Gender Roles and Binge Drinking Among Latino Emerging ... - PsycNET

Psychology of Addictive Behaviors 2014, Vol. 28, No. 3, 719 –726

© 2014 American Psychological Association 0893-164X/14/$12.00 http://dx.doi.org/10.1037/a0037406

Gender Roles and Binge Drinking Among Latino Emerging Adults: A Latent Class Regression Analysis Ellen L. Vaughan, Y. Joel Wong, and Katharine G. Middendorf

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Indiana University Bloomington Gender roles are often cited as a culturally specific predictor of drinking among Latino populations. This study used latent class regression to test the relationships between gender roles and binge drinking in a sample of Latino emerging adults. Participants were Latino emerging adults who participated in Wave III of the National Longitudinal Study of Adolescent Health (N ⫽ 2,442). A subsample of these participants (n ⫽ 660) completed the Bem Sex Role Inventory—Short. We conducted latent class regression using 3 dimensions of gender roles (femininity, social masculinity, and personal masculinity) to predict binge drinking. Results indicated a 3-class solution. In Class 1, the protective personal masculinity class, personal masculinity (e.g., being a leader, defending one’s own beliefs) was associated with a reduction in the odds of binge drinking. In Class 2, the nonsignificant class, gender roles were not related to binge drinking. In Class 3, the mixed masculinity class, personal masculinity was associated with a reduction in the odds of binge drinking, whereas social masculinity (e.g., forceful, dominant) was associated with an increase in the odds of binge drinking. Post hoc analyses found that females, those born outside the United States, and those with greater English language usage were at greater odds of being in Class 1 (vs. Class 2). Males, those born outside the United States, and those with greater Spanish language usage were at greater odds of being in Class 3 (vs. Class 2). Directions for future research and implications for practice with Latino emerging adults are discussed. Keywords: binge drinking, gender roles, Latinos, latent class regression

ethnic/racial differences in heavy drinking among emerging adults, with Latinos drinking less than Caucasians but more than African Americans and Asian Americans (Johnston, O’Malley, Bachman, & Schulenberg, 2007; Paschall, Bersamin, & Flewelling, 2005). To understand these differences, research must attend to the cultural characteristics that may serve as important risk and protective factors for alcohol use among Latino emerging adults. Despite emerging adulthood as a developmental period of increased risk for alcohol use, there is a paucity of research that investigates binge drinking among Latino emerging adults. Latino emerging adults face the same potential risks and consequences of excessive alcohol use as other emerging adults including driving while intoxicated, aggression, risky sex, and the development of alcohol abuse and dependence (National Institute on Alcohol Abuse and Alcoholism, 2007; Vicary & Karshin, 2002; Wells, Speechley, Koval, & Graham, 2007). However, the health disparities literature has noted that despite lower drinking rates than their White counterparts, Latinos experience greater consequences from their use (Chartier & Caetano, 2010). Thus, understanding risk and protective factors unique to the growing population of Latino emerging adults deserves critical attention and has important implications for the development of prevention and intervention efforts. Emerging adults are in a stage of exploration that involves the development of identity in multiple life domains (e.g., relationships). For Latino youth, a concomitant process of cultural adaptation may be integral to identity exploration and may be related to alcohol use outcomes. Theoretical advancements in the understanding of acculturation include adaptation in the contexts of both receiving and heritage cultures in three domains (Schwartz, Unger,

Binge drinking is a prevalent emerging adult risk behavior (Chen, Dufour, & Yi, 2004 –2005). Peak engagement in binge drinking often coincides with the developmental tasks of emerging adulthood, including the development of long-term academic and career goals and the formation of important interpersonal relationships (Arnett, 2006; Maggs & Schulenberg, 2004 –2005). Much of the literature on drinking during emerging adulthood focuses on college students, although emerging adults may also be transitioning into other contexts such as the workforce, military, marriage, or parenthood. These transitions are not without risk as emerging adults often engage in risk behaviors. There are well-documented

Ellen L. Vaughan, Y. Joel Wong, and Katharine G. Middendorf, Department of Counseling and Educational Psychology, Indiana University Bloomington. This research used data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by Grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 other federal agencies and foundations. Special acknowledgment is due Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health Website (http://www.cpc .unc.edu/addhealth). No direct support was received from Grant P01HD31921 for this analysis. Correspondence concerning this article should be addressed to Ellen L. Vaughan, Indiana University, Department of Counseling and Educational Psychology, 201 N Rose Ave, Bloomington, IN 47405-1006. E-mail: [email protected] 719


720

VAUGHAN, WONG, AND MIDDENDORF

Zamboanga, & Szapocznik, 2010). These domains involve behavioral acculturation, values acculturation, and identity-based acculturation. Adherence to strict gender roles might reflect both values and identity-based acculturation, whereas language usage represents behavioral acculturation. This more nuanced conceptualization of acculturation allows for not only understanding how acculturation, broadly, is related to binge drinking, but also how different aspects of acculturation may contribute to drinking. Acculturation has been studied as a predictor of alcohol use in the Latino population (Zemore, 2007). Studies have shown that higher levels of acculturation are a risk factor for alcohol abuse and dependence (Caetano, Ramisetty-Mikler, & Rodriguez, 2009; Safer & Piane, 2007). Using the National Epidemiologic Study of Alcohol and Related Conditions, Strunin, Edwards, Godette, and Heeren (2007) found that being born in Mexico was protective against alcohol abuse, but not alcohol dependence. Another study showed that U.S.-born Puerto Ricans and South/Central Americans have higher rates of alcohol abuse and dependence than those that are foreign-born (Caetano et al., 2009). Furthermore, among emerging adults, acculturation may impact behavior risk differently for men and women. Studies have linked greater acculturation to more alcohol use among Latina college students than their Latino counterparts (Rafaelli et al., 2007; Safer & Piane, 2007). However, another study of Mexican American college students found no relation between acculturation, and heavy alcohol use was found for both men and women (Zamboanga, Rafaelli, & Horton, 2006). Taken together, these studies suggest that there is variation in the role of acculturation on drinking behavior and that it might impact the drinking behavior of women differently from men. Although there are consistent gender differences in alcohol use across all racial/ethnic groups, when compared with their Caucasian counterparts, the gap is particularly large for Latino emerging adults (Corbin, Vaughan, & Fromme, 2008; O’Malley & Johnston, 2002). Gender roles may be important in understanding the differences in binge drinking for Latinos and Latinas. Several review articles have suggested that this gender difference is likely driven by traditional cultural influences that discourage drinking among women (Galanti, 2003; Gilbert & Collins, 1997). These authors hypothesize that cultural prohibitions against drinking for women are protective. Cultural adaptation may shift these prohibitions and, as Latinas spend more time outside their culture of origin, they may be more likely to binge drink. However, there have been few studies on gender roles and drinking among emerging adults. Gender roles among Latinos are often referred to as marianismo for women, and machismo for men. Recent development of a scale measuring marianismo yielded multiple dimensions of this construct (Castillo, Perez, Castillo, & Ghosheh, 2010). Domains of marianismo include behavioral expectations in five domains: being the family pillar, being virtuous and chaste, being subordinate to others, silencing to maintain harmony, and being a spiritual pillar. Recent developments in the conceptualization of machismo broaden this construct to include traditional machismo and caballerismo (Arciniega, Anderson, Tovar-Blank, & Tracey, 2008). These two domains reflect different aspects of male identity with traditional machismo reflecting dominance, competitiveness, and aggressiveness. Caballerismo reflects a man’s obligation to take care of his family, respect elders, and be fair. This, along with other studies, suggests that machismo is much more than the

stereotypic hypermasculine notion of being dominant, independent, and authoritarian (Falicov, 2010; Torres, Solberg, & Carlstrom, 2002). The research investigating gender roles and alcohol use among Latino emerging adults has been sparse. Venegas, Cooper, Naylor, Handson, and Blow (2012) found that adherence to traditional gender roles did not significantly predict heavy episodic drinking among Latino college students attending a Hispanic serving institution. These authors note that their sample at a Hispanic serving institution may have had little variation in its cultural constructs. Arciniega and colleagues (2008) did not find that their measures of traditional masculinity and caballerismo were significantly related to alcohol use. It is possible that their adult sample ranging in age from 18 to 74 years did not capture the exploration in gender roles one might see in an emerging adult sample that is exploring different aspects of identity. A broader study of male gender roles and alcohol use and its consequences in a non-Latino sample suggests that traditional gender attitudes for men are associated with greater consumption and that such attitudes are associated with more alcohol use problem for women (McCreary, Newcomb, & Sadava, 1999). Furthermore, emerging adulthood is also associated with peak alcohol use for many. Thus, investigation of gender roles and alcohol use in an emerging adult sample may provide important information for the development of intervention and prevention programs for this group. In the current study, we conducted secondary data analysis using data from the National Longitudinal Survey of Adolescent Health (Add Health). Wave III of this study administered the Bem Sex Role Inventory—Short (BSRI–S). Studies on the impact of gender roles are difficult to interpret because of the differing measures used to operationalize gender roles. The current study used Choi, Fuqua, and Newman’s (2009) gender role categories. Through factor analysis, these authors analyzed the BSRI–S and found three gender role categories: femininity, social masculinity, and personal masculinity. Feminine characteristics included being warm, compassionate, gentle, and tender. Social masculinity, which included characteristics such as being forceful, dominant, and aggressive, represented a dimension of social control over others. In contrast, personal masculinity, which included characteristics such as being willing to take a stand and defends own beliefs, was a dimension representing a personal or internal locus of self-control. In Choi and colleagues’ sample of emerging adults, women were higher in femininity and men were higher is social masculinity, but women and men had equal levels of personal masculinity. However, this sample was disproportionately European American, with only 2% of the sample identifying as Latino. The current study investigated the role of gender roles in binge drinking in a nationally representative sample of Latino emerging adults who participated in Add Health. We employed latent class regression to test whether there are different classes of participants for whom the relationship between gender roles and binge drinking differs. There is empirical and theoretical literature linking acculturation and indicators of cultural context (e.g., language usage, nativity) to behavioral outcomes, including alcohol use (e.g., Schwartz et al., 2010; Zemore, 2007). Although these constructs have been less studied among Latino emerging adults, research in this area is growing (Rafaelli et al., 2007; Schwartz et al., 2011; Vaughan, Kratz, Escobar, & Middendorf, 2013; Venegas et al., 2012). Although the Add Health data do not include comprehen-

GENDER ROLES AND BINGE DRINKING

sive measures of acculturation, they include items related to acculturation. Thus, we also tested whether these classes varied significantly as a function of gender, language usage, and nativity (immigrant vs. U.S.-born).

Method


Participants and Procedure Data from Wave III from the Add Health study were used for this secondary data analysis (Harris et al., 2009). Participants in this study included individuals who were Latino emerging adults between the ages of 18 and 25 years who had responded to measures assessing nativity, language usage, and gender roles. In Wave III of Add Health, 2,442 participants were Latino and between the ages of 18 and 25 years. These participants were on average 21.87 (SE ⫽ 0.23) years old, 48.1% were female, and 23.4% were born outside the United States. Finally, 29% of the sample reported at least one episode of binge drinking in the past 2 weeks. The BSRI–S was administered only to a subsample of Wave III participants. For the current study, 660 participants were administered the BSRI–S. Wave III participant responses were gathered via in-home interviews.

Measures Demographic variables. Demographic variables included gender and nativity for this Latino sample. Nativity was assessed by evaluating yes or no responses to the question “Were you born in the United States?” Gender was dummy coded (male ⫽ 1, female ⫽ 0). Language usage variables. Language usage was operationalized by assessing language usage with family and friends. For this purpose, a composite of two questions investigating language usage in these two contexts was evaluated. The two questions were “What language do you use most with your family and close relatives?” and “What language do you use most with your close friends?” For this study, possible responses included English, Spanish, and half-English and half another language. Response options were coded Spanish ⫽ 0, half English– half another language ⫽ 1, and English ⫽ 2. These two items were summed, yielding a response range of 0 to 4 (M ⫽ 2.93, SE ⫽ 0.11). Two items composed the media language variable. Participants were asked, “Do you often buy recorded music that is sung in a language other than English?” and “Do you often read newspapers, watch TV shows, or listen to radio programs that are in a language other than English?” Response options for each item are coded 1 ⫽ yes and 0 ⫽ no. These items were summed and the dichotomized to reflect any media usage in another language. Gender roles. The BSRI–S (Bem, 1974) assesses gender role orientation using self-report. Traditionally, gender role outcomes have been characterized as masculine, feminine, or androgynous. More recently, Choi et al. (2009) analyzed the BSRI–S via factor analysis and found three gender role factors, two masculine and one feminine. This study used these three factors: personal masculinity, social masculinity, and femininity to analyze gender role orientation. The personal masculinity factor assesses the personal domain of masculinity and consists of five items including is willing to take a stand, defends own beliefs, is independent, has

721

leadership abilities, and has a strong personality. Internal consistency for the personal masculinity scale was .80 in the current sample. The social masculinity factor assesses the social aspects of masculinity and contains the following four items: forceful, dominant, aggressive, and assertive. Internal consistency for the social masculinity scale was .66 in the current sample. The femininity factor comprises nine items: affectionate, warm, compassionate, gentle, tender, sympathetic, sensitive to the needs of others, soothe hurt feelings, and understanding. For the current sample, internal consistency for the femininity scale was .92. Alcohol use variables. Binge drinking was dichotomized from two questions. Men were asked, “During the past two weeks, how many times did you have five or more drinks on a single occasion, for example, in the same evening?” Women were asked, “During the past two weeks, how many times did you have four or more drinks on a single occasion, for example, in the same evening?” Respondents’ answers ranged from 0 to 14 for both questions. Binge drinkers were defined as those who reported one or more episodes of binge drinking in the past 2 weeks. Likewise, any alcohol use problem was a dichotomized outcome generated from three items reflecting interpersonal problems related to alcohol use: “You had problems at school or work because you had been drinking,” “You had problems with your friends because you had been drinking,” and “You had problems with someone you were dating because you had been drinking.” Experiencing any alcohol use problem were those respondents who indicated that any of these items had occurred one or more times in the past year. Typical number of drinks was included in our follow-up analyses of alcohol use problems. This was measured by the item “Think of all the times you have had a drink during the past 12 months. How many drinks did you usually have each time? A ‘drink’ is a glass of wine, a can of beer, a wine cooler, a shot glass of liquor, or a mixed drink.”

Statistical Analyses We used latent class regression (LCR) to address our research questions. LCR empirically tests whether there are one or more classes of participants with different regression coefficients for the association between a series of predictor variables and an outcome (Vermunt & Magidson, 2008). A key methodological advantage of LCR is that unlike conventional regression methods, there is no assumption that individuals are from a single population; consequently, LCR can test for parameter differences across unobserved subgroups (Ding, 2006). LCR has been used in several Add Health studies as well as drug and alcohol studies. In a recent Add Health study, Wong and Maffini (2011) examined the optimal number of latent classes that reflected differential associations between relational bonds and suicide attempts among Asian American adolescents. In another study, Reboussin, Hubbard, and Ialongo (2007) used LCR to identify subgroups (latent classes) of African American adolescents who demonstrated similar patterns in predictors of marijuana involvement. In our current study, the predictors were femininity, social masculinity, and personal masculinity; the dichotomous outcome was binge drinking; and the covariates were language usage, usage of media in another language, gender, and nativity (U.S.-born vs. immigrant). To conduct our LCR analysis, we used Latent Gold 4.5 software (Vermunt & Magidson, 2008), which can handle



722

stratification, cluster sampling, and probability sampling weights in the Add Health data. The model testing strategy in LCR involves examining the most parsimonious one-class model and then testing an increasing number of latent classes (two to five class models) to identify the model that best represents an adequate fit to the data. To select the most appropriate model, we considered several criteria (Nylund, Asparouhov, & Muthén, 2007; Vermunt & Magidson, 2008). First, we examined the fit of the models tested using the Akaike information criterion (AIC), Bayesian information criterion (BIC), and conditional bootstrap estimates of the loglikelihood (⫺2LL). Lower AIC and BIC estimates indicate better model fit, whereas conditional bootstrapping tests whether the improvement in model fit as a result of an additional latent class is statistically significant. Second, we examined the accuracy in which cases were classified into classes through the use of classification errors (proportion of cases estimated to be misclassified in the latent classes) and the entropy R2 (accuracy of class membership based on the manifest variables). A low proportion of classification errors and a high entropy R2 suggest a high degree of classification accuracy. Third, we focused on the interpretability of our findings based on our identified models. Interpretable models tend to have latent classes that are meaningfully distinct from one another in terms of significant associations between predictors and outcomes and do not have classes with unusually small memberships (e.g., less than 1% of the total sample). Fourth, we relied on the principle of parsimony; that is, our goal was to identify the most parsimonious model (i.e., the model with the fewest number of latent classes) that was an adequate fit to the data.

Results Preliminary Results Given that we assessed gender role orientation using the three factors of the BSRI–S identified by Choi et al. (2009), we examined whether the factor structure identified by these authors applied to our sample of Latina/o emerging adults. Therefore, we conducted a confirmatory factor analysis on the 18-item BSRI based on an oblique three-factor model. We used Mplus Version 6.12 (Muthén & Muthén, 2011), which can accommodate the probability sampling weights in the Add Health data. Hu and Bentler (1999) recommend the following cutoff scores for fit indices to evaluate the adequacy of model fit: the comparative fit index (CFI; a value ⱖ.95), the root mean square error of approximation (RMSEA; a value ⱕ.06), and the standardized root mean square residual (SRMR; a value ⱕ.08). However, other researchers have argued that these guidelines on fit indices may be overly restrictive (Marsh, Hau, & Wen, 2004). For instance, a CFI value of .80 –.89 may be considered an adequate and marginal fit (Meyers, Gamst, & Guarino, 2013). Our results indicated an adequate model fit to the data: ␹2/df ⫽ 1.78; CFI ⫽ .89; RMSEA ⫽ .038 (90% CI [.030, .045]); SRMR ⫽ .072. Moreover, all the item loadings on the latent variables were significant, p ⬍ .001. These results are fairly consistent with the model fit statistics found by Choi and colleagues (2009). Their model fit statistics on the aforementioned fit indices were ␹2/df ⫽ 2.94, CFI ⫽ .93, RMSEA ⫽ .07. They did not report SRMR. In sum, these findings provide support for the applicability of the Choi et al. three-factor model of the BSI–S to our sample.

LCR Results Our LCR model testing results comparing one to five latent class models are shown in Table 2. Among the five classes, the three-class model had the lowest AIC value and the second lowest BIC value. The conditional bootstrap estimates of the loglikelihood also suggested that the three-class model provided the most optimal representation of the data. The two-class model was a significantly better fit than the one-class model, and the threeclass model was a significantly better fit than the two-class model. However, the four-class model was not a significantly better fit than the three-class model, and the five-class model was not a significantly better fit than the four-class model. With regard to classification accuracy, the three-class model had a relatively low proportion of classification errors (0.09) and the highest entropy R2 value among the five classes (.78). In terms of interpretability, the proportion of participants in each latent class was as follows: Class 1 ⫽ 55%, Class 2 ⫽ 27%, Class 3 ⫽ 18%. Hence, we did not find any disproportionately small classes (e.g., 1% of the sample). In addition, as demonstrated in our subsequent analyses, the latent classes were meaningfully distinct from one another in terms of the associations between the predictor variables and outcomes. With regard to parsimony, the three-class model accounted for the associations between the predictors and outcome using fewer classes than the four- and five-class models. Collectively, the above-mentioned criteria of model fit, classification accuracy, interpretability, and parsimony pointed to the three-class model as the most optimal representation of the data. The parameter estimates for our three-class model are presented in Table 3. Overall, the three gender role variables—femininity, social masculinity, and personal masculinity— differed significantly across the three latent classes in their associations with binge drinking. In Class 1, personal masculinity was significantly associated with lower odds of binge drinking, but femininity and social masculinity were not significant predictors. Therefore, we labeled Class 1 the protective personal masculinity class. Because none of the gender role variables was significantly related to binge

Table 1 Percentages of Binge Drinking and Alcohol Use Consequences by Demographic Characteristics

Variable

Binge drinkers (n ⫽ 654)a

Any alcohol use problem (n ⫽ 451)a

34.1 16.8

21.9 11.8

26.6 20.5

19.0 11.1

23.5 24.9 30.3

6.9 22.7 20.6

17.4 25.5 29.2 24.9

4.3 19.1 8.6 17.2

Gender Male Female Nativity Born in the United States Born outside the United States Language usage with family Spanish English Half English–half another language Language usage with peers Spanish English Half English–half another language Total sample a

ns are unweighted; percentages are weighted.


Table 2 Latent Class Analyses: Model Comparison Model 1 2 3 4 5

class class class class class

BIC

AIC

⫺2LL difference

Bootstrap p

Classification errors

647.98 659.01 649.39 695.17 732.90

630.76 603.06 554.69 561.73 560.73

45.71 66.36 10.96 19.01

.004 ⬍.001 .314 .118

0.06 0.09 0.14 0.17


Note. BIC ⫽ Bayesian information criterion; AIC ⫽ Akaike information criterion.

drinking in Class 2, we labeled this class the nonsignificant class. In Class 3, femininity was not significantly related to binge drinking. However, social masculinity was significantly associated with higher odds of binge drinking, whereas personal masculinity was significantly related to lower odds of binge drinking. Given the contrasting findings for social and personal masculinities, we labeled Class 3 the mixed masculinity class. Multinomial logistic regression was conducted to determine whether gender, nativity, and language usage significantly predicted class membership. Unadjusted models were run for each independent variable (gender, nativity, language usage with family and friends, and usage of media in another language) using class as the dependent variable. Class 2 was specified as the reference group because it was characterized by nonsignificant relationships between gender roles and binge drinking. Unadjusted models were significant for gender, language usage, and nativity. Thus, these variables were retained in the adjusted model. The adjusted model was significant, Wald F(6, 67) ⫽ 17.58, p ⬍ .001, and correctly classified 77% of cases. Women had greater odds of being in Class 1 (OR ⫽ 60.85, 95% CI [15.59, ⬎100]) and lower odds of being in Class 3 (OR ⫽ 0.04, 95% CI [0.02, 0.11]) versus Class 2. Those born outside the United States had greater odds of being in Class 1 (OR ⬎ 100, 95% CI [44.37, ⬎100]) and Class 3 (OR ⫽ 8.75, 95% CI [2.97, 25.73]) versus Class 2. Finally, for every unit increase in language usage participants had greater odds of being in Class 1 (OR ⫽ 21.99, 95% CI [9.15, 25.87]) versus Class 2 and lower odds of being in Class 3 (OR ⫽ 0.67, 95% CI [0.46, 0.97]) versus Class 2. Thus, the protective personal masculinity class was characterized by greater odds of including those who were female, born outside the United States, and reporting greater English language usage. The mixed masculinity class was characterized by

723

greater odds of those who were men, being born outside the United States, and less English language usage. In a follow-up analysis, we used LCR to test whether there were different classes of participants for whom the relationship between gender roles and drinking-related problems differed among participants with a lifetime history of alcohol use. The variables were identical to those in the previous LCR analysis, except that the outcome variable was drinking-related problems, and typical number of drinks consumed per drinking occasion was added as a predictor. Our LCR model testing one to five latent classes revealed that the one-class model and the four-class model had the lowest BIC and AIC values, respectively. However, the four-class model had a relatively high proportion of classification errors (0.22). Also, the conditional bootstrap estimate of the loglikelihood indicated that the two-class model was not a significantly better fit to the data than the one-class model (p ⫽ .108). Given the importance of parsimony, we concluded that the oneclass model provided the most optimal representation of the data. None of the predictors (number of drinks, femininity, social masculinity, and personal masculinity) were significantly related to drinking-related problems in the one-class model.

Discussion Latino emerging adults report rates of heavy drinking that are higher than their Asian and Black counterparts, but lower than Whites and those who identify as other (Paschall et al., 2005). Furthermore, there have been persistent gender differences in alcohol use among Latino emerging adults (Corbin, Vaughan, & Fromme, 2008; Paschall et al., 2005). Efforts to explain these differences have often pointed to traditional gender roles as protection against drinking for Latinas (Galanti, 2003; Gilbert & Collins, 1997). We used LCR, a descriptive method that allows for identifying groups in which there may be different patterns in the relationships between gender roles and binge drinking. Results indicated there were three latent classes of individuals in our sample. In two classes, masculinity was associated with binge drinking and, for a third class, there was no relationships between gender roles and binge drinking. Taken together, this suggests that the overall relationship between gender roles and binge drinking might be more nuanced than simple adherence or nonadherence to traditional gender roles. Furthermore, only masculinity played a role in binge drinking for both men and women. Our results identified a protective personal masculinity class wherein emerging adults who endorsed a strong sense of self or

Table 3 Latent Class Regression Model Predicting Binge Drinking

Variable

Class 1: Protective personal masculinity OR [95% CI]

Class 2: Nonsignificant OR [95% CI]

Class 3: Mixed masculinity OR [95% CI]

p-value difference

Femininity Social masculinity Personal masculinity

1.12 [0.92, 1.37] 1.07 [0.92, 1.25] 0.69ⴱ [0.52, 0.91]

0.12 [0.01, 1.84] 0.001 [⬍0.001, 33.71] 4817.45 [0.03, ⬎100]

0.44 [0.13, 1.56] ⬎100ⴱ [4.15, ⬎100] 0.001ⴱ [⬍0.001, 0.28]

.015 .004 .001

Note. OR ⬍1.00 indicates that a variable predicted lower odds of binge drinking; OR ⬎1.00 indicates that a variable predicted higher odds of binge drinking. The p-value difference reflects differences between the classes in the prediction of binge drinking. ⴱ Variable was a significant predictor of binge drinking for the given class (p ⬍ .05).


724


being firm in their beliefs had lower odds for binge drinking. These emerging adults may be better able to choose nonalcoholic beverages or employ protective strategies that limit the amount of alcohol they consume. Follow-up analyses found that this class comprised those participants who were female, born outside the United States, and noted greater English language usage. This finding is an important addition to the cultural literature with respect to acculturation. Zemore (2007), using meta-analysis, found that greater acculturation is associated with more drinking. The emerging adult literature has found this to be particularly salient for women (Rafaelli et al., 2007; Safer & Piane, 2007). In this case, it appears that there is a group of emerging adult women, who are born outside the United States, but who have greater English language usage, for whom personal masculinity protects them from binge drinking. This group of women may be protected by a shift in gender roles that allows them to assert independence in decisions around drinking. It is unclear whether the erosion of traditional gender roles impacts their drinking behavior as femininity was unrelated to drinking. In the mixed masculinity class, personal and social masculinity played different roles in predicting binge drinking. Like the protective masculinity class, being independent or standing up for one’s beliefs was protective against binge drinking. However, endorsing characteristics of being aggressive and dominant increased the odds of binge drinking. This class largely comprised men, those born outside the United States, and those with less English language usage. In the context of the Latino male gender roles literature (e.g., Arciniega et al., 2008), this result seems to accurately reflect how machismo and caballerismo might be related to alcohol use. That is, the more competitive aspects of machismo increase risk for binge drinking, whereas aspects of caballerismo such as leadership in meeting one’s obligation to family is protective against binge drinking. In addition, our finding that men had greater odds (than women) of being in the only latent class in which social masculinity was a risk factor for binge drinking converges with Iwamoto and Smiler’s (2013) results that endorsement of masculine norms was more strongly associated with alcohol use for boys than for girls. Interestingly, our results indicate that for both men and women, gender roles play an important role in binge drinking for emerging adults. In addition, this was more likely the case for those born outside the United States. For both men and women born outside the United States, it appears that having a firm sense of one’s beliefs and willingness to act on them is protective. It may be that if one is confident in her/his beliefs, then one is better able to act in a way that is consistent with those beliefs. For women, this may reflect greater adoption of less traditional gender roles. However, for men, being aggressive, forceful, and dominant may contribute to more risky drinking. These characteristics might be associated with more stereotypic portrayals of men’s drinking behavior particularly for those born outside the United States. In the United States, drinking behavior by men, especially young men, is often portrayed as competitive. The link between alcohol use and its consequences for both college and noncollege emerging adults is well established (Cleveland, Mallett, White, Turrisi, & Favero, 2013; Quinn & Fromme, 2011). Thus, we conducted follow-up analyses testing the link between gender roles and alcohol use problems. Results revealed a one-class model wherein our hypothesized independent variables

were not significant predictors of interpersonal alcohol use consequences. It may be that for this population gender roles are not linked to alcohol use problems or consequences. Alternatively, it could be that the links between gender roles and alcohol use consequences are more complex and indirect. Testing of culturally specific dimensions of gender roles and potential mediators represents an important avenue for future research. There are numerous strengths of the current study worth noting. Although the literature has hypothesized that gender roles play a role in Latino and Latina drinking behavior (Galanti, 2003; Gilbert & Collins, 1997), few studies have investigated this assertion among emerging adults. This study provides some evidence that masculine gender roles may contribute to both risk and protection for binge drinking. Surprisingly, feminine gender roles were not related to binge drinking for men or women. One possibility for this is that the measurement of femininity by the BSRI–S reflects descriptors such as being affectionate or warm. It is possible that these descriptors are not well linked to behavioral expectations for Latino emerging adults with respect to gender. Perhaps using more culturally specific measures of marianismo (e.g., Castillo et al., 2010) might yield information on how gender roles might be related to binge drinking. Although there are numerous strengths of the current study, it is not without limitations. Although an adaptation of the BSRI–S has been used to study gender roles and alcohol use among adolescents in Mexico, it does not measure culturally specific gender roles (Kulis, Marsiglia, Lingard, Neiri, & Nagoshi, 2008). There are measures of machismo, caballerismo, and marianismo, which reflect culturally specific gender roles (e.g., Arciniega et al., 2008; Castillo et al., 2010). For example, the measure of marianismo taps the ways in which Latina women must ensure family harmony and unity (Castillo et al., 2010). The Arciniega et al. (2008) measure of machismo and caballerismo reflects Latino men’s roles such as showing strength and commitment to family. These measures reflect explicit behavioral and attitudinal expectations for Latino men and women. Finally, the current study did not capture whether the participants were struggling with their gender roles. Indeed, the extent to which one experiences gender role conflict (O’Neil, 2008) may be more predictive of health risk behaviors such as drinking than one’s gender roles. Second, the Add Health study did not directly measure acculturation. Although language usage and nativity have been used as indicators of acculturation, they do not fully reflect the current theoretical understanding of this complex construct (see Schwartz et al., 2010). Third, the current study was cross-sectional and captured only a snapshot of participants’ current gender roles and drinking behavior. Finally, the current study used a dichotomous measure of binge drinking. Thus, the link between gender roles and binge drinking cannot be extrapolated to other drinking behaviors. The aforementioned limitations hold promise for future investigations including extensions to potential indirect relationships between gender roles and problems associated with alcohol use for Latino emerging adults. Gender role conflict may put Latino emerging adults at great risk for binge drinking as well as its consequences. Emerging adulthood is a time of exploration, which might include testing new ways of being that challenge traditional gender roles. Exploration of gender roles in different contexts (e.g., college) may yield heavier alcohol use and, thus, greater consequences. For college students, per-



ceptions of other’s drinking or social norms may be important. Male Latinos may engage in high-risk drinking because they are embedded in a male social network that values competitiveness, and they may perceive high levels of drinking by those in their network. Research suggests that there may be differences between Caucasian and Latino college students in whether ethnicity-specific norms predict drinking behaviors (LaBrie, Atkins, Neighbors, Mirza, & Larimer, 2012). LaBrie et al. (2012) found that Latinos’ drinking was more connected to the drinking of the typical student than ethnicity-specific norms on a campus with fewer Latino students. If the typical college student is perceived as a Caucasian male (LaBrie et al., 2012; Lewis & Neighbors, 2006), Latino emerging adults’ drinking behavior while attending predominantly White institutions may be best linked to those students. Drinking by male college students may be driven, in part, by gender role expectations that prize heavy drinking among male students. Recent research on alcohol use consequences suggest that some emerging adults do not perceive some consequences as negative (White & Ray, 2013). Men rate consequences as less bothersome than women. Perhaps gender roles play a role here in that some consequences of drinking are markers of the competitive aspects of masculinity and, within certain contexts, are not experienced as negative. For Latino emerging adults in other contexts, such as Hispanic serving institutions or the work world, the links between gender roles, drinking, and drinking consequences may be strikingly different and less clear. Thus, future research might investigate the extent to which changes in gender roles impact drinking and consequences over time and in particular contexts. The current research also has implications for prevention and intervention with Latino emerging adults. Emerging adults are often negotiating transitions into adult roles (Arnett, 2006). This transition may be occurring in a variety of contexts including, but not limited to, college, the workplace, or the military. Each of these contexts is embedded within its own cultural context, with its own prescriptions about race/ethnicity, gender roles, and alcohol use. For psychologists, it might be important to explore how gender roles and changes in those roles might impact decision making about alcohol use. For example, Latino emerging adults who are transitioning into a college environment may have vastly different experiences with respect to alcohol use if they are transitioning into a predominantly White institution versus a Hispanic serving institution. Latino students may be negotiating decisions around alcohol use in the context of a college campus where alcohol use is quite prevalent and messages about use are connected to gender roles. Individualized interventions that identify the important reference groups for a student at risk for binge drinking may be helpful. This might be followed by exploration of what these reference groups tell emerging adults about whom they are and the links between identity and risky behaviors. Given the paucity of research on cultural as well as established risk factors for drinking among Latino emerging adults, implications for prevention are speculative and based on an incomplete literature. However, psychologists working with Latino emerging adults should be attentive to the role that cultural factors such as gender roles play in high-risk drinking behaviors.

725 References

Arciniega, G. M., Anderson, T. C., Tovar-Blank, Z. G., & Tracey, T. (2008). Toward a fuller conception of machismo: Development of a traditional machismo and caballerismo scale. Journal of Counseling Psychology, 55, 19 –33. doi:10.1037/0022-0167.55.1.19 Arnett, J. J. (2006). Emerging adulthood: Understanding the new way of coming of age. In J. J. Arnett & J. L. Tanner (Eds.), Emerging adults in America: Coming of age in the 21st century (pp. 3–19). Washington, DC: American Psychological Association. doi:10.1037/11381-001 Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting and Clinical Psychology, 42, 155–162. doi:10.1037/ h0036215 Caetano, R., Ramisetty-Mikler, S., & Rodriguez, L. (2009). The Hispanic Americans Baseline Alcohol Survey (HABLAS): The association between birthplace, acculturation and alcohol abuse and dependence across Hispanic national groups. Drug and Alcohol Dependence, 99, 215–221. doi:10.1016/j.drugalcdep.2008.08.011 Castillo, L. G., Perez, F. V., Castillo, R., & Ghosheh, M. R. (2010). Construction and initial validation of the Marianismo Beliefs Scale. Counselling Psychology Quarterly, 23, 163–175. doi:10.1080/ 09515071003776036 Chartier, K., & Caetano, R. (2010). Ethnicity and health disparities in alcohol research. Alcohol Research & Health, 33, 152–160. Chen, C. M., Dufour, M. C., & Yi, H. (2004 –2005). Alcohol consumption among young adults ages 18 –24 in the United States: Results from the 2001–2002 NESARC Survey. Alcohol Research & Health, 28, 269 –280. Choi, N., Fuqua, D. R., & Newman, J. L. (2009). Exploratory and confirmatory studies of the structure of the Bem Sex Role Inventory Short Form with two divergent samples. Educational and Psychological Measurement, 69, 696 –705. doi:10.1177/0013164409332218 Cleveland, M. J., Mallett, K. A., White, H. R., Turrisi, R., & Favero, S. (2013). Patterns of alcohol use and related consequence in non-college– attending emerging adults. Journal of Studies on Alcohol and Drugs, 74, 84 –93. Corbin, W. R., Vaughan, E. L., & Fromme, K. (2008). Ethnic differences and the closing of the sex gap in alcohol use among college-bound students. Psychology of Addictive Behaviors, 22, 240 –248. doi:10.1037/ 0893-164X.22.2.240 Ding, C. S. (2006). Using regression mixture analysis in educational research. Practical Assessment, Research & Evaluation, 11, 1–11. Falicov, C. J. (2010). Changing constructions of machismo for Latino men in therapy: “The devil never sleeps.” Family Process, 49, 309 –329. doi:10.1111/j.1545-5300.2010.01325.x Galanti, G. (2003). The Hispanic family and male–female relationships: An overview. Journal of Transcultural Nursing, 14, 180 –185. doi:10.1177/ 1043659603014003004 Gilbert, M. J., & Collins, R. L. (1997). Ethnic variation in women’s and men’s drinking. In R. W. Wilsnack & S. C. Wilsnack (Eds.), Gender and alcohol. New Brunswick, NJ: Rutgers Center of Alcohol Studies. Harris, K. M., Halpern, C. T., Whitsel, E., Hussey, J., Tabor, J., Entzel, P., & Udry, J. R. (2009). The National Longitudinal Study of Adolescent Health: Research design. Retrieved from http://www.cpc.unc .edu/projects/addhealth/design Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. doi:10.1080/10705519909540118 Iwamoto, D. K., & Smiler, A. P. (2013). Alcohol makes you macho and helps you make friends: The role of masculine norms and peer pressure in adolescent boys’ and girls’ alcohol use. Substance Use & Misuse, 48, 371–378. doi:10.3109/10826084.2013.765479 Johnston, L. D., O’Malley, P. M., Bachman, J. G., & Schulenberg, J. E. (2007). Monitoring the Future National Survey results on drug use, 1975–2006: II. College students and adults ages 19 – 45 (NIH Publication No. 07– 6206). Bethesda, MD: National Institute on Drug Abuse.


726


Kulis, S., Marsiglia, F. F., Lingard, E. C., Neiri, T., & Nagoshi, J. (2008). Gender identity and substance use among students in two high schools in Monterrey, Mexico. Drug and Alcohol Dependence, 95, 258 –268. doi:10.1016/j.drugalcdep.2008.01.019 LaBrie, J. W., Atkins, D. C., Neighbors, C., Mirza, T., & Larimer, M. E. (2012). Ethnicity-specific norms and alcohol consumption among Hispanic/Latino/a and Caucasian students. Addictive Behaviors, 37, 573– 576. doi:10.1016/j.addbeh.2012.01.007 Lewis, M. A., & Neighbors, C. (2006). Who is the typical college student? Implications for personalized normative feedback interventions. Addictive Behaviors, 31, 2120 –2126. doi:10.1016/j.addbeh.2006.01.011 Maggs, J. L., & Schulenberg, J. E. (2004 –2005). Trajectories of alcohol use during the transition to adulthood. Alcohol Research & Health, 28, 195–201. Marsh, H. W., Hau, K. T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11, 320 –341. doi:10.1207/ s15328007sem1103_2 McCreary, D. R., Newcomb, M. D., & Sadava, S. W. (1999). The male role, alcohol use, and alcohol problems: A structural modeling examination in adult women and men. Journal of Counseling Psychology, 46, 109 –124. doi:10.1037/0022-0167.46.1.109 Meyers, L. S., Gamst, G., & Guarino, A. J. (2013). Applied multivariate research: Design and interpretation. Thousand Oaks, CA: Sage. Muthén, L. K., & Muthén, B. O. (2011). MPlus (Version 6.12) [Computer software]. Los Angeles, CA: Author. National Institute on Alcohol Abuse and Alcoholism. (2007). What colleges need to know now: An update on college drinking research. (NIH Publication No. 07–5010). Retrieved from http://www.collegedrinking prevention.gov/1College_Bulletin-508_361C4E.pdf Nylund, K. L., Asparouhov, T., & Muthén, B. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling. A Monte Carlo simulation study. Structural Equation Modeling, 14, 535–569. doi:10.1080/10705510701575396 O’Malley, P. M., & Johnston, L. D. (2002). Epidemiology of alcohol and other drug use among American college students. Journal of Studies on Alcohol, (Suppl. 14), 23–39. O’Neil, J. M. (2008). Summarizing 25 years of research on men’s gender role conflict using the Gender Role Conflict Scale: New research paradigms and clinical implications. The Counseling Psychologist, 36, 358 – 445. doi:10.1177/0011000008317057 Paschall, M. J., Bersamin, M., & Flewelling, R. L. (2005). Racial/ethnic differences in the association between college attendance and heavy alcohol use: A national study. Journal of Studies on Alcohol, 66, 266 – 274. Quinn, P. D., & Fromme, K. (2011). Alcohol use and related problems among college students and their noncollege peers: The competing roles of personality and peer influence. Journal of Studies on Alcohol and Drugs, 72, 622– 632. Rafaelli, M., Stone, R. A. T., Iturbide, M. I., McGinley, M., Carlo, G., & Crockett, L. J. (2007). Acculturation, gender, and alcohol use among Mexican American college students. Addictive Behaviors, 32, 2187– 2199. doi:10.1016/j.addbeh.2007.02.014 Reboussin, B. A., Hubbard, S., & Ialongo, N. S. (2007). Marijuana use patterns among African-American middle-school students: A longitudi-

nal latent class regression analysis. Drug and Alcohol Dependence, 90, 12–24. doi:10.1016/j.drugalcdep.2007.02.006 Safer, A. M., & Piane, G. (2007). Analysis of acculturation, sex, and heavy alcohol use in Latino college students. Psychological Reports, 101, 565–573. doi:10.2466/PR0.101.2.565-573 Schwartz, S. J., Unger, J. B., Zamboanga, B. L., & Szapocznik, J. (2010). Rethinking the concept of acculturation: Implications for theory and research. American Psychologist, 65, 237–251. doi:10.1037/a0019330 Schwartz, S. J., Weisskirch, R. S., Zamboanga, B. L., Castillo, L. G., Ham, L. S. Huynh, Q. L., . . . Cano, M. A. (2011). Dimensions of acculturation: Associations with health risk behaviors among college students from immigrant families. Journal of Counseling Psychology, 58, 27– 41. Strunin, L., Edwards, E. M., Godette, D. C., & Heeren, T. (2007). Country of origin, age of drinking onset, and drinking patterns among Mexican American young adults. Drug and Alcohol Dependence, 91, 134 –140. doi:10.1016/j.drugalcdep.2007.05.013 Torres, J. B., Solberg, S. H., & Carlstrom, A. H. (2002). The myth of sameness among Latino men and their machismo. American Journal of Orthopsychiatry, 72, 163–181. doi:10.1037/0002-9432.72.2.163 Vaughan, E. L., Kratz, L. M., Escobar, O. S., & Middendorf, K. G. (2013). Latino subgroup as a moderator of the relationship between language usage and alcohol use in a national sample of Latino emerging adults. Journal of Latina/o Psychology, 1, 182–194. doi:10.1037/a0033384 Venegas, J., Cooper, T. V., Naylor, N., Handson, B. S., & Blow, J. A. (2012). Potential cultural predictors of heavy episodic drinking in Hispanic college students. The American Journal on Addictions, 21, 145– 149. doi:10.1111/j.1521-0391.2011.00206.x Vermunt, J. K., & Magidson, J. (2008). Latent GOLD (Version 4.5) [Computer software]. Belmont, MA: Statistical Innovations. Vicary, J. R., & Karshin, C. M. (2002). College alcohol abuse: A review of the problems, issues, and prevention approaches. The Journal of Primary Prevention, 22, 299 –331. doi:10.1023/A:1013621821924 Wells, S., Speechley, M., Koval, J. J., & Graham, K. (2007). Gender differences in the relationship between heavy episodic drinking, social roles, and alcohol related aggression in a U.S. sample of late adolescent young adult drinkers. The American Journal of Drug and Alcohol Abuse, 33, 21–29. doi:10.1080/00952990601082613 White, H. R., & Ray, A. E. (2014). Differential evaluations of alcoholrelated consequences among emerging adults. Prevention Science, 15, 115–124. doi:10.1007/s11121-012-0360-8 Wong, Y. J., & Maffini, C. S. (2011). Predictors of Asian American adolescents’ suicide attempts: A latent class regression analysis. Journal of Youth and Adolescence, 40, 1453–1464. doi:10.1007/s10964-0119701-3 Zamboanga, B. L., Rafaelli, M., & Horton, N. J. (2006). Acculturation status and heavy alcohol use among Mexican American college students: Investigating the moderating role of gender. Addictive Behaviors, 31, 2188 –2198. doi:10.1016/j.addbeh.2006.02.018 Zemore, S. E. (2007). Acculturation and alcohol among Latino adults in the United States: A comprehensive review. Alcoholism: Clinical and Experimental Research, 31, 1968 –1990. doi:10.1111/j.1530-0277.2007 .00532.x

Received May 3, 2013 Revision received February 17, 2014 Accepted June 9, 2014 䡲