Associations of Multiple Acculturation Domains with Smoking Status among Latino Adults

Abstract

Few efficacious interventions for tobacco use exist for Latinos. Identification of cultural factors relevant to smoking among Latinos can inform the development of efficacious interventions for Latino smokers. Acculturation is associated with smoking, especially among Latinas, but extant research is limited by operationalization of acculturation with unidirectional, single-domain proxies. We examined associations of multiple domains of acculturation with gender and smoking status among Latino adults. Cross-sectional data from 140 bilingual Latino adults was utilized. Acculturation was measured with the 4 subscales of the Multidimensional Acculturation Scale II (MAS-II). Logistic regression analyses tested interaction effects between MAS-II American and Latino Cultural Identification subscales, English and Spanish Proficiency subscales, and their interactions with gender, on smoking status. Higher English Proficiency was associated with greater odds of being a smoker at Spanish Proficiency scores of 4.5 or higher. Higher Latino Cultural Identification was associated with lower odds of being a smoker among women, but not men. Acculturation toward American culture, per se, may not be a risk factor for smoking; rather, its influence depends on Latino culture maintenance. Unlike in other areas of mental/behavioral health among Latinos, biculturalism may not be protective against smoking. The association between acculturation and smoking among Latinas may be a function of loss of Latino culture identification. Intervention programs should consider targeting these at-risk individuals. Longitudinal work that corroborates current findings and identifies mechanisms underlying these associations is needed.

Tobacco is the leading cause of morbidity and mortality in the United States (Jha et al., 2013; Mokdad, Marks, Stroup, & Gerberding, 2004). Although Latino persons are less likely to smoke cigarettes compared to non-Latino white persons (Centers for Disease Control and Prevention, 2015; Substance Abuse and Mental Health Services Administration, 2014), notable disparities exist among those who do smoke. For example, Latinos who smoke are less likely to be screened for smoking status (Cokkinides, Halpern, Barbeau, Ward, & Thun, 2008), be advised to quit during a health care encounter (Cokkinides et al., 2008; Houston, Scarinci, Person, & Greene, 2005; Levinson, Pérez-Stable, Espinoza, Flores, & Byers, 2004; Lopez-Quintero, Crum, & Neumark, 2006), or use cessation aids (Levinson et al., 2004; Trinidad, Pérez-Stable, White, Emery, & Messer, 2011; Webb Hooper, Rogers, & Okuyemi, 2015). There is a dearth of research on efficacious interventions for smoking cessation among this population, and the few interventions that have been tested among Latinos show limited effects on cessation (Lawrence, Graber, Mills, Meissner, & Warnecke, 2003; Webb, Rodriguez-Esquivel, & Baker, 2010). A significant barrier to intervention development is that little is known about what factors contribute to cessation among Latinos, including knowledge of risk and protective factors for smoking that may be uniquely or differentially important to understudied groups, such as Latinos (Fiore, Jaén, & Baker, 2008). Such research can inform the development or adaptation of interventions to better address the unique needs and preferences of Latinos (Alegría et al., 2006). The current study seeks to address this gap by examining the relation between acculturation and smoking among Latinos, with particular attention to multiple domains of acculturation.

Acculturation

Acculturation is a multidimensional process whereby individuals acquire the behaviors/practices, values, and identity associated with the host (e.g., American) culture while maintaining the behaviors/practices, values, and identity associated with the heritage (e.g., Latino) culture (Schwartz, Unger, Zamboanga, & Szapocznik, 2010). Each cultural dimension (e.g., acquisition of the host culture and maintenance of the heritage culture) has been shown to encompass at least three domains: (a) behaviors/practices—customs and language; (b) values—system of beliefs; and (c) identity—sense of attachment and esteem derived from attachments (Schwartz et al., 2010). The extent to which persons retain or adopt behaviors/practices, values, and identifications within or across cultural domains can vary, suggesting the need to examine acculturation separately across domains to more accurately understand its influence (Schwartz et al., 2010). In addition, a multidimensional conceptualization of acculturation accounts for biculturalism, the simultaneous identification with heritage and host culture influences (Schwartz & Unger, 2010; Schwartz et al., 2010), further suggesting the need to examine interrelationships among acculturation domains.

Acculturation and Smoking Among Latinos

Much research exists on the relationship between acculturation and smoking among Latinos (e.g., Bethel & Schenker, 2005; Kondo, Rossi, Schwartz, Zamboanga, & Scalf, 2016; Lara, Gamboa, Kahramanian, Morales, & Bautista, 2005; Wilkinson et al., 2005). However, it is limited by a predominant use of demographic proxies as indicators of acculturation (e.g., nativity, years of U.S. residency, familial generations in the United States) or short questionnaires that assess a single domain (usually the behavioral domain) and direction of acculturation (usually toward the U.S. culture). These approaches have been widely criticized for questionable construct validity, failure to capture the bidirectional nature of acculturation, and failure to capture multiple domains of acculturation (Abraído-Lanza, Armbrister, Florez, & Aguirre, 2006; Schwartz et al., 2010).

Notwithstanding these limitations, extant research has consistently found that greater acculturation toward American culture is associated with smoking among women, but not men (Bethel & Schenker, 2005; Kondo et al., 2016; Lara et al., 2005; Wilkinson et al., 2005). Researchers posit this is because of the internalization of American social norms that may appraise smoking among women as both more normative and acceptable compared with the Latino culture (Bethel & Schenker, 2005; Wilkinson et al., 2005). However, research that clarifies which domains of acculturation are most associated with smoking among Latino adults is lacking. The limited research is restricted to adolescents and is mixed, with one study finding Latino cultural identity to be associated with lower odds of smoking (Unger, Ritt-Olson, Wagner, Soto, & Baezconde-Garbanati, 2009), one study finding Latino cultural identity to be associated with greater odds of smoking among girls (Lorenzo-Blanco, Unger, Ritt-Olson, Soto, & Baezconde-Garbanati, 2011), and one study finding no association of specific acculturation domains with smoking (Lorenzo-Blanco, Unger, Ritt-Olson, Soto, & Baezconde-Garbanati, 2013).

Biculturalism and Smoking Among Latinos

Biculturalism refers to strong acculturation toward both the host and heritage cultures (Schwartz & Unger, 2010; Schwartz et al., 2010), and is argued to be the optimal cultural adaptation strategy by allowing persons to better navigate and cope with the diverse demands of living in two cultural worlds (Rogler, Cortes, & Malgady, 1991; Schwartz et al., 2010). Biculturalism has generally been associated with better psychosocial outcomes (Nguyen & Benet-Martínez, 2012). Studies among adults that have examined substance use have generally found that greater endorsement of the heritage culture domains is protective against substance use, while greater endorsement of American culture domains is risk-bearing (Schwartz et al., 2011; Unger, Schwartz, Huh, Soto, & Baezconde-Garbanati, 2014). Although these studies shed some light on the independent relations of the heritage and American acculturation domains with substance use, they do not speak to biculturalism, which would involve examination of synergistic effects (i.e., interaction effects). No known published study has simultaneously examined both the independent and synergistic effects of multiple directions and domains of acculturation for their relations to smoking among Latino adults.

Current Study

The purpose of the current study was to examine synergistic associations of acculturation measured across multiple domains and including orientation toward both American culture and Latino culture with smoking status among Latino men and women. Consistent with previous research demonstrating the protective effects of biculturalism, we hypothesized that participants with both strong American and Latino cultural orientations would exhibit the lowest likelihood of smoking. We also explored whether these associations differed between men and women. Consistent with previous research in substance abuse and smoking, we hypothesized that greater acculturation in the Latino cultural domains would be associated with lower likelihood of smoking and greater acculturation in the American cultural domains would be associated with higher odds of smoking among women, but not among men.

Method

Participants

Participants were 140 bilingual Latino residents of a major metropolitan area located in the south-central United States. Inclusion criteria were (a) self-identification as Latino/a, (b) age between 18 and 65 years, (c) possession of a valid home address and a functioning telephone number, and (d) demonstration of marginal to adequate health literacy in English, as indicated by a score of at least 45 on the Rapid Estimate of Adult Literacy in Medicine (Davis et al., 1993), and Spanish, as indicated by a score of at least 38 on the Short Assessment of Health Literacy for Spanish Adults (Lee, Stucky, Lee, Rozier, & Bender, 2010). Bilingual Latino participants were recruited to ensure variability in both American and Latino cultural orientations. Bilingualism was utilized as an indicator of biculturalism, as research has consistently shown that language proficiency accounts for most or a large proportion of variance in acculturation measures (e.g., Cuéllar, Arnold, & Maldonado, 1995; Rodriguez, Mira, Paez, & Myers, 2007).

Exclusion criteria were (a) having another household member enrolled in the study, (b) use of an illicit substance in the past 30 days, (c) currently pregnant or breastfeeding, (d) use of other tobacco products other than cigarettes, (e) participation in a smoking cessation program in the last 90 days, or (f) current use of nicotine replacement products. The first criterion was utilized to avoid the potential confound of clustering in the data, and the second through fifth criteria were utilized to avoid because they are special populations of smokers for whom specific study is warranted and preferable, as our hypotheses regarding the relationships with smoking are not necessarily applicable in the context of the unique factors of these populations. Therefore, the study of smoking cessation mechanisms among adolescent smokers requires a separate focus on factors that are relevant for that particular population. The last two criteria were utilized because the use of nicotine replacement therapy may interfere with self-report on other constructs that were assessed in the larger data collection project from which the data utilized in this study were obtained. (e.g., tobacco dependence, cigarette craving, cigarette consumption).

Procedures

The current study utilized data from a longitudinal project designed to examine psychosocial and contextual influences on multiple health risk behaviors among Latinos. Participants were recruited via local newspaper and radio advertisements and in-person outreach. Participants provided data across two in-person visits (on Day 1 and Day 7) to the research site and 6 days (Days 2-7) of ecological momentary assessment (EMA; Shiffman, 2009). The current study utilized only data collected at the first in-person assessment. Participants were screened for all eligibility criteria except health literacy via telephone prior to the first visit. Health literacy was assessed during the first in-person assessment to determine final eligibility. Participants could be compensated up to $120 for their time and effort. Participants were compensated $30 for the first in-person assessment, $20 for the second in-person assessment, and up to $70 for EMA assessments, which was prorated according to percentage of EMA completed. This research was approved by the Institutional Review Board of The University of Texas MD Anderson Cancer Center.

Measures

Demographics

Participant demographics included self-reported age, gender, years of education completed, and proportion of life spent in the United States. Proportion of life spent in the United States was computed as the quotient of self-reported years living in the United States and age.

Acculturation

Acculturation domains were assessed with the Multidimensional Acculturation Scale-II (MAS-II; Rodriguez et al., 2007). The MAS-II is a 22-item self-report measure that comprises the following four subscales: English Proficiency, Spanish Proficiency, American Cultural Identification, and Latino Cultural Identification. Sample items for each subscale include: “How well do read English?”; “How well do you read Spanish?”; “How important is it for you to be integrated into the American mainstream?”; and, “How important is it for you to be connected to your ethnic community?” Depending on the item, participants were asked to either rate their responses on a scale from 1 (cannot speak it or understand it) to 6 (very fluent) or on a scale from 1 (not well/not at all) to 5 (very well/very much). Mean scores can range from 1.1 to 5.2 on the language proficiency scales, and 1 to 5 on the cultural identity scales. Higher mean scores indicate greater acculturation on the respective factor.

Validity and reliability of MAS-II has been previously demonstrated with adults of Mexican origin (Rodriguez et al., 2007) and Mexican American and Central American college students (Rodriguez, Mira, Myers, Morris, & Cardoza, 2003). Composite reliability for the MAS-II subscales in this sample were as follows: English Proficiency = .87, Spanish Proficiency = .89, American Cultural Identification = .77, and Latino Cultural Identification = .83.

Smoking status

Participants were classified as either a non-smoker or smoker based on their response to the item “Do you currently smoke: (a) every day; (b) some days, or; (c) not at all?” Persons who reported smoking every day or some days were classified as smokers, and persons reporting not smoking at all were classified as nonsmokers. In all analyses, status as a smoker was coded as 1.

Data Analysis

Data was examined for normality utilizing critical values of 2.0 for skewness and 4.0 for kurtosis (Kim, 2013), and for multicollinearity using critical values of less than .10 for tolerance and greater than 10 for variance inflation factor (VIF; O’Brien, 2007). Sample descriptive statistics, bivariate correlations, and proportion of missing data for each variable were computed. Multiple imputation was utilized to account for all missing data, and results from 20 imputed data sets were pooled according to Rubin (1987). The primary data analytic approach for this study was logistic regression using maximum likelihood estimation with robust standard errors in Mplus Version 7.2. Smoking status (1 = “smoker”) was the outcome variable in each analysis. All analyses controlled for age, years of education completed, proportion of life in the United States, and gender (in analyses where it was not the moderator). Age, years of education completed, proportion of life in the United States, and gender were included as covariates in our analyses because past research has shown that acculturation and smoking patterns differ as a function of these variables among Latinos (e.g., Bethel & Schenker, 2005; Cuéllar et al., 1995; Pérez-Stable et al., 2001; Rodriguez et al., 2007).

Interaction tests of acculturation dimensions consisted of two analyses. One analysis included English Proficiency, Spanish Proficiency, and the interaction term of these two scales as predictor variables. The other included American Cultural Identification, Latino Cultural Identification, and the interaction term of these two scales as predictor variables. For the purposes of interpretation, English Proficiency and American Cultural Identity were designated the focal predictors and Spanish Proficiency and Latino Cultural Identity were designated the moderators in their respective analyses. Significant interactions were probed to identify regions of significance utilizing Johnson-Neyman plots (Johnson & Fay, 1950; Preacher, Rucker, & Hayes, 2007), which were supplemented with graphical depictions of predicted probabilities of being a smoker as a function of the focal predictor. Four analyses were additionally conducted examining gender, each of the four acculturation domains, and the interaction of each acculturation domain with gender. For the purposes of interpretation, the acculturation scales were designated as the focal predictors, and gender was designated as the moderator.

Results

Preliminary Analyses

Skewness and kurtosis statistics were in the acceptable range for all continuous variables except years of education completed (skewness = −1.13, kurtosis = 5.15). Tolerance and VIF were in the acceptable range for all variables. There was no participant missing data for age, gender, education, and smoking status. Participant missing data for the remaining variables ranged from 0.7% to 7.14%.

Participant Characteristics

Seventy-eight participants (55.7%) self-identified as Mexican/Mexican American, 18 (12.9%) as Central American, nine (6.4%) as South American, three (2.1%) as Puerto Rican, one (0.7%) as Spaniard, 21 (15.0%) of multiethnic Latino ethnicities, six (4.3%) as multiracial between Latino and other race/ethnicities (i.e., White and Native American), and four (2.9%) did not report their ethnicity. Other sample descriptive statistics and bivariate correlations among variables are summarized in Table 1. Age, education, male gender, and Spanish Proficiency each demonstrated associations with smoking status in the small to moderate range (per Cohen, 1992).

Table 1.

Sample Descriptives and Variable Correlations

Variable	1	2	3	4	5	6	7	8	9
1. Age	36.06 (12.30)
2. Education	−.11	13.12 (2.47)
3. Gender (1=male)	.16	−.11	62 (44%)
4. Proportion of life in United States	.04	−.10	.20*	.83 (0.28)
5. American identification	.23**	−.05	−.06	.15	4.25 (0.69)
6. English proficiency	.03	.47**	.11	.46**	.09	4.67 (0.64)
7. Latino identification	.04	.03	−.17*	−.11	.48**	−.12	4.27 (0.72)
8. Spanish proficiency	−.22**	.23**	−.20*	−.52**	−.13	−.16	.20*	4.20 (0.84)
9. Smoking status (1=smoker)	.25**	−.50**	.33**	.05	.02	−.07	−.16	−.24*	63 (44%)

Note. Correlations among two continuous variables are Pearson’s r; correlations with one or two categorical variables are Spearman’s rho. Means and standard deviations for continuous variables and Ns and sample proportions for dichotomous variables are listed in the diagonal.

^*p < .05.

^**p < .01.

Interactions of Acculturation Dimensions

The interaction of American Cultural Identity and Latino Cultural Identity was not significantly associated with smoking status, nor were the main effects of either scale. There was a significant interaction effect of English Proficiency and Spanish Proficiency on smoking status (unstandardized estimate [b] = 1.29, standard error of estimate [SE b] = 0.53, p = .02, adjusted odds ratio [AOR] = 3.64, 95% confidence interval [CI] [1.29, 10.26]; Table 2). A Johnson-Neyman plot of the conditional effect of English Proficiency on smoking status indicated that the relationship between English Proficiency and smoking status is positive and statistically significant at Spanish Proficiency scores of 4.5 and higher (see Figure 1). Figure 2 depicts the probabilities of being a smoker predicted by English Proficiency at the sample’s mean Spanish Proficiency score and two Spanish Proficiency scores within the region of significance.

Table 2.

Smoking Status Regressed on Demographics and Acculturation

Acculturation domain and predictor	Estimate (SE b)	p value	AOR	95% CI
Language proficiency
Age	.03 (.02)	.12	1.03	[.99, 1.06]
Education	−.59 (.15)	.00	.55	[.41, .74]
Gender	1.18 (.43)	.01	3.25	[1.40, 7.52]
Proportion of years in the United States	−2.35 (1.15)	.04	.10	[.01, .90]
English proficiency	−4.88 (2.31)	.03	.01	[.00, .70]
Spanish proficiency	−6.15 (2.53)	.01	.00	[.00, .21]
English Proficiency × Spanish Proficiency	1.29 (.53)	.02	3.64	[1.29, 10.26]
Cultural identity
Age	.03 (.02)	.13	1.03	[.99, 1.06]
Education	−.43 (.13)	.00	.65	[.51, .83]
Gender	1.11 (.41)	.01	3.03	[1.34, 6.81]
Proportion of years in the United States	−.67 (.98)	.49	.51	[.08, 3.45]
American cultural identity	−.67 (1.20)	.58	.51	[.05, 5.43]
Latino cultural identity	−1.26 (1.31)	.34	.28	[.02, 3.69]
American Cultural Identity × Latino Cultural Identity	.21 (.30)	.49	1.23	[.68, 2.22]

Note. N = 140. Separate regression analyses were conducted with each acculturation domain. SE b = standard error of unstandardized estimate; AOR = adjusted odds ratio; CI = confidence interval.

Figure 1.

Conditional effect of English Proficiency on smoking status. Solid line represents the unstandardized regression coefficient of the conditional effect. Dotted lines represent upper and lower 95% confidence intervals. Gray area indicates region of Spanish Proficiency scores at which the conditional effect of English Proficiency is statistically significant.

Figure 2.

Relation between English Proficiency and predicted probability of being a smoker at high levels of Spanish Proficiency. Spn = Spanish Proficiency score. Solid lines represent significant relations; dotted line represents nonsignificant relation.

Interaction Tests of Acculturation Domains and Gender

The interaction effect of gender and Latino Cultural Identity on smoking status was marginally significant (b = 0.91, SE b = 0.50, p = .07, AOR = 2.48, 95% CI [0.93, 6.62]). Follow-up analyses indicated that higher Latino Cultural Identity scores were significantly associated with lower odds of being a smoker among women (b = −0.82, SE b = 0.40, p = .04, AOR = 0.44, 95% CI [0.20, 0.96]) but not among men (b = 0.08, SE b = 0.32, p = .79, AOR = 1.08, 95% CI [0.58, 2.03]). Figure 3 depicts the predicted probability of being a smoker as a function of Latino Cultural Identity scores for men and women. No other interaction effects or main effects of gender and acculturation were statistically significant.

Figure 3.

Relation between Latino Cultural Identity and predicted probability of being a smoker for men and women. Solid line represents significant relation; dotted line represents nonsignificant relation.

Discussion

The current study is among the first to simultaneously examine multiple domains and directions of acculturation for their independent and multiplicative relations with smoking status, as well as their interaction effects with gender, among Latino adults. Previous research has shown that English language preference is associated with higher odds of smoking and Spanish language preference is associated with lower odds of smoking, particularly among women (Bethel & Schenker, 2005; Kondo et al., 2016; Lara et al., 2005). The current study demonstrated a more complex relationship. Based on past research demonstrating the benefits of biculturalism on psychosocial outcomes (e.g., Nguyen & Benet-Martínez, 2013), we hypothesized that participants with bicultural orientations would exhibit the lowest odds of smoking. However, our findings did not support these predictions. Study results provide no evidence of an association of bicultural self-identification with smoking status. What is more, results regarding bilingualism suggest a disadvantageous association with smoking status.

Greater English proficiency among persons with high Spanish proficiency is associated with increased risk of being a smoker. Among persons with low levels of Spanish proficiency, there was no association between English proficiency and smoking status. Although the current findings cannot speak to cause and effect, they bring forth the hypothesis that, among persons who have low proficiency in Spanish, adoption of the English language is not a risk factor for smoking. Only for those who have strong Spanish language skills is the adoption of English associated with increased likelihood of smoking. Thus, efforts to understand through what mechanisms English language adoption may be linked to increased risk of smoking may only be relevant to native Spanish speakers who regard English as their second language. This may include Latinos who are immigrants to the United States or U.S-born Latinos who were reared and/or live in environments wherein they are first exposed to and immersed the Spanish language early on in life.

Study findings further raise the question of what factors may account for the association between English proficiency and smoking status among persons with high Spanish proficiency. One possibility is that increasing English language proficiency may result in more access to and interactions with, Americanized social networks that endorse more positive smoking norms (Unger et al., 2000). For example, in a study among 1,411 Latinas, bilingual and English-dominant speakers were less likely, compared with Spanish-dominant speakers, to recognize the detriments of smoking, but more likely to see the individual and social advantages of smoking (Campbell & Kaplan, 1997). Indeed, language is regarded as the primary mechanism through which cultural information is transmitted (Kang, 2006).

Other possible mechanisms through which bilingualism may relate to smoking behavior might be the experience of acculturative stress (Rodriguez, Myers, Mira, Flores, & Garcia-Hermandez, 2002) or bicultural stress (Romero & Roberts, 2003). Acculturative stress and bicultural stress generally refer to distress resulting from navigating and experiencing conflicts between two cultural orientations. Bilingual individuals may have greater exposure to such experiences, and this stress may increase the probability of engaging in unhealthy behaviors, such as smoking. Limited research among adolescents, for whom acculturative stress appears to be associated with odds of smoking, supports this idea (Lorenzo-Blanco & Unger, 2015). No known research has examined relations between these cultural stressors and smoking among Latino adults. Further research is needed to examine and elucidate the social-cognitive mechanisms between being bilingual and increased risk of smoking among Latino adults.

Neither the independent effects of American and Latino cultural identification nor their interaction were associated with smoking status in the current study. Thus, we did not find support for acculturation in self-identification. These null findings are nevertheless significant, in that they underscore the importance of accounting for the multidimensional nature of acculturation whose facets may have differential relationships with a given outcome, as demonstrated here. It is also notable that current findings among Latino adults differ from previous work among Latino adolescents. At least one study has demonstrated cultural identification to be negatively associated with smoking among Latino adolescents (Castro, Stein, & Bentler, 2009). These findings raise the possibility that acculturation in the self-identification domain may have differential importance to smoking across development levels.

The current study was unable to replicate the gender differentiated association of acculturation with smoking status. This is particularly surprising in light of the robustness of this finding in the broader published research. However, considerable reductions in the ability to detect an effect are to be expected with a dichotomous moderator (vs. a continuous moderator, as in the main analyses power; Aguinis, Boik, & Pierce, 2001). One of these four analyses approached the common threshold for statistical significance. Follow-up analyses revealed that greater Latino cultural identification was associated with lower likelihood of being a smoker among women but not among men. Although these follow-up analyses should be regarded as exploratory, they make way for the hypothesis that the association between acculturation and smoking among Latinas observed in the broader published research may be a function of decreases in Latino culture identification rather than increases in American culture identification. This hypothesis is consistent with speculation that Latinas may relinquish the more restrictive attitudes and social norms regarding women’s smoking associated with their heritage culture (Bethel & Schenker, 2005) as well as with research indicating that greater Latino culture identification is associated with lower odds of substance use (Schwartz et al., 2011; Unger et al., 2014).

Limitations

The current study has some limitations. First, although it allowed for the detection of significant interaction effects among two continuous variables, a sample size of 140 may be considered small and limit our ability to detect significant interactions with a categorical variable, such as gender, as such analyses generally suffer from lower power (Aguinis et al., 2001). Relatedly, although a highly worthwhile additional set of analyses would have been to examine the three-way interaction of gender and measures of each direction of acculturation, or to examine the association of acculturation domains with smoking topography variables (e.g., daily vs. nondaily smoker, number of cigarettes per day), sample size precluded such analyses.

Second, the cross-sectional design of the study precludes inferences regarding causality. Longitudinal work that corroborates current findings and identifies mechanisms underlying these associations is needed. Third, as all participants were bilingual in English and Spanish, the sample demonstrated moderate to high mean levels of both American and Latino cultural identifications. Therefore, current findings are limited to persons who have some degree of identification with each culture and some degree of proficiency with each language and may not generalizable to persons who are strongly monocultural/monolingual. While a bicultural sample is logically appropriate for questions regarding biculturalism, it is not clear whether or not it is appropriate to include persons who are strongly monocultural/monolingual in studies that examine the interactive effects of acculturative directions. Such an approach assumes monocultural/monolingual status part of the same process as biculturalism. Acculturation theory would benefit from research that clarifies whether monolingual/monocultural persons should be treated as a distinct group or as the lowest end of a spectrum of bilingualism/biculturalism.

Fourth, multiple subgroups were combined to represent our Latino adult sample. Latino subgroups have similar cultural aspects; however, group differences still exist that may differentially impact their experiences of acculturation (Alegría et al., 2004; Schwartz et al., 2010). Fifth, although the multidimensional assessment of acculturation is a major strength of the study, the measure utilized here is limited by the fact that it does not assess acculturation in the values domain and that it specifically assesses language proficiency rather than acculturation in the behavioral domain, more broadly. Last, we did not separate out former smokers and perform independent analyses for this group because of their small number (there were four former smokers in the nonsmokers group). Former smokers may be characteristically different from nonsmokers and differentially impacted by acculturation.

Conclusions

In summary, the current study utilized a multidimensional conceptualization of acculturation – an approach that is still rare in tobacco research. The current study is a notable step forward in research on acculturation and smoking, which has largely relied on narrow measures and proxies of acculturation that represent single domains and directions of acculturation. Study results may provide some clarity and specificity in the nature of the relationships among gender, acculturation, and smoking among Latinos. What accounts for the associations between acculturation domains and smoking demonstrated here remains to be understood. However, explanations that focus on the potential independent effect of acculturation toward American culture are limited in the context of the current findings, which underscore the importance of the synergism between cultural orientations.