’ stated motivations for using SNSs are quite similar to more traditional forms of communication—to stay in touch with friends, make plans, get to know people better, and present oneself to others. We begin with a summary of theories that describe the role of SNSs in adolescents’ interpersonal relationships, as well as common methodologies used in this field of research thus far. Then, with the social changes that occur throughout adolescence as a backdrop, we address the ways in which SNSs intersect with key tasks of adolescent psychosocial development, specifically peer affiliation and friendship quality, as well as identity development. Evidence suggests that SNSs differentially relate to adolescents’ social connectivity and identity development, with sociability, self-esteem, and nature of SNS feedback as important potential moderators. We synthesize current findings, highlight unanswered questions, and recommend both methodological and theoretical directions for future research. Keywords Adolescent psychosocial development
Social networking sites
Friendships
Identity
Self-esteem L. A. Spies Shapiro (&)
G. Margolin Psychology Department-SGM 922, University of Southern California, 3620 McClintock, Los Angeles, CA 90089-1061, USA e-mail: lspies@usc.edu Introduction Although computers initially were developed for adults, adolescents have fully embraced these technologies for their own social purposes and typically are the family experts on how to use electronic media and social networking sites (SNSs). Adolescents and young adults initially dominated SNSs such as MySpace and Facebook, with parents often following their children into this youthdriven phenomenon. The preponderance of adolescents has access to and engages in use of SNSs: Based on relatively recent data, although perhaps presently an underestimate, 73 % use social networking sites (Lenhart 2009, 2012; Lenhart et al. 2010). Moreover, despite the terms of service of Facebook restricting its use to those age 13 or older, it is estimated that 7.5 million younger children also have accounts (‘‘That Facebook Friend’’ 2011). The sheer amount of time that adolescents and young adults spend using electronic media is perhaps the most revealing: on average, 11–18 year olds spend over 11 h per day exposed to electronic media (Kaiser Family Foundation 2010). Late adolescents and emerging adults average approximately 30 min per day just on Facebook alone (Pempek et al. 2009). Many adolescents begin and end their day by checking SNS posts. Furthermore, SNS use commonly disrupts adolescents’ solitary activities as well as their ongoing face-to-face interactions. The presence of SNS use in many adolescents’ lives is thus indisputable; however, the impact on adolescents’ individual development and social lives is only starting to be understood. Scientific study of adolescence has long targeted the development of one’s identity and the formation of friendships and peer relationships as important topics of study (Institute of Medicine 2010). Two of the key tasks in adolescence are ‘‘to stand out—to develop an identity and 123 2 pursue autonomy… and to fit in—to find comfortable affiliations and gain acceptance from peers’’ (Brown 2008). Although seemingly divergent goals, the interplay between the need for one’s own personal identity and the need for close personal ties and strong group affiliations permeates all domains of adolescents’ everyday lives (Crosnoe and Johnson 2011) and clearly intersects with SNS use. The literature on SNSs and adolescents’ quest to fit in examines whether SNS use extends and deepens adolescents’ ongoing relationships or expands their contacts in new directions. Whereas childhood friendships are rooted in shared interests and activities, close friendships in adolescence involve trust, self-disclosure, and loyalty (Collins and Steinberg 2006; Brown and Larson 2009). SNSs potentially offer additional avenues for support and communication— crucial to the development of age-appropriate adolescent relationships; yet, there are questions to be addressed about why adolescents might differentially benefit from SNSs. Social networking sites offer adolescents new opportunities as well as new challenges to express to the world who one is. In one-on-one communications within SNSs (e.g., ‘‘Facebook messages’’), adolescents can express their likes and dislikes as well as their worldviews and get immediate feedback. With SNSs, adolescents express their views and the recipients of this information include both known as well as unknown targets. Although there has been variability over time in the specific format of SNS profiles, adolescents have the option of choosing what self-identifying information to provide. Thus, with the advent of SNSs, most adolescents will widely share, with varying degrees of accuracy, honesty, and openness, information that previously would have been private or reserved for select individuals. Key questions include whether adolescents accurately portray their identities online, and whether use of SNSs might impact adolescents’ identity development. Straddling these two developmental tasks, adolescents also can join Internet ‘‘groups’’ reflecting the aspects of their identity that they wish to explore or deepen. Thus, SNSs may simultaneously amplify dimensions of selfidentify and extend group identities. Moreover, SNSs create more publicly prominent avenues for adolescents to commit to preferred activities, groups, and, in some cases, beliefs. Social comparison is another dimension of SNSs that is highly relevant to adolescents. Invitations to social gatherings, such as a spontaneous party, and good news, such as a won football game or a college acceptance, can be shared and congratulated but also serve as a point of comparison for one’s own accomplishments. Similarly, distressing or objectionable information—including unflattering and compromising pictures, untrue information, or unfortunate news, for example, a car accident or an arrest—can spread throughout adolescents’ social network and beyond in a 123 Clin Child Fam Psychol Rev (2014) 17:1–18 nanosecond. Teens’ tendencies to share information impulsively, coupled with the power of SNSs for rapid and widely distributed communications, can have important ramifications for teens’ personal and interpersonal worlds. Goals and Scope of the Present Paper This literature review examines how SNSs intersect with and impact adolescents’ social and identity development. After first summarizing theoretical perspectives that provide a framework for SNS use and implications for adolescents, we then review the extant literature on SNSs and (a) adolescent social relationships, as well as (b) identity development. Finally, we offer suggestions for future directions, which call for more nuanced investigations of SNSs that focus less on positive versus negative impacts and more on the mechanisms by which SNSs both reflect and shape varied dimensions of adolescents’ lives. Beyond the Scope Here There are also several important topics in the SNS literature that deserves brief mention because of their relevance for adolescent development but are outside of the scope of this paper. Specifically, these topics include cyber-bullying, the Internet and parent–child relationships, Internet addiction, and the impact of SNSs on sleep and academic performance. Cyber-Bullying Considerable attention has been directed to the negative side of SNSs, namely online bullying, harassment, and humiliation, which have been extensively detailed in other reviews (Strom and Strom 2005; Tokunaga 2010). Beyond findings that the vast majority of adolescents encounter some degree of negative experience through SNSs, the viciousness of online bullying is exacerbated due to the depersonalized yet public nature of technology-based postings coupled with the pervasiveness of SNSs (Bazelon 2013). SNSs and Parent–Child Relations Social networking site use further complicates parent–child relationships during adolescence (Mesch 2003, 2006; Subrahmanyam and Greenfield 2008; Punamaki et al. 2009). As adolescents’ preoccupation with SNSs potentially takes priority over and interferes with everyday family activities, SNS use among adolescents has been linked with greater parent–youth conflict (Subrahmanyam and Greenfield 2008; Mesch 2006) and less time with parents (Lee 2009). Although some parents are unaware of what their teens are Clin Child Fam Psychol Rev (2014) 17:1–18 posting online, other parents utilize SNSs to maintain greater contact with their teens, requiring them to be connected to them through their SNS of choice (Kanter et al. 2012). It is worth noting that changes in the parent–child relationship associated with SNSs are likely to influence peer relations and vice versa. However, the extant literatures on how computer-mediated communications impact these two domains are relatively distinct with only a few exceptions (e.g., Punamaki et al. 2009; Subrahmanyam and Greenfield 2008). Internet Addiction A note about extreme use of the Internet is in order, as distinctions often are blurred between Internet addiction and subthreshold, albeit heavy use, of SNSs. Extreme degrees of Internet and electronic media use are increasingly recognized as Internet addiction, a disorder with symptoms that are analogous to those of substance use and gambling disorders. Findings from epidemiological studies of Internet addiction in youth vary, with prevalence rates ranging widely from less than one percent to 38 % (Aboujaoude 2010; Leung 2004). Some adolescents may be more vulnerable to develop symptoms of Internet addiction than others, including those experiencing other psychological symptoms and disorders such as depression, ADHD symptoms, or hostility (Ha et al. 2007; Yen et al. 2007). Researchers recently proposed the concept of ‘‘Facebook addiction’’ and developed a scale to measure the symptoms of addiction related to Facebook use specifically (Andreassen et al. 2012). Sleep Disturbance Adolescents’ use of the computer, including use of computer-mediated communication, has been related to disruptions in sleep. A study of computer use in relation to adolescents’ sleep quality, perceived health, and tiredness upon awakening found that for young adolescent boys, intensive computer use was associated with less sleep and more irregular sleep, which in turn related to poorer perceived health (Punamaki et al. 2007). Similarly, a sample of high school seniors with Internet addiction and overuse reported greater daytime sleepiness (Choi et al. 2009). Research demonstrates that the use of computers before bed relates to sleep disruptions in adolescents (see Cain and Gradisar 2010 for review). Academic Disturbance There is recent evidence that SNS use can also hinder academic performance (Huang and Leung 2009; Jacobsen 3 and Forste 2011; Kirschner and Karpinski 2010). Proposed mechanisms for the link between SNS use and lower academic performance include less total time studying as well as inefficient studying due to multitasking (Jacobsen and Forste 2011; Junco and Cotten 2012; Kirschner and Karpinski 2010) and could also reflect the sleep disturbance. College students who used Facebook had lower GPAs and spent less time studying than those who did not use Facebook (Kirschner and Karpinski 2010), despite no differences regarding total time spent online. Another study showed that two-thirds of students reported using electronic media during class, while studying, or while doing homework, with amount of electronic media use negatively associated with self-reported GPA (Jacobsen and Forste 2011). Some recent evidence shows that Facebook use specifically relates to lower college GPAs (Junco and Cotten 2012), whereas other studies indicate that computer use is detrimental to the academic performance of some, but not all adolescents (Hofferth and Moon 2011). Theories Relating SNSs to Psychosocial Development Theories examining SNSs and adolescent development address for whom and under what circumstances SNSs accord advantages versus disadvantages for adolescent development. Two theoretical questions in particular are examined here in order to conceptualize how SNSs impact adolescents’ social connectivity as well as their identity development. In What Ways Does SNS Use Advance the Goal of Establishing Close Interpersonal Ties for Adolescents? Some theories contend that SNS use is generally beneficial for the enhancement of adolescents’ social connections. For example, the stimulation hypothesis (McKenna and Bargh 2000) describes how adolescents in general have an easier time self-disclosing in online versus face-to-face communication, which is a less threatening format in which adolescents can share more freely. With self-disclosure facilitating relationship closeness, this theory also posits that online communications lead to closer, higher-quality friendships among adolescents. Second, the rich-get-richer hypothesis posits a stratified advantage for SNS use, that is, for highly sociable adolescents, there are added benefits from extending options for communication through electronic means (Kraut et al. 2002), and iterative effects such that more online communication relates to more cohesive relationships overall (Lee 2009). However, it is also hypothesized that individuals with limited offline social networks and poor social skills do not develop quality friendships through online connections and may spend time 123 4 engaging in low-quality connections in lieu of cultivating relationships in real life. It further has been suggested that spending excessive amounts of time on SNSs actually can lead to symptoms of depression, which then increase the risk for social isolation (O’Keeffe et al. 2011). Thus, these theoretical perspectives provide potential frameworks for hypotheses about who benefits most from SNS use and who might experience SNS use as detrimental. Another perspective proposes differential impacts associated with SNS use but actually gives the advantage to those who are disenfranchised in face-to-face communications. The social compensation hypothesis (McKenna et al. 2002) proposes that adolescents who are uncomfortable interacting with peers in face-to-face contexts are better able to develop social networks and meet their social needs online where certain channels of communication, including voice tone, eye contact, and facial expressions, are not available. That is, the more limited number of communication channels of SNSs may offer unique benefits to those who are uncomfortable with face-to-face interaction, whereas others do not directly benefit (McKenna et al. 2002). All three of the theories mentioned thus far focus on relationship benefits as contrasted with the earlier, and largely discredited reduction hypothesis, stating that forming friendships with strangers online that are low in quality detracts from time spent cultivating pre-existing offline friendships (Locke 1998). This earlier theory, however, emerged in response to Internet use more generally, before the advent of SNSs, and before large numbers of adolescents had access to the Internet. In What Ways Can SNS Use Foster Identity Development for Adolescents? There are two dimensions of SNS use that may contribute to adolescents’ development of self-identity. First, SNS use provides opportunities for self-disclosure and, in some circumstances, demands self-disclosure, which plays a role in adolescents’ identity development. Decisions about how adolescents identify themselves, the feedback received on these decisions, and how they view their own profile in comparison with others’ profiles are potential factors in individual identity. The hyperpersonal model for computer-mediated communication, for example, posits that adolescents engage in selective self-presentations online; moreover, the feedback from these presentations may, in turn, alter individuals’ self-perceptions (Walther et al. 2011). Second, the Internet makes it feasible for some adolescents to affiliate with other, likeminded individuals online when such opportunities may not be possible in face-to-face interaction. The Internet allows adolescents to 123 Clin Child Fam Psychol Rev (2014) 17:1–18 make connections with persons like themselves, that is, ethnic or sexual minority youth (Larson et al. 2002), particularly if such affiliations are not available through local peer networks. Adolescents can join ‘‘groups’’ reflecting aspects of their identity that they wish to explore or deepen and thereby foster a group identity. Relatedly, adolescents can explore and expand their ideas and interests into new arenas through the Internet, for example, communicating with others from more diverse backgrounds and expanding into new intellectual, political, and social networks that create opportunities for transnational and global connections (Markstrom 2010). Such connections can broaden as well as deepen self-identity while, at the same time, enhance feelings of belongingness and affiliation. Literature Review Method for Review To examine the intersection of SNS use and adolescent development, we conducted a search on PsycInfo and Google Scholar using several inclusion criteria. First, we searched for articles examining the use of SNSs from 2006 to the present, as 2006 is the year that Facebook opened to any individual over 13 who had a valid email address (Abram 2006). Keywords searched included the combinations of ‘‘adolescent’’ with ‘‘Internet communication,’’ ‘‘electronic communication,’’ ‘‘social networking site,’’ ‘‘computer-mediated communication,’’ ‘‘Facebook,’’ and ‘‘MySpace,’’ in conjunction with search terms related to the key tasks of adolescent development, including ‘‘peer relationships,’’ ‘‘friendship quality,’’ ‘‘identity,’’ ‘‘intimacy,’’ and ‘‘autonomy.’’ We also examined related articles from reference lists of the resulting studies from the above searches. In this literature review, we include articles that address the relationship between SNS use and tasks of adolescent development, focusing on peer relationship and identity development, specifically with an emphasis on studies including adolescent samples. In addition, select articles examining college samples were included that link SNS use to outcomes relevant to adolescent development. Because this literature is growing at an unusually rapid pace (Wilson et al. 2012), our review identifies and synthesizes representative articles of the present topics of review. With the frequent introduction of new technology and applications, characteristics of SNSs also change rapidly. Whereas ‘‘MySpace’’ was once the SNS of choice and the subject of early research (Kujath 2011), this SNS is rarely used today. Facebook is now the SNS of choice (Chubb 2010), but this, too, is potentially losing popularity (Guynn Clin Child Fam Psychol Rev (2014) 17:1–18 and Faughnder 2012) with other SNSs briefly taking hold, for example, Formspring, an innovative, more anonymous SNS that originated in 2009 and then shut down in spring 2013. In particular for this research domain, the rapidly evolving modifications in technology and consequent alterations in adolescents’ use of the technology present challenges when designing, conducting, and comparing studies on SNS use (Wilson et al. 2012). Different findings not only are attributable to different research methods but also to changes in SNSs and their functionality. Measurement of Adolescents’ Use of and Response to SNSs Self-Report The majority of studies to date examining adolescents’ use of SNSs are based on investigator-developed questions to elicit respondents’ self-report of SNS behaviors. Behaviors most commonly assessed include frequency of use, with questions typically inquiring about general use on average (e.g., Pempek et al. 2009; Reich et al. 2012), duration of use (e.g., Chou and Edge 2012), and, less frequently, time of use, for example, ‘‘after lights out’’ (Van den Bulck 2007). Internet use and related behaviors are sometimes measured with the Internet Addiction Test, a 20-item scale assessing compulsive use, mood changes, and impairment of functioning due to Internet use (IAT; Young 1998, for example, ‘‘Do you feel depressed, moody, or nervous when you are offline, which goes away once you are back online?’’). Shorter self-report questionnaires also are available including Morahan-Martin and Shumacher’s (2000) 13-item scale that assesses similar issues, including distress, academic decline, and interpersonal problems related to Internet use. Of note, these measures assess Internet use in general and do not single out SNS use. Several investigators (e.g., Kirschner and Karpinski 2010; Punamaki et al. 2009) acknowledge potential limitations associated with possible self-report biases. For example, questions assessing adolescents’ number of Facebook friends may inadvertently pull for inflated answers because of adolescents’ desire to appear more popular. Adolescents also might underestimate the degree to which SNS use interferes with their daily activities, similar to the underreporting of other problem behaviors. Furthermore, it may be difficult for adolescents to report how much time they spend on SNSs, particularly if they are multitasking with homework, watching television, or even eating dinner with their family; yet, self-report questionnaires do not always assess the context in which SNS use is taking place. Adolescents also may keep their profile page open throughout the day even though their activities on SNSs may occur in bursts. 5 Experimental Studies Other studies used experimental conditions that manipulate some feature of SNSs to investigate the impacts of that feature. In a study designed to capture the public nature of SNS use, Gonzales and Hancock (2011) asked participants to complete questionnaires either in front of a mirror or in front of their Facebook profile page. Thomaes et al. (2010) manipulated the feedback participants received (positive versus negative) in response to a personal profile they created on the Internet as part of a game, and self-esteem was measured at three points during the laboratory procedure. After creating a profile, participants were exposed to feedback from confederates judging their profiles. Haferkamp and Kramer (2011), in contrast, highlighted the social comparison aspect of SNSs by testing the effects of viewing others’ SNS profile pages on individuals’ body image and career satisfaction; these investigators presented participants with online profiles depicting those who were either attractive or unattractive and those with either high or low occupational success. These standardized, simulated online interactions are informative in isolating precise features of SNSs, although perhaps fall short on ecological validity, particularly compared to research that examines actual records of adolescents’ SNS use (Forest and Wood 2012; Tynes et al. 2008). Objective Assessments of SNS Use A small but growing number of SNS studies objectively examine the specific content of SNS interactions, including content from adolescents’ Facebook postings. A study that utilized individuals’ ten most recent Facebook posts (as reported by the participant, not obtained from the profile page) involved systematic coding of the posts for positivity, negativity, and the amount of ‘‘likes’’ that the posts received (Forest and Wood 2012). In general, public profiles on Facebook allow for the observation of certain dimensions related to teen communication. Nonetheless, questions can be raised about the accuracy of information and the influence of self-presentational guidelines on Facebook content. Moreover, teens are increasingly encouraged to use privacy settings to restrict the information that is publicly available. Summary of Findings: SNS Use, ‘‘Fitting in,’’ Acceptance and Affiliation Table 1 presents the empirical research examining the association between SNS use and adolescent peer relationships. We include 13 representative studies that describe the nature of adolescents’ SNS use and answer questions about ways in which SNS use is associated with 123 6 Clin Child Fam Psychol Rev (2014) 17:1–18 Table 1 Social networking site use, fitting in, acceptance, and affiliation Authors (year) Sample characteristics Measures Primary results Ellison et al. (2007) N = 286 Undergraduate students Affiliation measures Adapted measures of bonding, bridging, and maintained social capital (self-report) SNS measures Facebook usage intensity, Facebook profile elements, purpose of Facebook use (self-report) Ellison et al. (2011) N = 450 Undergraduate students Affiliation measures Measures of bridging social capital (6-item selfreport questionnaire) and bonding social capital (5-item self-report questionnaire) SNS measures Facebook use, number of Facebook friends (‘‘total’’ vs. ‘‘actual’’), Facebook connection strategies measure (self-report) Forest and Wood (2012) N = 80 (Study 1) N = 177 (Study 2) N = 98 (Study 3) Undergraduate Facebook users Affiliation measures Likeability of Facebook user Positivity and negativity of status updates (coded) Self-esteem (self-report) SNS measures 10 most recent status updates (self-report) Number of ‘‘likes’’ and comments on status updates (self-report) Grieve et al. (2013) N = 344 (Study 1) N = 274 (Study 2) Australian university Facebook users Age M = 28.12 Affiliation measures Social Connectedness Scale-revised (20-item self-report questionnaire). SNS measures Facebook social connectedness (20-item self-report questionnaire) Kwon and Wen (2010) N = 229 Users of Korean SNSs Age: 66.2 % in their 20s, 21.0 % in their 30s N = 487 Teens with household telephones and a SNS profile Age 12–17 Affiliation measures Social identity (self-report) SNS measures Perceived ease of use, perceived encouragement, and perceived usefulness of SNSs (self-report) Intensity of Facebook use relates to greater perceived bridging social capital after adjusting for demographic factors, b = .34, p \ .0001, as well as to greater bonding social capital, b = .37, p \ .001. Students with lower self-esteem, b = .34, p \ .0001, and general life satisfaction, b = .31, p \ .0001, perceived greater bridging social capital with greater intensity of Facebook use. Analogous results were found for bonding social capital, b = .34, p \ .001 (life satisfaction), b = .37, p \ .0001 (self-esteem). Greater social information-seeking behaviors on Facebook related to greater perceived bridging social capital, b = .22, p \ .0001, and bonding social capital, b = .18, p = .0006. There were diminishing returns for those with high numbers of actual friends on Facebook, demonstrating a curvilinear relationship between actual Facebook friends ([500) and types of social capital. Youth with low self-esteem viewed Facebook as a safer place for self-expression than did youth with high self-esteem, b = -.31, p = .005. Youth with low self-esteem had higher coded negativity, b = -.31, p = .001, and lower positivity, b = .26, p = .004, in their Facebook status updates in comparison with youth with high self-esteem. Youth with low self-esteem were rated as less likeable by coders than those with high selfesteem, b = .22, t(71) = 2.01, p = .048. A factor analysis revealed that Facebook connectedness is distinct from social connectedness. Facebook facilitates social connections and relates to lower depression, r = -.22, p \ .001 and anxiety, r = -.17, p \ .001, and greater life satisfaction, r = .26, p \ .001. Social identity was positively related to perceived SNS usefulness, b = 6.03, p \ .01, which in turn related to greater SNS use, b = 3.95, p \ .01. Social identity related positively to perceived encouragement via SNSs, b = 2.81, p \ .01. Lenhart and Madden (2007) McMillan and Morrison (2006) 123 N = 72 College students Age 19–25 Affiliation measures Interview questions related to social connectivity online SNS measures Interview questions assessing use of SNSs, including motivations for use Affiliation measures Autobiographical narratives coded for building and forming social relationships online SNS measures Autobiographical narratives coded for use of computer-mediated communication and feelings toward computer-mediated communication 82 % of teens reported using SNSs to send private messages to friends. 91 % of teens reported using SNSs to stay in touch with friends they see frequently. 72 % of teens reported using SNSs to make plans with friends. The coded narratives revealed that participants viewed computer-mediated communication as something that helped them form bonds with others. The narratives also underlined the view of computer-mediated communication as something that facilitates participation in various activities, including special interest groups. Clin Child Fam Psychol Rev (2014) 17:1–18 7 Table 1 continued Authors (year) Sample characteristics Measures Primary results Pempek et al. (2009) N = 92 Undergraduate students Age M = 20.6 (1.07) Participants reported using Facebook for 27.9 min on average per day. Coded diary entries revealed that 85 % of students reported using Facebook to communicate with friends. Participants reported viewing others’ profiles and pictures more often than posting information or updating their own profiles. Quinn and Oldmeadow (2013) N = 443 Age 9–13 Primary and secondary school students north of England Reich et al. (2012) N = 251 High school students in Los Angeles Age 13–19 M = 16.3 (1.2) N = 97 Canadian university students Age M = 21.69 Affiliation measures Diary and questionnaire (54 items) assessing social activities (e.g., getting to know people better) on Facebook SNS measures 7-day diary measure assessing Facebook use frequency and duration Self-report measure assessing Facebook activities (54 items) Affiliation measures Belonging measure (10-item self-report questionnaire) SNS measures Intensity of SNS use (6-item self-report questionnaire) Affiliation measures Lists of top 10 friends through SNS, IM, and faceto-face interactions SNS measures Experimenter-developed self-report survey assessing use of and attitudes toward SNSs Affiliation measures NEO-PI-R to measure personality (including extroversion and openness to experience) and group affiliations on Facebook SNS measures The Facebook Questionnaire (basic use, attitudes toward Facebook, and posting of identifying information; 28-item self-report questionnaire) CMC competence measure (motivation, knowledge, and efficacy; 13-item self-report questionnaire) Affiliation measures Perceptions of SNS use on relationships List of top 10 offline friends List of top 10 online friends SNS measures Typical Internet activities, motivation for Internet use, and SNS activities (self-report questions developed by research team) Ross et al. (2009) Subrahmanyam et al. (2008) N = 131 Undergraduate students in Los Angeles Valkenburg et al. (2006) N = 881 Age 10–19 Dutch users of SNS Affiliation measures Social self-esteem (12-item self-report questionnaire). SNS measures Use of SNSs (3-item self-report questionnaire) Frequency and tone of reactions to profiles (4item self-report questionnaire) relationship quality with friends. Eight studies are based on undergraduate students (Ellison et al. 2007, 2011; Forest and Wood 2012; Grieve et al. 2013; McMillan and Intensity of SNS use was positively associated with feelings of belonging for boys, b = .37, p \ .001, but not girls. Older boys who do not use SNSs (vs. SNS users) report lower perceptions of belonging to their group of friends, b = -.30, p = .004. 43 % of adolescents felt that SNS use made their friendships closer. 17 % of adolescents listed SNS friends that had no overlap with their face-to-face and IM friends. Girls used IM and SNSs more than boys, X2(1, N = 67) = 9.3, p = .002. Extroversion related to greater numbers of group affiliations on Facebook, t(42) = 2.44, p = .019. Individuals with high motivation to use computermediated communication spent more time on Facebook, t(36) = 4.45, p \ .001, and checked their Facebook wall more frequently, t(36) = 3.77, p = .001. 20 % of participants reported that SNSs bring them closer to their friends. A small number of participants indicated that SNSs cause them problems (2.5 %). 73 % of participants reported that SNSs did not impact their relationships. Youth utilize SNSs to keep in touch with friends they do not see often (reported by 81 % of youth) 49 % of students listed the same names for their closest online friends and offline friends. Adolescents’ social self-esteem related to the tone of profile feedback they received, with positive feedback relating to enhanced self-esteem, and negative feedback relating to deflated selfesteem, b = .48, p \ .01 Adolescents who consistently received negative feedback from their profile reported lower close friendship self-esteem, r(881) = .40, p \ .001. Morrison 2006; Pempek et al. 2009; Ross et al. 2009; Subrahmanyam et al. 2008), two are based on adolescents (Lenhart and Madden 2007; Reich et al. 2012), two also 123 8 include younger children (Quinn and Oldmeadow 2013; Valkenburg et al. 2006), and one includes a broader range of ages from adolescents to 30 year olds (Kwon and Wen 2010). Measurements of SNS use assess frequency, intensity, and duration of SNS use, in addition to more detailed measures of specific SNS content shared. SNSs and Friendship Quality Although SNSs have provided notable structural changes to adolescents’ social relations, adolescents and young adults’ stated motivations for using SNSs are quite similar to more traditional forms of communication—to stay in touch with friends, make plans, and get to know people better (Lenhart and Madden 2007; Pempek et al. 2009). That said, evidence suggests that connectedness through SNSs may be slightly different from general social connectedness. A factor analysis indicated ‘‘Facebook connectedness’’ as a distinct construct from general connectedness that was uniquely related to general well-being and negatively related to depression and anxiety (Grieve et al. 2013). Nonetheless, the most common use of SNSs is to maintain and extend existing offline friendships (McMillan and Morrison 2006; Reich et al. 2012; Subrahmanyam et al. 2008). On average, in a college sample, 49 % of respondents’ top face-to-face friends were also their SNS friends (Subrahmanyam et al. 2008). In addition, the content of most SNS communication focuses on everyday events related to school, mutual friends, and upcoming activities. Only 29 % reported using SNSs to ‘‘look for new people.’’ The online–offline friend overlap findings were replicated in a study of high school students, with only 17 % of adolescents listing SNS friends that had no overlap with their face-to-face and IM friends (Reich et al. 2012). Thus, although young people might list hundreds of ‘‘friends’’ on SNS sites, the majority of their SNS time involves extensions of their offline relationships. Cross-sectional studies examining the relationship between frequency of SNS use and friendship quality, specifically, show that SNS use is associated with enhanced relationship quality and intimacy (Ellison et al. 2007; Grieve et al. 2013; McMillan and Morrison 2006; Reich et al. 2012), suggesting some support for the concept that SNSs enhance social connections, i.e., the stimulation hypothesis. To investigate how SNS connectedness is linked to relationship quality, McMillan and Morrison (2006) coded emerging adults’ narratives about computermediated communication. Findings demonstrated that participants viewed this type of communication as something that facilitates planning social activities, maintaining ties with friends, and feeling part of a community. Interestingly, Korean SNS users with a high (versus low) sense of social identity, in particular, found SNSs to be a useful 123 Clin Child Fam Psychol Rev (2014) 17:1–18 and good resource for social support (Kwon and Wen 2010). In a study of younger participants, Reich and colleagues (2012) reported that 43 % of their high school student participants believed that SNS use made them feel closer to their friends. Similarly, 20 % of college students indicated that SNSs brought them closer to their friends, whereas only 2.5 % indicated that it had a negative impact; however, the majority (73 %) indicated SNS use did not have an effect on their relationships (Subrahmanyam et al. 2008). Despite overall support for SNSs enhancing friendship quality, interactive effects suggest that some individuals may benefit more than others. For example, in support of the ‘‘rich-get-richer’’ hypothesis, Canadian undergraduates scoring high on extroversion who likely have more offline friends reported more affiliations on Facebook than undergraduates scoring low on extroversion (Ross et al. 2009). Examining closeness to friends more thoroughly, Ellison et al. (2007) reported that those who use Facebook intensely (i.e., actively engage in Facebook activities frequently and feel emotionally connected to the use of Facebook) perceive greater bonding social capital, or strong social ties and closeness to others who can provide emotional support. Interestingly, in support of the social compensation hypothesis, an interaction effect revealed that those with low self-esteem and low life satisfaction particularly benefitted from Facebook use in terms of more emotional support. However, a later study indicated a ceiling related to social capital benefits; after individuals report having more than 500 ‘‘actual’’ (vs. online only) Facebook friends, there are diminishing returns on social capital gains. Focusing on early adolescents’ sense of belongingness, a highly salient social construct for adolescents, Quinn and Oldmeadow (2013) demonstrated that SNS use is related to a greater sense of belongingness for boys, but not girls in a sample of young students north of England. Older boys who did not use SNSs reported a lower sense of belongingness than SNS users, which may be indicative that this sample of boys also had poorer social skills or other personality differences impacting their social connections both online and offline. Research incorporating objective SNS content to examine the role of self-esteem in SNS use and social connectedness supports the ‘‘rich-get-richer’’ hypothesis. Facebook users with low, compared to high, self-esteem posted status updates that were rated as lower in positivity and higher in negativity by trained undergraduate coders (Forest and Wood 2012). Based on the recently posted status updates and the number of ‘‘likes’’ and comments received by those posts, Facebook users with low selfesteem also were not as ‘‘likable’’ to the trained coders. In a related vein, Dutch adolescents who frequently received negative feedback from their SNS profiles also reported Clin Child Fam Psychol Rev (2014) 17:1–18 lower social self-esteem (Valkenburg et al. 2006). In general, the direction of effects is unclear and may indeed be reciprocal, with low self-esteem individuals posting more negative messages and receiving less positive feedback, which then fuels the low self-esteem. It is worth noting, however, that even the Facebook users with low selfesteem reported that Facebook was a safe way for them to self-disclose. Thus, although individuals with low selfesteem may view SNSs as a useful way to feel connected to others, supporting the social compensation hypothesis (McKenna et al. 2002), those with poor social skills may be at risk for opening themselves up to harmful feedback from others. Summary Generally, SNS use appears to benefit and not detract from adolescents’ sense of peer affiliation, but adolescents’ offline level of social functioning is a consideration in the overall impact. Those adolescents who have strong offline social skills also appear to have more online connections and contacts. Whether online communication actually improves the overall quality of their relationships or simply resembles their already strong relationships is difficult to tease apart without longitudinal studies. On the other hand, some adolescents who have more limited social success offline appear to derive enhanced relationship satisfaction online, particularly if they find online communications more comfortable than offline social interaction. The Internet may provide a leveling effect in relationship satisfaction for certain individuals, as described in the social compensation hypothesis (McKenna et al. 2002). There are, however, some caveats to the ameliorating influences of online communication, that is, adolescents who post more negative messages, which may include those with low self-esteem or poor social skills, open themselves up to negative feedback from others. Thus, there is evidence for an overall positive association between SNS use and adolescents’ sense of social connectivity. While those who are less socially inclined may report feeling more socially connected through SNSs, as described in the social compensation hypothesis, those who are less socially inclined may also be likely to receive less positive input from others via SNSs. It remains unclear whether this translates into fewer social benefits from SNSs, or whether SNSs actually are detrimental to less socially skilled adolescents. Review of Findings: SNS Use and Identity Table 2 displays 14 studies that examine the link between SNS use and constructs related to identity. Five of these studies examined adolescent samples (Hillier and Harrison 2007; Tynes et al. 2008; Valkenburg and Peter 2007; 9 Valkenburg et al. 2011; Yu et al. 2011), four included adolescents as well as older participants (Back et al. 2010; Haferkamp and Kramer 2011; McLaughlin et al. 2012; Silenzio et al. 2009), and five are based on college samples (Chou and Edge 2012; Christofides et al. 2009; Grasmuck et al. 2009; Walther 2007; Walther et al. 2011). SNSs, Self-Disclosure, and Self-Presentation Self-disclosure, which involves an iterative process of sharing personally relevant information and receiving feedback, is central to identity formation. SNSs bring both sides of this information-sharing into a highly public arena. Based on Canadian participants’ self-reports, Facebook disclosures are ‘‘likely’’ or ‘‘very likely’’ to include information about salient recent or upcoming happenings—by sharing pictures with friends, information about relationship status, and mention of their birthday (Christofides, Muise, and Desmarais 2009). Some adolescents report that online interactions are more conducive to self-disclosure than face-to-face interactions (Valkenburg and Peter 2007). Valkenburg, Sumter and Peter (2011) indicated that online self-disclosure may be a ‘‘rehearsal’’ for other types of selfdisclosure, that is, online self-disclosure to known friends in early adolescent years was associated with greater offline self-disclosure at the next wave of data collection, one-half year later; yet, offline self-disclosure did not lead to greater online self-disclosure. With adolescents controlling what information and photographs they wish to share to a broad audience through SNS profiles, there is considerable speculation that some adolescents may post misinformation or at least idealized versions of themselves. To examine this possibility, researchers asked individuals to report on themselves as they are and as they ideally would like to be. Additionally, several close friends also completed personality measures about the participant, and objective research assistants coded individuals’ actual SNS profiles. Results demonstrated that adolescents did not portray their ‘‘ideal selves’’ through their SNS profiles, and that certain personality characteristics, such as extroversion and openness, came across accurately through SNS profiles (Back et al. 2010). There is also evidence, however, that while individuals may not express idealized versions of themselves via SNSs, they may alter or highlight different aspects of themselves. Walther (2007) created several conditions in which undergraduate students were told that an online message would be received by different individuals who varied by age, status, and relevance to their own life (e.g., professor from their university, high school student from another state, or college student from another university). Time spent on the message, as well as number of edits and level of message complexity, were all objectively measured 123 10 Clin Child Fam Psychol Rev (2014) 17:1–18 Table 2 SNS use and identity Authors (year) Back et al. (2010) Sample characteristics Measures Primary results N = 236 Identity measures SNS users from the United States and Germany Ten-item Personality Inventory Observers accurately rated participants’ personalities based on viewing their SNS profile, particularly for extroversion, r = .39, p \ .001, a...
Purchase answer to see full attachment