Sage Journals: Discover world-class research

Abstract

It is well established that immigrant adolescents have lower smoking rates than their native-born counterparts. Although smoking rates among immigrants have been theorized to increase with U.S. acculturation, this hypothesis has seldom been tested using longitudinal data spanning multiple developmental stages. The authors address this limitation using data from the National Longitudinal Study of Adolescent to Adult Health to model age-based smoking trajectories by gender and nativity status among Asian Americans (ages 10–33 years), adjusting for a range of control covariates. Trajectory analyses indicate that the gap between immigrants and natives generally increases as individuals age, but this process varies by gender, with immigrant women exhibiting a significantly less steep smoking growth trajectory (b = −.011, p < .001) compared with native-born men (and all other nativity-gender combinations), whereas immigrant men show no significant smoking trajectory slope difference compared with native men. In summary, results suggest a gendered acculturation process for smoking behavior among Asian Americans.

Keywords

acculturation adolescents Asian Americans gender smoking

Although in decline since 1964, cigarette smoking remains the single most preventable risk factor for death in the United States (Centers for Disease Control and Prevention 2014). And although aggregate national estimates show relatively low smoking rates among Asian Americans, smoking remains a particularly pressing concern for Asian American adolescents and adults (Chae, Gavin, and Takeuchi 2006; Kandula et al. 2009). This is because smoking prevalence exhibits substantial variation across Asian American ethnicities, with some ethnic groups, such as Korean Americans, reporting strikingly higher smoking rates than the U.S. national average (Kim, Son, and Nam 2005).

Research indicates that smoking habits begin to develop during adolescence (Umberson, Crosnoe, and Reczek 2010) and that the risk for adopting the smoking habit is higher among acculturated Asian American adolescents, such as those with native-born status, more developed English-speaking skills, and longer duration of stay in the United States (Ma et al. 2004; Unger et al. 2000, 2004; Weiss and Garbanati 2006). However, the shortage of longitudinal studies of Asian American smoking, spanning multiple developmental periods, restricts our understanding of process of the developing smoking behavior by immigrants over time, as most estimates come from cross-sectional data (Umberson et al. 2010). Extant longitudinal studies of immigrant assimilation and/or acculturation have focused mostly on the role of socioeconomic achievements, especially among Latinx immigrants (Haller, Portes, and Lynch 2011; Zhou and Xiong 2005), leaving the acculturative effects on behavioral health in Asian Americans relatively understudied.

We investigate temporal changes in smoking across adolescence to adulthood, examining the effects of a series of acculturation, social, and psychological variables. The negative acculturation hypothesis, which holds that immigrants will attain a similar behavioral health profile to their native-born counterparts with a lengthier stay in the host country (Ro 2014), serves as the overarching theoretical framework of this study. We also emphasize the gendered aspect of acculturation and smoking behavior by considering the selective acculturation hypothesis (Weiss and Garbanati 2006). As we are assessing longitudinal smoking patterns from adolescence to adulthood, we strengthen the overall theoretical framework by discussing the “time and timing” principle of the life-course perspective, which suggests time and timing of occurrence as critical to shaping long-term health and behavioral outcomes such as smoking (Gong et al. 2011).

Background

Acculturation and Smoking among Asian Americans

The lower rates of smoking among Asian American adolescent immigrant boys and girls compared with their native-born counterparts are often interpreted through the lens of the immigrant health and behavioral paradox (Marks, Ejesi, and García Coll 2014). That is, researchers view the healthier behavior as a paradox because immigrant adolescents fare better in behavioral health than their native-born counterparts, despite living in households with fewer socioeconomic resources and numerous postmigration challenges (Marks et al. 2014; Raymond 2011; Ro 2014). Much research has focused on explaining this paradox, which contradict the expected pattern of increased smoking risk associated with low socioeconomic status (SES) and stress (e.g., Marks et al. 2014). Some scholars suggest that more immigrant youth than native-born youth reside in intact families and receive greater parental supervision (Bui 2013; Marta and Ron 2011). Social control effects stemming from greater religious involvement in immigrant households, relative to native-born households, may also buffer immigrant adolescents against adopting negative health behavior such as smoking (Salas-Wright et al. 2016).

Against the backdrop of the immigrant behavioral paradox in smoking, mounting evidence has emerged on the progressive decline of protective health behaviors over time in immigrant populations. Studies in this vein indicate that smoking increases among immigrants with greater acculturation, as indicated by the lengthier duration of stay (Gorman, Lariscy, and Kaushik 2014; Ma et al. 2004; Maxwell, Bernaards, and McCarthy 2005; Zhang and Wang 2008) and greater English language use (Ma et al. 2004; Tang, Shimizu, and Chen 2005; Unger et al. 2000). The findings are consistent in Asian American women samples, as demonstrated by a meta-analysis of nine cross-sectional studies demonstrating that smoking rates among Asian immigrant women tend to rise with a greater degree of linguistic acculturation and longer stay in the United States (Choi et al. 2008).

Despite these suggestive cross-sectional findings, it remains an open question why immigrant populations, which show lower initial smoking prevalence, move toward attaining comparable smoking rates as their U.S.-born counterparts. The negative acculturation hypothesis proposes that the protective elements, including familial protection, religious control, and ethnic and community resources, dwindle with a greater degree of acculturation into mainstream U.S. society (Ro 2014). A growing body of research supports the negative acculturation hypothesis showing that immigrants do indeed tend to develop negative health behaviors such as poor diet, substance use, and smoking with longer duration of stay in the United States (Gfroerer and Tan 2003; Ro 2014). Research considering generational differences in health behavior also yield empirical support for negative acculturation as smoking rates rise in the successive generation (Acevedo-Garcia et al. 2005; Ahmmad and Adkins 2021). Thus, a growing body of empirical research has indicated that the salubrious influence of Asian sociocultural practices tends to diminish over the first three immigrant generations (Acevedo-Garcia et al. 2005).

Although the empirical support for the negative acculturation hypothesis is growing, the findings remain inconclusive because of overreliance on cross-sectional data and study designs (Marks et al. 2014). In these studies, scholars have included length of residence and generational status as key temporal indicators to understand the acculturative process and associated health behavior change (Acevedo-Garcia et al. 2005; Ahmmad and Adkins 2021; Ro 2014). Unfortunately, the retrospective reports on immigrants’ duration of stay in the United States are subjective and vulnerable to recall bias (Bell et al. 2019). As most of the cross-sectional studies have relied on adult immigrant samples, these study findings may be confounded by the effects of immigrant maintenance of transnational networks with countries of origin. For example, frequent contacts with extended family members in countries of origin through remote communication and in-person visits may slow the acculturative process (Ferguson and Bornstein 2012). However, unlike adult immigrants, adolescent immigrants have fewer opportunities to maintain transnationalism, passing their formative years in the United States and embedding in extended U.S.-based social networks, learning to speak in English, and adopting the American culture lifestyle (Motti-Stefanidi 2018). Thus, analyzing immigrant adolescents transitioning to adulthood in the United States may help overcome many of the limitations that characterize research focused on adult samples.

Moreover, given the lack of research using prospective data following individuals from adolescence to adulthood, social factors contributing to longitudinal changes in health behavior remain unclear (Marks et al. 2014). Previous research has also been limited by inadequate consideration of potential confounding due to background characteristics, such as immigrant household demographic factors. Indeed, much research has indicated that immigrant household characteristics such as parental income, education, and smoking are powerful determinants of adolescent smoking (Motti-Stefanidi 2018; Salas-Wright et al. 2016). Considering these limitations, leading researchers have called for longitudinal analyses using panel data to understand the long-term pattern of loss in adolescent immigrant health behavior (Marks et al. 2014).

Using data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), we address these challenges, exploring how Asian American smoking behavior varies across nativity, gender, and age. In examining the mechanisms of smoking behavior development, we consider acculturative effects on smoking and the effects of depressive symptoms on smoking prevalence across the early to middle life course. Moreover, given that the factors protecting against smoking remain unexplored in the Asian American population, we investigate the effects of family support and religiosity. These findings on risk and protective factors of smoking can collectively inform policymaking around smoking prevention focused on diverse racial and ethnic experiences and advance our understanding of longitudinal acculturation processes in the United States.

Selective Acculturation and Gendered Acculturation in Smoking Behavior

Gender is an essential social determinant of health behavior in immigrant and racial/ethnic minority populations in the United States, in part because both the adoption of behavior such as smoking and the acculturative process could be gendered (Ahmmad and Adkins 2021; Weiss and Garbanati 2006). Among Asian American adolescents, gendered roles and expectations can predispose individuals to divergent patterns of social contacts, peer-network formation, or acculturation, thereby producing different levels of risk for smoking across men and women (Gorman et al. 2014; Marsiglia et al. 2010). Thus, we discuss current theories focused on gender and smoking in the Western context and the gendered acculturation hypothesis, which theorizes gender-stratified risks for smoking among racial/ethnic minorities in the United States.

Smoking rates between men and women differ widely in developing nations; however, smoking rates between men and women in Western countries have tended to converge over time (Pampel 2001). Early scholars viewed the narrowing of smoking rates across gender in the West primarily as due to the societal movement toward gender equality, marked by reduced social constraints on women and, consequently, higher participation of women in tertiary education and accompanying gains in income, power, and prestige (Pampel 2006). Subsequent studies, however, revealed empirical support for the gender equality hypothesis to be incomplete. For instance, as Pampel (2001) suggested, shifts toward parity in smoking in the West are associated with the long history of cigarette use and the resultant process of cigarette diffusion from men to women (Pampel 2001). Pampel examined the cigarette diffusion hypothesis across different cultural contexts. In his findings, the United States and France, for instance, both had long histories of cigarette use and initially high cigarette consumption among men and very low smoking prevalence among women. Increasingly, male smoking rates have receded while female smoking rates have risen, perhaps because of aggressive commercial marketing targeting women (Amos 1996), resulting in a pattern of convergence in cigarette smoking rates between men and women (Amos 1996; Pampel 2001, 2006).

However, for investigating smoking among immigrants and ethnic minorities, the predominant perspectives on gender and smoking, cigarette diffusion and gender equality may offer incomplete explanations, because contemporary ethnic minorities’ assimilation in the United States is segmented rather than linear (Pampel et al. 2019; Ryabov 2015; Xie and Greenman 2005). That is to say, for non-White post-1965 immigrants originating predominantly from Asia and Latin America, straightforward assimilation into a White mainstream has been uncertain (Zhou and Xiong 2005). For instance, Asian Americans, even those residing in the United States for multiple generations, have sometimes been characterized as “perpetual foreigners” for visible non-European physical features (Ahmmad and Adkins 2021), which might affect the Asian American assimilation process. Leading scholars examining Asian American assimilation in the United States have argued that clustering into disadvantaged neighborhoods and ethnic enclaves by recent immigrants provides fewer opportunities for mainstream integration (Haller et al. 2011; Zhou and Xiong 2005).

For assessing smoking in Asian Americans, the gendered acculturation hypothesis, suggesting different pathways of behavioral acculturation between men and women, can be a useful framework. Focused primarily on non-European immigrants, the hypothesis points mainly to the divergent risks of negative acculturation for women and men regardless of their socioeconomic achievements (Pampel et al. 2019; Ryabov 2015; Xie and Greenman 2005). The gender-based norms around smoking or drinking exercised in many countries of origin may underlie the difference in acculturation, which may sustain through immigrant social networks within ethnic enclaves or immediate social contexts (Ahmmad and Adkins 2021; Kandula et al. 2009; Walton 2012). Research conducted among Mexican American immigrants shows that a cultural “double standard” exists around alcohol use, such that women often experience social stigma when consuming alcohol, while men are often encouraged to drink (Marsiglia et al. 2010). Likewise, a considerable culture of stigma discourages women from smoking in many developing nations, including many Asian and Latin American immigrant-sending countries (Gorman et al. 2014). The gendered norms and enforcement of social control might substantially shape adolescent Asian American immigrant girls’ smoking during the transition to adulthood.

Gendered Acculturation in Smoking across the Life Course

Whether the gendered acculturation is limited to particular developmental periods or continues across the life course is largely unknown. However, the “time and timing” principle of the life-course perspective, which emphasizes the timing of occurrence as a crucial determinant of one’s long-term health and behavioral outcomes, may be useful in this regard (Gong et al. 2011). As adolescents transition to adulthood, they may experience a weakening of social control and, in turn, gain increased opportunities to broaden social contacts or expedite the acculturation process (Bui 2013; Marta and Ron 2011; Salas-Wright et al. 2016). With the attenuation of nation of origin cultural influence and rising acculturation in the transition from adolescence to adulthood, whether men and women are equally at risk for smoking can better be framed in the light of the “time and timing” principle of the life-course framework (Gong et al. 2011). The “time and timing” principle in the life-course framework suggests that the exposure in adolescence may have a lasting impact throughout one’s life course, or those who develop a habit of regular smoking in adolescents may continue into adulthood. Previously, the timing of migration or “age at migration” has been found to strongly determine immigrants’ health behaviors, with those immigrating in adolescence vs. adulthood having worse behavioral health outcomes in adulthood (Barr et al. 2016; Gong et al. 2011; Roshania, Narayan, and Oza-Frank 2008; Schooling et al. 2004).

Similarly, the timing of gendered acculturation around smoking may have a lasting effect across the life course of the immigrant population, resulting in diverging smoking profiles across Asian immigrant women and men. Thus, this study examines changing smoking rates from adolescence to adulthood among Asian Americans considering immigration status and gender, while controlling for a range of ethnicity, acculturation, SES, and psychosocial measures as control covariates.

Hypotheses

Hypothesis 1: Smoking rates among Asian American adolescent immigrants will be lower than those of native-born Asian Americans across the life course.

Hypothesis 2: The changes in smoking rates will differ across Asian American immigrant men and women. The slope of the age trajectory of smoking rates will be more strongly positive for Asian American immigrant men than Asian American immigrant women. The slope of the age trajectory of smoking rates will be approximately equal for Asian American immigrant and native-born men.

Hypothesis 3: Psychosocial resources such as depressive symptoms and acculturative measures (e.g., language use and coethnic peer network) will partially explain the smoking gaps between Asian American men and women across the life course.

Methods

Data

We examine data from waves 1 to 4 of Add Health (Harris et al. 2009). Add Health is a longitudinal panel study, and has added no new participants since the initial sample ascertainment in wave 1. A general description of the study and sample may be found in Ahmmad and Adkins (2021). The age range of the analysis sample spans 10 to 33 years. The Asian American nonmissing sample size in wave 1 was 1,584; some of the respondents did not participate in later waves of data collection, and no new participants were added. Therefore, the original sample size decreased to 1,089, 1,195, and 1,060 participants in waves 2, 3, and 4, respectively. The ethnic categories within the Asian American group included Chinese, Filipino, Japanese, Asian Indian, Korean, Vietnamese, and other Asian. Add Health allowed multiple ethnic, as well as multiple racial, identifications. For instance, Asian Americans who identified the Chinese could also identify as Filipino (categorized here as “multiethnic Asian”), as well as also identifying as Black (categorized here as “multiracial Asian”).

Survey Design

Survey weights are used to account for nonprobability sampling design, and Add Health requires the use of weights for its probability-based nonrandom, multistage sampling (Bollen et al. 2016; Park 2017). Add Health assigned survey weights for multilevel clustering, nonrandom sampling and for longitudinal, repeated observations. However, the use of survey weights has been inconsistent across disciplines, implying that the application of weights is guided by disciplinary convention rather than scientific validation (Bollen et al. 2016). Thus, survey methodologists have explicitly suggested that the use of survey weight should be determined by empirical tests, not on the basis of uncritical dogma, because unnecessary use of weights leads to diminished statistical power (Bollen et al. 2016). Thus, the universal use of weights for Asian American subsample may not be scientific, as found in a statistical test conducted by Park (2017). The weight informativeness test determined that the use of survey weights assigned for the general population to Asian American subsample would be unnecessarily penalizing (Park 2017). Moreover, using survey weights is complicated for the Asian American subsample because Add Health oversampled only Chinese Americans, whereas it also collected data from Filipinos, Vietnamese, Japanese, Asian Indians, along with others. Existing research in Asian American health has also used inconsistent sampling weights. For instance, Ryabov (2015) used only school-level weights to account for school-level clustering, leaving out the option for using person weights and region stratification weights (Ryabov 2015). Given these considerations, survey weights were not applied in this analysis.

Measures

The dependent variable, regular smoking, was derived from a dichotomous question that asks if the respondent smoked at least one cigarette daily for the past 30 days. This question was asked in all waves of the Add Health survey. Gender and nativity, the primary independent variables of interest, were measured by respondents’ self-report to a question asking the respondents sex and nation of origin.

The acculturation variable, use of English in personal conversation, is available in waves 1 to 3. There was relatively little within-participant variation in English use across waves; therefore, on the basis of the English use information from waves 1 to 3, we estimated English use for wave 4 as the average of preceding waves. The language use variable was dichotomized, with 1 indicating English use and 0 indicating no use of English, in personal conversation. The other acculturation variables include proportion Asians in friend network and proportion Asians in school, constructed using combined information from in-school friend nomination module and in-school survey data in wave 1. Participants nominated up to 10 close friends, which we matched with racial identification reports from in-school survey deriving coethnic friend proportion measure (number of known Asian friends divided by total number of known-race friends). Similarly, we divided the total number of students identifying as Asians by the total number of students to derive the proportion Asians in school.

We use an eight-item depressive symptom index (the Center for Epidemiologic Studies Depression Scale) that was consistently measured in all four waves. The questions regarding depressive symptoms ask how often it was true during the past week that the respondent (1) was bothered by the things that usually did not bother them; (2) could not shake off the blues, even with help from family and friends; (3) felt that they were just as good as others; (4) had trouble keeping attention; (5) was depressed; (6) was too tired to do things; (7) felt that people disliked them; and (8) felt sad. The items have a reliability score (.77) across waves and load highly for depressive symptoms (Adkins et al. 2009; Wight, Sepúlveda, and Aneshensel 2004).

The measures of education and household income include variables assessed across the life course. Education at wave 1 represents parents’ education level, whereas education levels at wave 3 and 4 were derived from respondents’ reports on the highest education level achieved so far. Education level is a dichotomous variable with 0 representing high school graduation and below (and 1 representing some college and above) among resident parents in wave 1; in wave 3, 0 indicates respondents’ education level as 12th grade and below (and 1 represents some college and above); in wave 4, 0 indicates respondents’ high school graduation and below (and 1 represents some college and above). Household income is a continuous measure. Like education, household income during wave 1 was based on parental income. Household income during waves 3 and 4 was based on respondents’ reports on income of current residents or household members who contributed to the household budget. Household smoking was measured as a binary variable on the basis of parental reports regarding their personal smoking or any resident smoking in the household.

The analyses also controlled for several other covariates. One of them is parental control, measured using seven items on resident parents’ approval (yes = 1, no = 0) of respondents’ choices of people to socialize with, clothing, amount of TV watching, TV programs, weekend time spending, weekday bedtime, and diet (see Table A1 in the Appendix for a detailed description of items). Cronbach’s α for the parental control scale was .64. Religiosity was assessed using a single item measuring the respondent’s perception of religion. The question assessed the importance of religious faith or religion, ranging from not important (1) to very important (4). All composite variables used in this analysis (e.g., parental control) were constructed as averages of the input items.

Statistical Analysis

Descriptive statistics of the sample (Table 1) present the total sample and subsamples stratified by nativity and gender and include significance testing of nativity-gender differences in all analysis variables using F tests. Specifically, linear and logistic mixed model F tests were used for time-variant continuous and dichotomous variables, respectively, and the F test of linear and logistic regression was used for time-invariant continuous and dichotomous variables, respectively.

Table 1.

Descriptive Statistics of the Sample: National Longitudinal Study of Adolescent to Adult Health (m = 30).

Total Sample (n = 6,332)					Immigrant Men (n = 1,352)	Immigrant Women (n = 1,392)	Native-Born Men (n = 1,948)	Native-Born Women (n = 1,640)	Subgroup Difference
Variable	Mean or %	SD	Minimum	Maximum	Mean or %	Mean or %	Mean or %	Mean or %	Subgroup Difference
Smoking (ever regular)	18%	.39	0	1	16%	10%	24%	21%	***
Female	48%	.50	0	1	—	—	—	—	—
Foreign born	43%	.50	0	1	—	—	—	—	—
Age	20.53	5.56	10.25	33.34	21.02	20.81	20.24	20.22	***
Chinese	17%	.38	0	1	17%	16%	19%	15%	ns
Filipino	35%	.48	0	1	49%	44%	25%	27%	***
Japanese	3%	.16	0	1	1%	0%	5%	3%	***
Asian Indian	2%	.14	0	1	2%	1%	2%	2%	ns
Korean	6%	.23	0	1	9%	10%	3%	3%	***
Vietnamese	5%	.22	0	1	6%	9%	3%	2%	***
Other Asian	11%	.32	0	1	10%	8%	12%	14%	ns
Multiethnic Asian	8%	.27	0	1	3%	3%	12%	11%	***
Multiracial Asian	14%	.35	0	1	3%	7%	19%	22%	***
Parents smoking in wave 1	32%	.47	0	1	32%	28%	32%	36%	ns
Education	71%	.45	0	1	71%	72%	69%	71%	ns
Household income	6.73	3.23	−3.56	17.13	6.64	6.42	6.95	6.82	**
English use in intimate settings	66%	.47	0	1	46%	45%	80%	84%	***
Number of Asian friends (maximum = 10)	.78	1.20	0	10	.96	.92	.68	.63	**
Proportion Asians in high school	.25	.19	−.2	.67	.23	.24	.27	.25	*
Religion important	2.89	.92	−.25	5.63	2.91	3.03	2.77	2.89	***
Parental control	.24	.24	−.47	1.01	.25	.26	.24	.21	ns
Depressive symptoms	.73	.39	−.55	3	.71	.77	.69	.77	***

Note: Some negative values are due to multiple imputations. m = multiple imputations.

p < .05. **p < .01. ***p < .005.

Inferential analyses of smoking apply generalized linear mixed models (observations = 6,332). A linear mixed model (i.e., identity link function) is used in the primary inferential analysis, and a random intercept and an age slope are specified. Cluster-robust standard errors (Huber-White sandwich) are used to account for violations of error distribution assumption. The full model is visualized in a simplified path diagram in Figure A1 in the Appendix (Guo and Hipp 2004). Margin plots are used to visualize and further examine differences between each of the four subgroups (i.e., native men, native women, immigrant men, and immigrant women), providing more granular examination of disparities in smoking across ages 10 to 33. Additionally, we conducted a sensitivity analysis specifying a logistic mixed model (i.e., logit link function) with a random intercept. Models also specifying a random slope in the logistic sensitivity analysis were also considered but failed to converge, which is likely indicative that the slope is more accurately modeled as fixed in the logistic model. Stata 16.1 was used for all analyses, and the commands mixed and melogit were used to specify the inferential models.

Missingness ubiquitously occurs in survey data, and it is more prevalent in panel data because of sample attrition over time. Ignoring the missing values may lead to biased parameter estimation and compromised statistical inference. The conventional listwise deletion method, which analyzes complete case analysis only, produces biased coefficients in the “missing at random” scenario, and generally reduces statistical power. To handle missing values in the data, we apply multiple imputation using chained equations for full Asian American subsample analysis (White, Royston, and Wood 2011). Multiple imputation using chained equations is a well-validated method, with the desirable property of maintaining the distributional characteristics of variables (i.e., logit imputation model for binary variables), using Markov-chain Monte Carlo iterative methods. All multiple imputation analyses included a conservative 30 imputations (m = 30).

Results

Descriptive Statistics

Table 1 presents descriptive statistics of the total sample and the subsamples stratified by nativity-gender combination. In the total sample, 18 percent of the respondent observations reported daily smoking, with differences across gender and nativity status. The lowest smoking rate is reported among immigrant women (10 percent), and the highest is among native-born men (24 percent). The percent of immigrant men smoking daily is 16 percent, whereas 21 percent of native-born women reported daily smoking. Approximately 43 percent of respondents were foreign born.

Concerning acculturation, about 45 percent of immigrant men and women reported using English in their intimate settings, whereas 80 percent or more of native-born Asian American women and men reported English use in intimate settings. Compared with their native-born counterparts, immigrant men and women indicated that a greater proportion of their friendship networks were friends of Asian descent. Immigrant women rated highest on the question of the importance of religion in their lives.

Education and household income levels show minor differences across gender and nativity. Immigrant and native-born women reported higher levels of depressive symptoms than their men counterparts. The divergent growth rates of smoking across gender and nativity are described in the multivariate analyses. Significant nativity-gender differences were observed for 18 of the 25 analysis variables considered, as shown in the rightmost column of Table 1.

Variation in Mean Levels, and Age Trajectory Slope, of Smoking across Gender and Nativity among Asian Americans

Model 1 in Table 2 indicates the average disparities in smoking rates across the four nativity-gender combinations. This shows that on average, across the observed time period, immigrant Asian American men and (b = −0.069, p < .001) and women (b = −0.113, p < .001) show significantly lower smoking rates than native-born Asian American men referent across the age range considered. Next, in model 2 in Table 2, we model the longitudinal trajectory of smoking behavior, which indicates that the majority of the disparities in smoking are due to differences in slopes, not differences in the intercept (i.e., the smoking rates observed at the baseline age), except in the case of immigrant Asian American men, for whom nominal intercept and slope differences are present but small and nonsignificant. In particular, immigrant Asian American women have much flatter smoking trajectory slopes (b = −0.010, p < .001), and native-born Asian American women also have significantly flatter trajectory slopes (b = −0.006, p < .05) than the native-born Asian American men referent.

Table 2.

Estimates from Linear Mixed-Model Growth Curve Predicting Smoking Behavior: National Longitudinal Study of Adolescent to Adult Health (n = 6,332, m = 30).

	Model 1		Model 2		Model 3		Model 4		Model 5
	Coefficient	t	Coefficient	t	Coefficient	t	Coefficient	t	Coefficient	t
Native men (reference)	—	—	—	—	—	—	—	—	—	—
Native women	−.030	−1.404	.033	1.138	.027	.942	.021	.749	.013	.464
Immigrant men	−.073***	−3.529	−.058	−1.945	−.050	−1.638	−.040	−1.307	−.034	−1.115
Immigrant women	−.138***	−7.389	−.035	−1.284	−.028	−1.015	−.015	−.529	−.013	−.473
Age			.014***	7.803	.014***	7.792	.014***	7.233	.014***	7.409
Native men × age (reference)			—	—	—	—	—	—	—	—
Native women × age			−.006*	−2.403	−.006*	−2.399	−.006*	−2.264	−.006*	−2.228
Immigrant men × age			−.002	−.842	−.002	−.851	−.003	−1.126	−.003	−1.038
Immigrant women × age			−.010***	−4.289	−.010***	−4.288	−.011***	−4.437	−.011***	−4.330
Chinese (reference)					—	—	—	—	—	—
Japanese					.092***	5.312	.083***	4.703	.080***	4.251
Filipino					.041	1.061	.029	.740	.048	1.184
Asian Indian					.001	.014	.006	.141	−.009	−.232
Korean					.112***	3.906	.105***	3.766	.098***	3.509
Vietnamese					.013	.545	.016	.648	.019	.760
Other Asian					.124***	4.841	.113***	4.350	.102***	3.928
Multiethnic Asian					.130***	4.246	.116***	3.879	.112***	3.568
Multiracial Asian					.130***	5.074	.116***	4.709	.101***	4.040
Parent smokes							.109***	6.064	.107***	6.024
Education							−.015	−1.089	−.018	−1.288
Household income							.009***	5.075	.008***	4.801
Religion important							−.006	−.853	−.005	−.829
English use									.034**	2.719
Proportion Asians in friend network									−.009	−1.254
Proportion Asians in school									−.060	−1.307
Depressive symptoms									.069***	5.262
Parental control									.014	.586
Intercept	.235***	16.453	.092***	4.633	.008	.388	−.044	−1.408	−.093**	−2.741
Random intercept SD (ln transformed)	−1.492***	−40.701	−4.256***	−83.351	−4.258***	−84.143	−4.246***	−83.158	−4.254***	−81.713
Random age SD (ln transformed)	—		−1.684***	−28.404	−1.719***	−27.958	−1.767***	−27.057	−1.793***	−26.593
Residual SD (ln transformed)	−1.176***	−59.828	−1.249***	−57.274	−1.249***	−57.493	−1.254***	−57.649	−1.255***	−57.903

p < .05. **p < .01. ***p < .001.

To allow more granular inference of gender and nativity differences by age, we calculated marginal effects, with 95 percent confidence intervals, for each nativity-gender combination (i.e., native men, native women, immigrant men, and immigrant women) across the full age range, which are visualized in Figure 1. This figure shows all two-way comparisons for the four nativity-gender groups. It illustrates the extremely low levels of smoking behavior throughout the time period for immigrant Asian American women, which are significantly lower than for all other nativity-gender combination, except at the youngest ages in the sample (i.e., age < 18 years). Immigrant Asian American men also show a lower risk for smoking compared with native-born Asian American men across the life course, but their age slope is roughly comparable with that of native-born men. Native-born women exhibit a pattern wherein they overlap native-born men at early ages, before significantly diverging around age 22, while initially showing elevated smoking rates relative to immigrant men at early ages, before converging around age 24. Finally, Figure 1F illustrates that immigrant men and immigrant women had quite similar smoking rates at baseline but diverge across adolescence and young adulthood, with immigrant men smoking significantly more than immigrant women after approximately age 18.

Figure 1.

Margins plot of predicted probabilities from linear mixed-effects models showing smoking growth trajectories across age, all two-way gender-nativity comparisons (on the basis of Table 2, model 2): (A) native men versus native women, (B) native men versus immigrant men, (C) native men versus immigrant women, (D) native women versus immigrant men, (E) native women versus immigrant women, and (F) immigrant men versus immigrant women.

The Roles of Ethnic Differences, SES, Acculturation, and Psychosocial Factors in Asian American Smoking Trajectories

Models 3 to 5 in Table 2 sequentially introduce three sets of covariates to assess their associations with smoking in this sample, as well as to examine the degree to which these factors explain the observed nativity-gender differences in smoking. In short, there is no evidence that the inclusion of these covariates results in attenuation of the observed nativity-gender differences in smoking. Rather, the significant nativity-gender differences described earlier for model 2 in Table 2 are virtually unchanged in terms of coefficient direction, significant, and magnitude in the remaining models 3 to 5.

Model 3 in Table 2 adds intra-Asian ethnic dummy indicators to the growth curve specified in model 2 to model conditional ethnic differences in smoking rates. Consistent with previous research (Ahmmad and Adkins 2021), model 3 shows elevated smoking rates among Japanese, Korean, and other Asian ethnicities, multiethnic Asians, and multiracial Asians (b ≈ 0.01 for all, p < .001), relative to the Chinese American referent. Model 4 in Table 2 examines the marginal effects of the SES indicators, religiosity, and parental smoking, adding these variables to the model 3 specification. Household income had an independent positive effect on smoking among Asian Americans (b = .009, p < .001), whereas education had a statistically nonsignificant effect. Parental smoking was significantly positively associated with respondent smoking (b = .109, p < .001), and religiosity was found to be nonsignificant. Finally, model 5 examines the influence of acculturation and psychosocial factors. Linguistic acculturation, indicated by English use in personal settings, had significant independent effects on smoking (b = .034, p < .01). Other acculturative measures, such as coethnic friend density in school, coethnic friend preference, as well as the parental control measure, did not show statistically significant associations with smoking. Depressive symptoms showed strong independent effects on smoking growth (b = .069, p < .001).

Sensitivity Analyses

To test the robustness of our substantive results, we conducted sensitivity analyses in which we changed the link function of the estimated mixed models from linear (i.e., identity link function) to logistic (i.e., logit link function) and reestimated the full five-model sequence. These sensitivity results are briefly discussed here and documented in Table A2, which gives all parameter estimates and inference in tabular form, and Figure A2, which provide a visualization of the marginal effects, with 95 percent confidence intervals, in the same format as Figure 1 in the main text. In short, the sensitivity analysis results are substantively extremely similar to the main analysis sequence, yielding virtually identical nativity-gender patterns in the main analysis (Figure 1) and the sensitivity analysis (Figure A2). To support research rigor and reproducibility, the analysis script is available through the Open Science Framework (https://osf.io/2cqrh/).

Discussion

These analyses add to the ongoing investigation on negative acculturation and the immigrant behavioral paradox in smoking among Asian Americans from adolescence to adulthood (ages 10–33 years). Results from linear mixed models show that Asian American adolescent immigrants smoke at increasing rates while transitioning from adolescence to adulthood. However, findings highlight the gendered pattern of smoking over time, extending empirical support to the selective acculturation hypothesis.

Overall, immigrant adolescents exhibit lower smoking rates than native-born Asian Americans across the life course. However, the pattern in the smoking growth across gender and nativity appears complex. For example, immigrant men show a comparable age trajectory slope as native-born men. On the contrary, the immigrant women’s advantage in smoking strengthens as they age through young adulthood. The findings may inform public health policy discouraging smoking among adolescents and young adults exhibiting divergent acculturation and integration patterns in the United States.

The mixed support for the negative acculturation hypothesis leads to stronger evidence for selective acculturation across Asian American women and men (Eitle, Wahl, and Aranda 2009). Despite having similar linguistic assimilation and exposure to household smoking in early life and known risk factors for smoking, immigrant men and women exhibit vastly different smoking growth (see Figure 1). The divergent pathways of losing immigrant protective behavior support the selective acculturation hypothesis, suggesting that Asian Americans’ assimilation can be segmented across various social statuses, including gender. Respondents with similar linguistic assimilation show different smoking trajectories across the life course, primarily because of complex social expectations across gender. Compared with immigrant adolescent boys, immigrant adolescent girls may face stricter social norms against smoking (Marsiglia et al. 2010).

Additionally, immigrant adolescent boys may receive social pressure to adopt host lifestyles and socialize in an expanded social network that includes peers from other racial and ethnic minorities and native-born generations. On the contrary, immigrant girls may choose to reduce social contacts and participate less in culturally stigmatized health damaging behavior such as smoking (Gorman et al. 2014). The gendered pattern of smoking across the life course indicates that Asian American women continue to exercise gender norms generally practiced in Asian countries. Evidence suggests that smoking is highly gendered in many Asian countries, with smoking viewed as socially acceptable behavior for men but highly stigmatized for women (Gorman et al. 2014).

This investigation included several variables representing acculturation, including English use, proportion Asian friends in friend network, and proportion Asians in schools. English use appears to have the strongest and consistent effect among acculturation variables, with more English language use in personal conversation associated with more smoking among Asian Americans. In the negative acculturation hypothesis, hypothesized variables to be protective of immigrant behavioral health include religiosity and parental control. Analyses show that the respondents’ perceptions of “religion to be important in life” are not statistically significant. The lack of a significant association may have resulted from changing religious views as adolescents assimilate to American society, rendering the protective effects of religion increasingly weak. Results were also nonsignificant for parental control. Further research should explicitly examine the changing social relationships and smoking behavior over time.

Demographic and psychosocial variables in this study include depressive symptoms, household smoking during childhood, childhood, and personal SES. Household smoking shows the strongest association with smoking, followed by depressive symptoms and English use. SES has the weakest association with smoking, which is supported by existing literature investigating the diminished effect of education on immigrants’ health (Assari 2020). Surprisingly, a very negligible positive effect is observed for household income, which might indicate that wealth gain is associated with greater purchasing power and access to smoking. Evidence shows that tobacco manufacturing companies target ethnic minorities, especially Asian and Pacific Island Americans, in commercial advertisements (Muggli et al. 2002). The study’s findings on SES are consistent with the positive association between income and adolescents smoking in Europe (Perelman et al. 2017). Other studies using Add Health data also show alcohol consumption risks with higher household income in Asian Americans (Ahmmad and Adkins 2021).

The minor gap in the smoking growth between native-born men and women across adolescence to adulthood supports gender equality or diffusion of innovation hypotheses. Although the rates in smoking growth remain lower among native-born women from adolescence to adulthood, the tremendous growth in smoking among native-born women indicates that the social control against women smoking weakens across immigrant generations. Further analysis across generations of Asian Americans and gender need to confirm results on the behavioral worsening across the social statuses.

As found in previous studies (Ahmmad and Adkins 2021), there exists ethnic heterogeneity in smoking prevalence. Chinese ethnicity shows the lowest smoking rates compared with other Asian American ethnic minorities. The highest smoking rate is observed among Asian Americans who share heritage with other racial groups (i.e., multiracial Asian Americans), followed by multiethnic Asian Americans. The ethnic disparity in smoking rates may result from the within-group concentration of acculturated individuals. For example, multiracial or multiethnic Asian Americans are mostly native born, indicating their risks for adopting risky health behavior. Nevertheless, the Korean ethnic group among non-mixed-heritage Asian Americans reported the highest smoking rates across the life course, which may not be driven by American acculturation but by the cultural influence in countries of origin (Kim, et al. 2005).

The present findings should be interpreted considering the following study limitations. Subjective reports on substance use and misuse, including smoking, are likely to be underreported in this survey. Especially among women, social stigma around smoking may also lead to underreporting of such behavior. Another limitation in this study, which is common across Asian American health studies, is the smaller sample size. A larger sample of Asian Americans may produce more accurate results by minimizing sampling error. Because of inconsistent measurement and a smaller sample size available in wave 5, this study did not include wave 5 of Add Health data. Another limitation of this study is the use of the English language as the marker of acculturation. The measures of language use may not represent complete acculturation to U.S. society, resulting in the weaker explanatory power of acculturation on smoking growth. A fuller explanation of the smoking gap among nativity and gender may emerge from a comprehensive analysis of integration and assimilation beyond language use. We included additional measure of acculturation such as coethnic friend preference; however, this was observed in adolescence only. Inclusion of a longitudinal measure capturing friend preference across time might be more informative.

Footnotes

Appendix

Table A2.

Estimates from Logistic Mixed-Model Growth Curve Predicting Smoking Behavior: National Longitudinal Study of Adolescent to Adult Health (n = 6,332, m = 30).

	Model 1		Model 2		Model 3		Model 4		Model 5
	Coefficient	t	Coefficient	t	Coefficient	t	Coefficient	t	Coefficient	t
Native men (reference)	—	—	—	—	—	—	—	—	—	—
Native women	−.323	−1.712	.172	.530	.111	.347	.038	.122	−.054	−.172
Immigrant men	−.695***	−3.441	−.925*	−2.404	−.853*	−2.199	−.704	−1.880	−.626	−1.673
Immigrant women	−1.620***	−7.212	−1.016*	−2.411	−.993*	−2.348	−.871*	−2.110	−.837*	−2.032
Age			.126***	7.701	.126***	7.694	.113***	6.729	.119***	7.088
Native men × age (reference)			—	—	—	—	—	—	—	—
Native women × age			−.046	−1.897	−.046	−1.902	−.041	−1.734	−.041	−1.708
Immigrant men × age			.006	.216	.005	.177	−.004	−.140	−.000	−.003
Immigrant women × age			−.064*	−2.132	−.064*	−2.121	−.067*	−2.272	−.063*	−2.163
Chinese (reference)					—	—	—	—	—	—
Japanese					1.150***	4.319	1.087***	4.001	1.021***	3.656
Filipino					.602	1.240	.528	1.055	.678	1.307
Asian Indian					−.002	−.003	.064	.090	−.184	−.262
Korean					1.755***	4.921	1.702***	4.779	1.616***	4.534
Vietnamese					.207	.456	.206	.450	.219	.479
Other Asian					1.529***	5.050	1.433***	4.568	1.294***	4.138
Multiethnic Asian					1.447***	4.313	1.313***	3.937	1.209***	3.518
Multiracial Asian					1.472***	4.851	1.342***	4.428	1.126***	3.689
Parent smokes							1.060***	5.869	1.038***	5.776
Education							−.210	−1.313	−.236	−1.444
Household income							.089***	4.341	.083***	4.078
Religion important							−.127	−1.785	−.124	−1.735
English use									.495**	2.877
Proportion Asians in friend network									−.111	−1.169
Proportion Asians in school									−.591	−1.213
Depressive symptoms									.783***	5.413
Parental control									.264	.906
Intercept	−1.949***	−13.706	−3.400***	−13.748	−4.420***	−13.897	−4.671***	−11.161	−5.364***	−11.570
Random intercept SD (ln transformed)	4.405***	8.777	5.203***	8.647	4.835***	8.604	4.549***	8.322	4.371***	8.202

p < .05. **p < .01. ***p < .001.

Acknowledgements

We thank the Consortium of Families and Health Research and the Department of Sociology at the University of Utah for providing access to the data. This research uses 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 to 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 Web site (). No direct support was received from grant P01-HD31921 for this analysis.

ORCID iDs

Zobayer Ahmmad

Kim Korinek

Author Biographies

Zobayer Ahmmad is an associate instructor in the Department of Sociology at the University of Utah. He received his PhD in 2021 and worked as a postdoctoral fellow in the Kaiser Permanente Division of Research. Using quantitative methods and applying sociological frameworks, he investigates health inequities in the United States along the intersecting lines of race/ethnicity and gender. His recent publications have explored health inequalities across diverse Asian ethnicity and gender from adolescence to adulthood. He has published more than 13 refereed articles in key sociological and specialty journals, including the Journal of Ethnic and Migration Studies, the Journal of Immigrant and Minority Health, and the Journal of Racial and Ethnic Health Disparities.

Kim Korinek is a professor of sociology and director of the Asia Center at the University of Utah. Her research examines the mutually transformative effects of social demographic changes, such as population aging and population mobility, and individual- and family-level experiences of receipt of support, living arrangements, socioeconomic mobility, health care utilization, and other outcomes related to well-being. Her work has been published in several top journals in in the disciplines of sociology and demography, including American Sociological Review, Demography, Social Science & Medicine, and International Migration Review.

Ming Wen is a professor of sociology and dean of the Faculty of Social Sciences at the University of Hong Kong. She is also a professor in the Department of Sociology at the University of Utah. Her research centers on social determinants of health and illness throughout the life course in the United States and China, focusing mainly on studying health disparities across sociodemographic groups and exploring family and neighborhood effects on human developmental outcomes. Her recent work has appeared in Social Science & Medicine, the American Journal of Public Health, and the Milbank Quarterly.

Daniel E. Adkins is an associate professor of sociology and psychiatry and chair of the Graduate Program in Statistics at the University of Utah. He is an interdisciplinary quantitative methodologist working primarily on topics in social epidemiology and psychiatric genetics. He completed his PhD at the University of North Carolina in 2009 with highest honors, including the Howard W. Odum Award for Excellence. He subsequently completed a postdoctoral fellowship in statistical genetics and worked as a National Institutes of Health–funded assistant professor at Virginia Commonwealth University prior to moving to Utah in 2016. His research focuses primarily on integrating quantitative methods from the social sciences and genomics to advance understanding of health. He has published more than 75 peer-reviewed articles in academic journals across disciplines, including Nature Neuroscience, Genome Biology, Nucleic Acids Research, Lancet Psychiatry, JAMA Psychiatry, Molecular Psychiatry, the American Journal of Psychiatry, Psychological Assessment, Social Forces, Sociological Theory, and Sociological Methods and Research.

References

Acevedo-Garcia

Dolores

Pan

Jocelyn

Jun

Hee-Jin

Osypuk

Theresa L.

Emmons

Karen M.

2005. “The Effect of Immigrant Generation on Smoking.” Social Science & Medicine 61(6):1223–42.

Adkins

Daniel E.

Wang

Victor

Dupre

Matthew E.

Van den Oord

Edwin J.C.G.

Elder

Glen H.

Jr.

2009. “Structure and Stress: Trajectories of Depressive Symptoms Across Adolescence and Young Adulthood.” Social Forces 88(1):31–60.

Ahmmad

Zobayer

Adkins

Daniel E.

2021. “Ethnicity and Acculturation: Asian American Substance Use from Early Adolescence to Mature Adulthood.” Journal of Ethnic and Migration Studies 47(19):4570–96.

Amos

Amanda

. 1996. “Women and Smoking.” British Medical Bulletin 52(1):74–89.

Assari

Shervin

. 2020. “Family Socioeconomic Status and Exposure to Childhood Trauma: Racial Differences.” Children (Basel, Switzerland) 7(6):57.

Barr

R. Graham

Avilés-Santa

Larissa

Davis

Sonia M.

Aldrich

Tom K.

Gonzalez

Franklyn

Henderson

Ashley G.

Kaplan

Robert C.

, et al. 2016. “Pulmonary Disease and Age at Immigration among Hispanics. Results from the Hispanic Community Health Study/Study of Latinos.” American Journal of Respiratory and Critical Care Medicine 193(4):386–95.

Bell

Andrew

Ward

Patrick

Tamal

Md Ehsanul Haque

Killilea

Mary

. 2019. “Assessing Recall Bias and Measurement Error in High-Frequency Social Data Collection for Human-Environment Research.” Population and Environment 40(3):325–45.

Bollen

Kenneth A.

Biemer

Paul P.

Karr

Alan F.

Tueller

Stephen

Berzofsky

Marcus E.

2016. “Are Survey Weights Needed? A Review of Diagnostic Tests in Regression Analysis.” Annual Review of Statistics and Its Application 3(1):375–92.

Bui

Hoan N.

2013. “Racial and Ethnic Differences in the Immigrant Paradox in Substance Use.” Journal of Immigrant and Minority Health 15(5):866–81.

10.

Chae

David H.

Gavin

Amelia R.

Takeuchi

David T.

2006. “Smoking prevalence among Asian Americans: findings from the National Latino and Asian American Study (NLAAS).” Public health reports 121(6):755–763.

11.

Choi

Sarah

Rankin

Sally

Stewart

Anita

Oka

Roberta

. 2008. “Effects of Acculturation on Smoking Behavior in Asian Americans: A Meta-Analysis.” Journal of Cardiovascular Nursing 23(1):67–73.

12.

Eitle

Tamela McNulty

Wahl

Ana-María González

Aranda

Elizabeth

. 2009. “Immigrant generation, selective acculturation, and alcohol use among Latina/o adolescents.” Social Science Research 38(3):732–742.

13.

Ferguson

Gail M.

Bornstein

Marc H.

2012. “Remote Acculturation: The ‘Americanization’ of Jamaican Islanders.” International Journal of Behavioral Development 36(3):167–77.

14.

Gfroerer

Joseph C.

Tan

Lucilla L.

2003. “Substance Use among Foreign-Born Youths in the United States: Does the Length of Residence Matter?” American Journal of Public Health 93(11):1892–95.

15.

Gong

Fang

Jun

Fujishiro

Kaori

Takeuchi

David T.

2011. “A Life Course Perspective on Migration and Mental Health among Asian Immigrants: The Role of Human Agency.” Social Science & Medicine 73(11):1618–26.

16.

Gorman

Bridget K.

Lariscy

Joseph T.

Kaushik

Charisma

. 2014. “Gender, Acculturation, and Smoking Behavior among U.S. Asian and Latino Immigrants.” Social Science & Medicine 106:110–8.

17.

Guo

Guang

Hipp

John

. 2004. “Longitudinal Analysis for Continuous Outcomes: Random Effects Models and Latent Trajectory Models.” Pp. 347–68 in Handbook on Data Analysis, edited by Hardy

Bryman

London: Sage Ltd.

18.

Haller

William

Portes

Alejandro

Lynch

Scott M.

2011. “Dreams Fulfilled, Dreams Shattered: Determinants of Segmented Assimilation in the Second Generation.” Social Forces 89(3):733–62.

19.

Harris

Kathleen Mullan

Halpern

Carolyn Tucker

Whitsel

Eric A.

Hussey

Jon M.

Tabor

Joyce

Entzel

Udry

J.R.

2009. “The National Longitudinal Study of Adolescent to Adult Health: Research Design.” https://addhealth.cpc.unc.edu//documentation/study-design

20.

Kandula

Namratha R.

Wen

Ming

Jacobs

Elizabeth A.

Lauderdale

Diane S.

2009. “Association between Neighborhood Context and Smoking Prevalence among Asian Americans.” American Journal of Public Health 99(5):885–92.

21.

Kim

Sun Seog

Son

Haein

Nam

Kyoung A.

2005. “The Sociocultural Context of Korean American Men’s Smoking Behavior.” Western Journal of Nursing Research 27(5):604–23.

22.

Grace X.

Tan

Yin

Toubbeh

Jamil I.

Xuefen

Shive

Steven E.

Lan

Yajia

. 2004. “Acculturation and Smoking Behavior in Asian-American Populations.” Health Education Research 19(6):615–25.

23.

Marks

Amy K.

Ejesi

Kida

García Coll

Cynthia

. 2014. “Understanding the U.S. Immigrant Paradox in Childhood and Adolescence.” Child Development Perspectives 8(2):59–64.

24.

Marsiglia

Flavio F.

Kulis

Stephen

Hussaini

Syed Khaleel

Nieri

Tanya A.

Becerra

David

. 2010. “Gender Differences in the Effect of Linguistic Acculturation on Substance Use among Mexican-Origin Youth in the Southwest United States.” Journal of Ethnicity in Substance Abuse 9(1):40–63.

25.

Marta

Tienda

Ron

Haskins

. 2011. “Immigrant Children: Introducing the Issue.” The Future of Children 21(1):3–18.

26.

Maxwell

Annette E.

Bernaards

Coen A.

McCarthy

William J.

2005. “Smoking Prevalence and Correlates among Chinese- and Filipino-American Adults: Findings from the 2001 California Health Interview Survey.” Preventive Medicine 41(2):693–99.

27.

Motti-Stefanidi

Frosso

. 2018. “Resilience among Immigrant Youth: The Role of Culture, Development and Acculturation.” Developmental Review 50:99–109.

28.

Muggli

M. E.

Pollay

R. W.

Lew

Joseph

A. M.

2002. “Targeting of Asian Americans and Pacific Islanders by the Tobacco Industry: Results from the Minnesota Tobacco Document Depository.” Tobacco Control 11(3):201–209.

29.

National Center for Chronic Disease Prevention and Health Promotion (US) Office on Smoking and Health. 2014. “`The Health Consequences of Smoking—50 Years of Progress: A Report of the Surgeon General.” Centers for Disease Control and Prevention (US). https://www.hhs.gov/sites/default/files/consequences-smoking-exec-summary.pdf

30.

Pampel

Fred C.

2001. “Cigarette Diffusion and Sex Differences in Smoking.” Journal of Health and Social Behavior 42(4):388–404.

31.

Pampel

Fred C.

2006. “Global Patterns and Determinants of Sex Differences in Smoking.” International Journal of Comparative Sociology 47(6):466–87.

32.

Pampel

Fred

Khlat

Myriam

Bricard

Damien

Legleye

Stéphane

. 2019. “Smoking among Immigrant Groups in the United States: Prevalence, Education Gradients, and Male-to-Female Ratios.” Nicotine & Tobacco Research 22(4):532–38.

33.

Park

So-Young

. 2017. “Depressive Symptoms and Suicidal Ideation from Adolescence to Young Adulthood in Chinese American and Filipino American Youth.” Journal of the Society for Social Work and Research 8(4):621–43.

34.

Perelman

Julian

Alves

Joana

Pfoertner

Timo-Kolja

Moor

Irene

Federico

Bruno

Kuipers

Mirte A. G.

Richter

Matthias

, et al. 2017. “The Association between Personal Income and Smoking among Adolescents: A Study in Six European Cities.” Addiction 112(12):2248–56.

35.

Raymond

Buriel

. 2011. “Historical Origins of the Immigrant Paradox for Mexican American Students: The Cultural Integration Hypothesis.” Pp. 37–60 in The Immigrant Paradox in Children and Adolescents: Is Becoming American a Developmental Risk? edited by García Coll

Marks

A. K.

Washington, DC: American Psychological Association.

36.

Annie

. 2014. “The Longer You Stay, the Worse Your Health? A Critical Review of the Negative Acculturation Theory among Asian Immigrants.” International Journal of Environmental Research and Public Health 11(8):8038–57.

37.

Roshania

Reshma

Narayan

K. M. Venkat

Oza-Frank

Reena

. 2008. “Age at Arrival and Risk of Obesity among US Immigrants.” Obesity 16(12):2669–75.

38.

Ryabov

Igor

. 2015. “Relation of Peer Effects and School Climate to Substance Use among Asian American Adolescents.” Journal of Adolescence 42(1):115–27.

39.

Salas-Wright

Christopher P.

Vaughn

Michael G.

Schwartz

Seth J.

Córdova

David

. 2016. “An ‘Immigrant Paradox’ for Adolescent Externalizing Behavior? Evidence from a National Sample.” Social Psychiatry and Psychiatric Epidemiology 51(1):27–37.

40.

Schooling

Mary

Gabriel

Edward

Sai Yin

Hedley

Anthony J.

Lam

Tai Hing

. 2004. “Childhood Migration and Cardiovascular Risk.” International Journal of Epidemiology 33(6):1219–26.

41.

Tang

Hao

Shimizu

Robin

Chen

Moon S.

2005. “English Language Proficiency and Smoking Prevalence among California’s Asian Americans.” Cancer 104(S12):2982–88.

42.

Umberson

Debra

Crosnoe

Robert

Reczek

Corinne

. 2010. “Social Relationships and Health Behavior Across the Life Course.” Annual Review of Sociology 36(1):139–57.

43.

Unger

Jennifer B.

Cruz

Tess Boley

Rohrbach

Louise Ann

Ribisl

Kurt M.

Baezconde-Garbanati

Lourdes

Chen

Xinguang

Trinidad

Dennis R.

, et al. 2000. “English Language Use as a Risk Factor for Smoking Initiation among Hispanic and Asian American Adolescents: Evidence for Mediation by Tobacco-Related Beliefs and Social Norms.” Health Psychology 19(5):403–10.

44.

Unger

Jennifer B.

Trinidad

Dennis R.

Weiss

Jie Wu

Rohrbach

Louise Ann

. 2004. “Acculturation as a Risk Factor for Smoking among Asian American Adolescents: Is the Association Confounded by Nationality?” Journal of Ethnicity in Substance Abuse 3(1):65–79.

45.

Walton

Emily

. 2012. “Resurgent Ethnicity among Asian Americans: Ethnic Neighborhood Context and Health.” Journal of Health and Social Behavior 53(3):378–94.

46.

Weiss

Jie Wu

Garbanati

James A.

2006. “Effects of Acculturation and Social Norms on Adolescent Smoking among Asian-American Subgroups.” Journal of Ethnicity in Substance Abuse 5(2):75–90.

47.

White

Ian R.

Royston

Patrick

Wood

Angela M.

2011. “Multiple Imputation Using Chained Equations: Issues and Guidance for Practice.” Statistics in Medicine 30(4):377–99.

48.

Wight

Richard G.

Sepúlveda

Joslan E.

Aneshensel

Carol S.

2004. “Depressive Symptoms: How Do Adolescents Compare with Adults?” Journal of Adolescent Health 34(4):314–23.

49.

Xie

Greenman

Emily

. 2005. “Segmented Assimilation Theory: A Reformulation and Empirical Test.” Ann Arbor: Population Studies Center, University of Michigan.

50.

Zhang

Jing

Wang

Zhongmiao

. 2008. “Factors Associated with Smoking in Asian American Adults: A Systematic Review.” Nicotine & Tobacco Research 10(5):791–801.

51.

Zhou

Min

Xiong

Yang Sao

. 2005. “The Multifaceted American Experiences of the Children of Asian Immigrants: Lessons for Segmented Assimilation.” Ethnic and Racial Studies 28(6):1119–52.

Changes in Smoking Prevalence from Adolescence to Adulthood among Asian Americans: Evidence of Selective Acculturation across Gender

Abstract

Keywords

Background

Acculturation and Smoking among Asian Americans

Selective Acculturation and Gendered Acculturation in Smoking Behavior

Gendered Acculturation in Smoking across the Life Course

Hypotheses

Methods

Data

Survey Design

Measures

Statistical Analysis

Results

Descriptive Statistics

Variation in Mean Levels, and Age Trajectory Slope, of Smoking across Gender and Nativity among Asian Americans

The Roles of Ethnic Differences, SES, Acculturation, and Psychosocial Factors in Asian American Smoking Trajectories

Sensitivity Analyses

Discussion

Footnotes

Appendix

Acknowledgements

ORCID iDs

Author Biographies

References