Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Jan 19.
Published in final edited form as: Emerg Adulthood. 2018 Apr 19;7(4):258–269. doi: 10.1177/2167696818763949

Smoking Trajectory Classes and Impact of Social Smoking Identity in Two Cohorts of U.S. Young Adults

Amanda L Johnson 1, Andrea C Villanti 1,2, Valerie Williams 3, Jessica M Rath 1,4, Donna M Vallone 1,4,5, David B Abrams 1,5,6, Donald Hedeker 7, Robin J Mermelstein 8
PMCID: PMC10798807  NIHMSID: NIHMS1913170  PMID: 38250305

Abstract

This study describes cigarette smoking trajectories, the influence of social smoker self-identification (SSID), and correlates of these trajectories in two cohorts of U.S. young adults: a sample from the Chicago metropolitan area (Social Emotional Contexts of Adolescent and Young Adult Smoking Patterns [SECAP], n = 893) and a national sample (Truth Initiative Young Adult Cohort Study [YA Cohort], n = 1,491). Using latent class growth analyses and growth mixture models, five smoking trajectories were identified in each sample: in SECAP: nonsmoking (n = 658, 73.7%), declining smoking (n = 20, 2.2%), moderate/stable smoking (n = 114, 12.8%), high/stable smoking (n = 79, 8.9%), and escalating smoking (n = 22, 2.5%); and in YA Cohort: nonsmoking (n = 1,215, 81.5%), slowly declining smoking (n = 52, 3.5%), rapidly declining smoking (n = 50, 3.4%), stable smoking (n = 139, 9%), and escalating smoking (n = 35, 2.4%). SSID was most prevalent in moderate/stable smoking (35.5% SECAP), rapidly declining smoking (25.2% YA Cohort), and nonsmoking. Understanding nuances of how smoking identity is formed and used to limit or facilitate smoking behavior in young adults will allow for more effective interventions to reduce tobacco use.

Keywords: social smoking, trajectory, tobacco, young adult, latent class growth model

Smoking remains the primary cause of cancer morbidity and mortality in the United States (U.S. Department of Health and Human Services, 2014). The young adult years are the period of greatest escalation and entrenchment of smoking (Richardson, Williams, Rath, Villanti, & Vallone, 2014), and significant initiation of cigarette and other combustible tobacco use still occur after age 18 (Foldes et al., 2010; Richardson et al., 2014; U.S. Department of Health and Human Services, 2012). Several smoking trajectory studies in young adults were conducted in community-based samples (D. W. Brook et al., 2008; J. S. Brook, Pahl, & Ning, 2006; J. S. Brook, Zhang, Burke, & Brook, 2014; Caldeira et al., 2012; Chassin, Presson, Pitts, & Sherman, 2000; Lee, Brook, Finch, & Brook, 2016; Lessov-Schlaggar, Kristjansson, Bucholz, Heath, & Madden, 2012; Orlando, Tucker, Ellickson, & Klein, 2004; Riggs, Chou, Li, & Pentz, 2007; Stanton, Flay, Colder, & Mehta, 2004; White, Pandina, & Chen, 2002) and national samples (Chen & Jacobson, 2012; Costello, Dierker, Jones, & Rose, 2008; Dutra, Glantz, Lisha, & Song, 2017; Fuemmeler et al., 2013; Jackson, Sher, & Schulenberg, 2008; Pollard, Tucker, Green, Kennedy, & Go, 2010). A number of studies of smoking trajectories from adolescence into young adulthood identify three common groups: “nonsmokers,” “early stable smokers,” and “late starters”; in the resultant models, early stable and late starter smokers achieved the highest levels of smoking in adulthood (D. W. Brook et al., 2008; J. S. Brook et al., 2014; Chassin et al., 2000; Costello et al., 2008; Dutra et al., 2017; Fuemmeler et al., 2013; Jackson et al., 2008; Lessov-Schlaggar et al., 2012; Orlando et al., 2004; Pollard et al., 2010; Riggs et al., 2007). Consistent with evidence on the increase in light or intermittent smoking among young adults (Pierce, White, & Messer, 2009) and smoking initiation after age 18 (Foldes et al., 2010), “occasional smokers” or “stable light smokers” were added as a distinct trajectory in several studies (D. W. Brook et al., 2008; Caldeira et al., 2012; Costello et al., 2008; Fuemmeler et al., 2013; Hair et al., 2017). In contrast to nonsmokers and heavy smokers who remain relatively stable over time, one study of smoking transitions in young adults showed that light and intermittent smokers were equally likely to become nonsmokers or heavy smokers at the end of a 2-year follow-up period (White, Bray, Fleming, & Catalano, 2009). Another study in young adult nondaily smokers showed relatively stable patterns of low, moderate, and high smoking behavior over 3.5 years follow-up (Klein, Bernat, Lenk, & Forster, 2013). Among young adult smokers, the phenomenon of “social smoking,” estimated at 51–80% in young adult (college and noncollege) smokers (Jiang, Lee, & Ling, 2014; Lisha, Delucchi, Ling, & Ramo, 2015; Moran, Wechsler, & Rigotti, 2004; Song & Ling, 2011; Villanti, Rath, & Vallone, 2012), likely accounts for a large proportion of “occasional” and “stable light smokers” in this age-group. The extent to which social smoking puts young people at risk for long-term stable patterns of smoking behavior in adulthood is unknown. Prevention efforts targeting adolescents may result in a smaller number of youth becoming “early stable” smokers. New types of interventions may be needed, however, to address the “late starter,” “occasional,” and “stable light” smokers to prevent the escalation or maintenance of smoking behavior into adulthood (Villanti, Niaura, Abrams, & Mermelstein, in press). Reducing social smoking may be a key target for intervention efforts in the latter groups.

Identity exploration is central to emerging adulthood (Arnett, 2000), and this includes exploring one’s identity in the context of self-perceptions regarding smoking behavior. Among adolescents, greater self-identification as a smoker predicted greater escalation of smoking behavior (Hertel & Mermelstein, 2012). This may be due to the perceptions of addiction (Hoek, Maubach, Stevenson, Gendall, & Edwards, 2013), risk (Brown, Carpenter, & Sutfin, 2011), and craving. Importantly, a number of studies highlight the large proportion of young adult smokers who self-identify as “social smokers” rather than “smokers” (Guillory, Lisha, Lee, & Ling, 2017; Song & Ling, 2011; Villanti et al., 2017). Identifying as a social smoker may allow young people to dissociate their smoking behavior from the known harms of smoking and thus facilitate their escalation of smoking. The identification as a “social smoker” and rejection of the “smoker” identity may also influence cessation behavior, with social smokers perceiving less of a need to quit (Falomir & Invernizzi, 1999; Høie, Moan, & Rise, 2010; Moan & Rise, 2005; Song & Ling, 2011). To date, no studies have examined the impact of self-identification as a social smoker on trajectories of cigarette use in young adults.

Social and contextual factors have also been shown to significantly influence smoking trajectories in adolescence and young adulthood (Dutra et al., 2017; Fuemmeler et al., 2013; Jessor, 1991; Oetting, 1999). An evidence review published in 2000 documented the consistent effect of social factors on the onset of adolescent smoking and increases in smoking frequency and intensity (Mayhew, Flay, & Mott, 2000), with more recent longitudinal studies highlighting the effect of parental transmission of smoking behavior (Mays et al., 2014; Selya, Dierker, Rose, Hedeker, & Mermelstein, 2012). The extent to which socializing influences impact trajectories of young adult tobacco use has received less attention, though recent studies have identified committed relationships (Huh, Huang, Liao, Pentz, & Chou, 2013), peer approval and exposure to a social smoking environment (Colder, Flay, Segawa, & Hedeker, 2008), and other social and environmental predictors (Fuemmeler et al., 2013; Klein et al., 2013) of smoking frequency over time in young adults. In adolescent males, social motives are associated with smoker identity development but not in females (Hertel & Mermelstein, 2016). Currently, it is unknown whether these social and contextual factors are unique drivers of smoking behavior or whether they shape self-identification as a social smoker in young adults, which in turn, impacts smoking escalation or reduction over time.

Most of the existing trajectory studies include birth cohorts ranging from the 1950s through the early 1980s, representing groups of young people who came of age prior to the Master Settlement Agreement and increases in state-level tobacco control efforts, which have influenced smoking norms and likely affect the development of smoking identity in today’s young adults. Additionally, the increased onset of smoking in young adulthood, as opposed to adolescence (Thompson, Mowery, Tebes, & McKee, 2017), supports the utility of examining smoking trajectories in emerging adults. As noted previously, smoking identity was associated with smoking escalation in adolescents in one cohort (Hertel & Mermelstein, 2012). The purpose of the current study is to describe the influence of self-identification as a social smoker on smoking trajectories of young adults in the same cohort and examine whether findings were consistent in a national cohort of U.S. young adults. The replication of smoking rate trajectories in two contemporary cohorts of young adults—a high-risk sample and a national sample—would provide strong evidence for categories of young adult smokers that could be targeted for interventions. We also sought to examine whether social and contextual influences on smoking, specifically family, peer, and work/school influences, were correlated with smoking trajectories after controlling for self-identification as a social smoker.

Method

The two cohorts included in this study are (1) an urban, community-based sample in Chicago, IL (Social Emotional Contexts of Adolescent and Young Adult Smoking Patterns [SECAP]), and (2) a large, national sample of young adults (Truth Initiative Young Adult Cohort Study [YA Cohort]). Details on both samples and measures employed are described below using the Strengthening the Reporting of Observational Studies in Epidemiology STROBE reporting guidelines.

Participants

SECAP.

SECAP at the University of Illinois at Chicago is a longitudinal, multimethod study of the natural history of smoking. This project successfully recruited a diverse cohort at high risk for cigarette smoking (N = 1,263 at baseline, 2005–2006) in the greater Chicago metropolitan area when they averaged 15 years of age and has retained almost 85% (n = 1,066) through the 7-year follow-up (March 2013–August 2014; mean age of 23). The current analyses focus on four annual waves of data (Years 4–7) from a subset of 893 young adults who had ever smoked a cigarette and provided complete data on the self-identified smoking status measure at the 4-year follow-up. All procedures received approval from the institutional review board at the University of Illinois at Chicago.

Truth initiative young adult cohort study.

The current study also uses data from seven consecutive, biannual waves of the Truth Initiative Young Adult Cohort Study (July 2011–October 2014). The detailed methods of this study have been described elsewhere (Rath, Villanti, Abrams, & Vallone, 2012). The sample is comprised of a nationally representative sample of young adults ages 18–34 drawn from GfK’s KnowledgePanel® which is recruited via address-based sampling to provide a statistically valid representation of the U.S. population, including cell phone-only households. The validity of this methodology has been reported previously (Chang & Krosnick, 2009; Fowler, Gerstein, & Barry, 2013; Grande, Mitra, Shah, Wan, & Asch, 2013; Kumar, Quinn, Kim, Daniel, & Freimuth, 2012; Rhodes, Breitkopf, Ziegenfuss, Jenkins, & Vachon, 2015; Yeager et al., 2011). The sample is refreshed at each wave to retain the initial sample size.

The panel recruitment rate ranged from 13.9% to 14.9% across the seven waves. In 64.4–65.7% of the identified households, one member completed a core profile survey of key demographic information (profile rate). At each wave, one panel member per household was selected at random to be part of the study sample, and no members outside the panel were recruited. The completion rate ranged from 46.2% to 68.4% across waves. The cumulative response rate (a product of these three rates) ranged from 4.4% to 6.6%. This study was approved by the Independent Investigational Review Board, Inc., for Waves 1–3 and Chesapeake Institutional Review Board, Inc., for Waves 4–7. Online consent was collected from participants before survey self-administration. The present analysis focuses on a subset of participants 18–24 at study entry (n = 1,491) who entered at any of the seven waves of data collection, reported ever smoking a cigarette at study entry, had at least one smoking rate, and provided complete data on the self-identified social smoking status measure.

Outcome

Cigarette smoking rate.

Cigarette smoking rate was derived from two questions: “During the last 30 days, on how many days did you smoke or try cigarettes (even one puff)?” and “On the days you smoked cigarettes, about how many cigarettes did you smoke each day?”

Analyses used a continuous smoking rate derived from the number of days reported smoking a cigarette in the last 30 days multiplied by the average number of cigarettes smoked per day, divided by 30 days:

(# of days smoked in past 30 days )× (average number of cigarettes per day) (30 days ).

Ever smokers who did not report cigarette use in the past 30 days were categorized as having a smoking rate of 0 cigarettes/day.

Self-Identification as a Social Smoker

Since the categorical measure of self-identified smoking status was associated with smoking escalation among adolescents (Hertel & Mermelstein, 2012) and largely drove the formation of a class of social smokers in our earlier work in young adults (Villanti et al., 2017), a variant of this measure was used to define self-identification as a social smoker. In both cohorts, participants responded to the following item: “Which of the following best describes how you think of yourself?” with response choices of “smoker,” social smoker, occasional smoker, “ex-smoker,” “someone who tried smoking,” and “nonsmoker.” Self-identified social smoker and occasional smoker were combined to social smoker self-identification (yes/no) in line with Hertel and Mermelstein’s work and included as a covariate in the trajectory models.

Other Measures

Participants in SECAP provided information on age, gender, race/ethnicity, education, and employment status at baseline and all other covariates at the 5-year follow-up. In the YA Cohort, sociodemographic data were collected at study entry as part of the KnowledgePanel routine data collection. For the analysis, race/ethnicity was dichotomized due to sample size (White/non-White), and education was dichotomized to at least some college education (yes/no). Employment status was defined as full time, part time, or not at all.

Social and contextual correlates.

Participants also provided information on family, peer, and work/school influences on smoking, including current living situation (at home, campus housing, own apartment, condo, and house). In SECAP, smoking environment in the home was assessed by asking participants if smoking is allowed at home (yes/no), whether they live with a smoker (yes/no), and whether their partner smokes (yes/no). In the YA Cohort, participants were asked about living with a partner, roommate, or family member who smokes (yes/no), and parental smoking during childhood (one or both parents, neither parent).

In the YA Cohort, participants who reported working at least part time were asked about smoking rules at their work (not allowed at all, allowed in some places, and no rules or laws) and smoking norms at work (no one smokes at work, at least a few people smoke at work, and I work by myself most of the time).

Peer influences on smoking were assessed by 3 items in both cohorts. Participants were asked “Out of every 100 people your age, how many do you think smoke cigarettes at least once a week?” with response options in groups of 10 ranging from “10 or less” to “91–100.” The second item asked “How many of your four closest friends usually smoke at least one cigarette a week? (0–4)” and the third item assessed whether participants’ spouse/life partner or boyfriend/girlfriend smokes cigarettes.

Data Analysis

The current analysis is based on n = 893 participants in SECAP and n = 1,491 participants in the YA Cohort. The analysis focused on 4 years of data collection in SECAP and seven waves (spanning 4 years) in the YA Cohort. Latent class growth analyses (LCGA) and growth mixture models (GMMs) were conducted in 2015 using Mplus, Version 7.4 (http://www.statmodel.com; B. O. Muthén, 2002; Muthen & Muthen, 2002). LCGA fixes the variance and covariance estimates for the intercept and slope (growth factors) within each class to 0, and thus all individual growth trajectories within a class are homogenous. From there, we conducted GMMs with one to six classes to estimate the unique variances of the growth factors within each class (Jung & Wickrama, 2008). The variance and covariance of the quadratic and cubic terms were fixed to 0. Due to the sparseness of the data at some ages, we modeled trajectories as a function of wave rather than of participant age to avoid nonconvergence of the age estimation. Maximum likelihood model estimation was used to account for missing data, assuming missing at random (Little & Rubin, 1987, 1989; L. K. Muthén & Muthén, 1998). Both the LCGA and GMM controlled for self-identification as a social smoker to derive the latent trajectory classes. In the GMM, the means of growth factors after controlling for the impact of social smoker self-identification could vary across classes. The class-specific mean coefficients of the intercept, slope, quadratic, and cubic terms indicate class-specific trajectory shapes.

Model fit.

The optimal number of trajectory classes was determined by running models with a successive number of classes from one to six trajectory classes. Model fit was evaluated using five criteria: (1) the sample-size-adjusted Bayesian information criteria (BIC); (2) the Lo–Mendel–Rubin likelihood ratio and the Vuong–Lo–Mendell–Rubin likelihood ratio tests (Collins & Lanza, 2010); (3) odds of correct classification (D. Nagin, 2005); (4) sample size in the trajectory classes; and (5) interpretability. Given findings from a review and simulation study on fit statistics that highlighted that “in the majority of cases, the LMR and adjusted LMR tests preferred a model with 2 fewer classes than were actually present” (Tein, Coxe, & Cham, 2013, pp. 654–655), the optimal model was selected with the number of classes that minimized the BIC (B. O. Muthén, 2001; D. S. Nagin, 1999; Nylund, Asparouhov, & Muthén, 2007), retained at least 2% of participants in each class, and was interpretable.

Once the best fitting model was identified, trajectory class probabilities and trajectory assignment were transferred to Stata, Version 14.2 (http://www.stata.com). Time fixed social and contextual influences were used to characterize trajectories of cigarette smoking. Pearson χ2 tests and Fisher exact tests were used to identify significant differences in categorical characteristics across trajectory classes (significance at p < .05). Two-way analysis of variance tests were used to identify significant differences in the level of continuous characteristics across trajectory classes (significance at p < .05).

Results

Model Fit

The model fit indices for determining the optimal number of trajectory classes are presented in Table 1. Based on the sample size in each trajectory class, BIC, and interpretation of the trajectory classes, the five-trajectory model provided the most stable patterns and best classification in both data sets.

Table 1.

Model Fit Indices for Growth Mixture Models in SECAP and YA Cohort.a

SECAP YA Cohort
Class Model Class 1 Class 2 Class 3 Class 4 Class 5 Class 6 Class 1 Class 2 Class 3 Class 4 Class 5 Class 6
Two-class model n (%) 122 (14.1%) 771 (85.9%) 1,291 (86.6%) 200 (13.4%)
OCC 32.33 110.11 99.00 18.23
ssBIC 15,183.57 18,316.15
VLMR LRT 0.00 0.01
LMR LRT 0.00 0.01
Three-class model n (%) 105 (11.8%) 33 (3.7%) 755 (84.5%) 89 (6.0%) 1,242 (83.3%) 160 (10.7%)
OCC 13.29 75.92 110.11 8.26 141.86 10.49
ssBIC 14,812.46 17,799.32
VLMR LRT 0.01 0.51
LMR LRT 0.01 0.51
Four-class model n (%) 35 (3.9%) 750 (84.0%) 22 (2.5%) 86 (9.6%) 1,234 (82.8%) 152 (10.2%) 44 (3.0%) 61 (4.1%)
OCC 22.81 82.33 10.49 13.29 61.50 6.87 7.40 9.87
ssBIC 14,563.08 17,326.85
VLMR LRT 0.63 0.43
LMR LRT 0.64 0.43
Five-class model n (%) 658 (73.7%) 22 (2.5%) 79 (8.8%) 114 (12.8%) 20 (2.2%) 1,215 (81.5%) 52 (3.5%) 50 (3.4%) 139 (9.3%) 35 (2.3%)
OCC 57.82 110.11 12.51 16.24 37.46 70.43 4.29 9.31 9.42 7.77
ssBIC 14,253.21 16,890.59
VLMR LRT 0.17 0.24
LMR LRT 0.17 0.24
Six-class model n (%) 686 (76.8%) 97 (10.9%) 37 (4.1%) 23 (2.6%) 14 (1.6%) 36 (4.0%) 34 (2.3%) 55 (3.7%) 30 (2.0%) 1,192 (79.9%) 91 (6.1%) 89 (6.0%)
OCC 89.91 10.24 24.64 82.33 75.92 15.39 19.83 10.36 11.35 32.33 8.01 7.20
ssBIC 14,076.36 16,474.88
VLMR LRT 0.61 0.28
LMR LRT 0.61 0.29
Open in a new tab

Note. OCC = odds of correct classification; ssBIC = sample-size-adjusted Bayesian information criteria; VLMR = Vuong–Lo–Mendell–Rubin; LRT = likelihood ratio test; LMR = Lo–Mendell–Rubin; SECAP = Social Emotional Contexts of Adolescent and Young Adult Smoking Patterns.

a

Apart from ssBIC, model fit indices were not available for one-class growth mixture models. For SECAP, ssBIC of one-class model was 15,772.69. For YA Cohort, ssBIC of one-class model was 19,474.70.

Overall Sample Characteristics

In both samples, the majority of participants were over age 21, female, non-Hispanic White, college educated, currently employed, and reported past 30-day use of alcohol at study entry (data available upon request).

Smoking Trajectories

The mean smoking rate at each survey wave for each trajectory class is presented in Figure 1 for the YA Cohort and Figure 2 for SECAP. Five distinctive smoking trajectories emerged from the analysis of each sampl: in SECAP, a nonsmoking trajectory (n = 658, 73.7%), declining smoking trajectory (n = 20, 2.2%), moderate/stable smoking trajectory (n = 114, 12.8%), high/stable smoking trajectory (n = 79, 8.9%), and an escalating smoking trajectory (n = 22, 2.5%); and in the YA Cohort, a nonsmoking trajectory (n = 1,215, 81.5%), slowly declining smoking trajectory (n = 52, 3.5%), rapidly declining smoking trajectory (n = 50, 3.4%), stable smoking trajectory (n = 139, 9%), and an escalating smoking trajectory (n = 35, 2.4%). Importantly, cigarette use occurred at some level in all groups at all time points, including the nonsmoking classes. Even in the declining trajectory classes, cessation was not attained.

Figure 1.

Figure 1.

Open in a new tab

Smoking trajectory classes in Social Emotional Contexts of Adolescent and Young Adult Smoking Patterns (SECAP) data set (n = 893). All SECAP participants reported smoking rates for all four waves of data collection.

Figure 2.

Figure 2.

Open in a new tab

Smoking trajectory classes in the Truth Initiative Young Adult Cohort Study (n = 1,491). YA Cohort participants varied in number of waves of data collection. Mean (standard deviation) number of waves by trajectory class: nonsmoking: 3.5 (0.1) waves; slowly declining: 2.8 (0.3) waves; rapidly declining: 3.9 (0.4) waves; stable: 3.0 (0.2) waves; and escalating: 3.8 (0.5) waves.

The five-class conditional GMMs are summarized in Table 2. In both samples, estimated proportions of respondents identifying as social smokers at study entry significantly differed by trajectory. In SECAP, the estimated intercept parameter of the trajectory significantly differed by social smoker self-identification, with those who identified as social smokers at study entry reporting significantly smaller initial smoking rates compared to those who did not identify as social smokers.

Table 2.

Conditional growth mixture model output including social smoker self-identification (SSID) as a covariate in the 5-class model.

SECAP Overall Nonsmoking Declining Moderate/Stable High/Stable Escalating
Mean percentage of SSID (%)a,b 34.6 15.9 35.5 10.1 21.0
Mean intercept parameter of smoking rate (I)a,b 0.66 4.23 2.76 5.99 7.52
Mean slope parameter of smoking rate (S)a,b 0.18 1.09 1.41 4.68 10.73
Mean quadratic parameter of smoking rate (Q)a,b −0.22 5.11 −0.72 −2.77 −8.60
Mean cubic parameter of smoking rate (C)a,b 0.04 −1.94 0.15 0.52 1.97
I on SSIDc −0.43
S on SSID 0.18
Q on SSID 0.01
YA Cohort Overall Nonsmoking Slowly Declining Rapidly Declining Stable Escalating
Mean percentage of SSID (%)a,b 24.9 10.4 25.2 15.2 21.8
Mean intercept parameter of smoking rate (I)a,b 0.34 16.67 18.72 8.21 1.02
Mean slope parameter of smoking rate (S)a,b 0.09 −2.62 −14.46 −1.34 16.52
Mean quadratic parameter of smoking rate (Q)a,b −0.02 0.78 3.91 0.66 −5.80
Mean cubic parameter of smoking rate (C)a,b 0.00 −0.13 −0.32 −0.07 0.56
I on SSID −0.05
S on SSID 0.17
Q on SSID −0.02
Open in a new tab

Note. Smoking rate = number of days reported smoking a cigarette in the last 30 days multiplied by the average number of cigarettes smoked per day, divided by 30 days. SECAP = Social Emotional Contexts of Adolescent and Young Adult Smoking Patterns.

a

Estimated means for latent variables.

b

Boldface values indicate significant two-tailed p value (<.05) for estimated means for the latent variables.

c

Significant p value (<.05).

In SECAP, the moderate/stable smoking trajectory class reported the largest proportion of participants identifying as social smokers at study entry (35.5%), whereas the rapidly declining smoking trajectory class in the YA Cohort reported the largest proportion of social smokers (25.2%; Table 2). In both samples, the next largest proportions of social smokers were present in the nonsmoking class followed by the escalating smoking class.

Social and Contextual Correlates of Trajectory Class

In both data sets, social and contextual influences on smoking remained correlated with trajectory class, even after controlling for self-identification as a social smoker. Living with a smoker, having a greater number of friends who smoke and having a partner who smokes were significantly associated with smoking trajectory class in both cohorts (Table 3). Respondents in both nonsmoking classes reported the smallest proportions of living with a smoker, the smallest perceived peer smoking prevalence estimates, and the fewest number of close friends who smoke. Respondents in the nonsmoking class in the YA Cohort reported the lowest prevalence of partner smoking across the five classes (19.3%).

Table 3.

Social Influence Correlates of Smoking Trajectory Class in Young Adult Ever Smokers in SECAP and YA Cohort.

SECAP YA Cohort
Class Model Nonsmoking Declining Moderate/Stable High/Stable Escalating Nonsmoking Slowly Declining Rapidly Declining Stable Escalating
Total n = 658 n = 20 n = 114 n = 79 n = 22 p a Total n = 1,215 n = 52 n = 50 n = 139 n = 35 p a
N = 893 73.7% 2.2% 12.8% 8.9% 2.5% N = 1,491 81.5% 3.5% 3.4% 9.3% 2.4%
Living situation
 At homeb 59.7% 56.8% 60.0% 65.4% 71.8% 70.0% 41.1% 42.6% 30.3% 24.3% 38.0% 38.5% *
 Campus housing 5.9% 7.0% 6.7% 3.9% 0.0% 5.0% 4.8% 5.3% 3.0% 8.1% 1.9% 0.0%
 Own apartment, condo, and house 34.4% 36.2% 33.3% 30.8% 28.2% 25.0% 54.0% 52.1% 66.7% 67.6% 60.2% 61.5%
Live with smoker
 No 58.9% 64.2% 60.0% 44.3% 42.3% 40.0% *** 81.4% 85.3% 59.6% 66.0% 63.3% 74.3% ***
 Yes 41.1% 35.8% 40.0% 55.7% 57.8% 60.0% 18.6% 14.7% 40.4% 34.0% 36.7% 25.7%
Smoking allowed at home
 No 79.3% 82.4% 60.0% 68.9% 71.8% 85.0% **
 Yes 20.7% 17.6% 40.0% 31.1% 28.2% 15.0%
Parental smoking status during childhood
 Yes, one 32.3% 30.7% 42.3% 38.0% 39.6% 37.1% ***
 Yes, both 22.2% 17.7% 34.6% 44.0% 43.9% 40.0%
 No, neither 45.5% 51.6% 23.1% 18.0% 16.6% 22.9%
Smoking rules at work
 Not allowed at allb 20.0% 22.0% 8.3% 13.6% 6.9% 10.0% **
 Allowed in at least some areas 74.9% 72.2% 91.7% 81.8% 91.4% 90.0%
 No rules or laws 5.1% 5.8% 0.0% 4.6% 1.7% 0.0%
Smoking culture at work
 No one smokes at workb 21.5% 23.5% 5.0% 15.0% 11.1% 10.0%
 At least a few people smoke at work 73.4% 71.8% 85.0% 80.0% 81.5% 85.0%
 I work by myself most of the time 5.1% 4.7% 10.0% 5.0% 7.4% 5.0%
Perceived smoking prevalence among peers
 Mean level (SE)C 5.5 (0.1) 5.3 (0.1) 6.3 (0.4) 5.9 (0.2) 6.2 (0.3) 5.8 (0.4) ** 5.0 (0.1) 4.8 (0.1) 6.2 (0.4) 5.5 (0.4) 6.4 (0.3) 5.2 (0.6) ***
Number of closest friends who smoke
 Mean (SE)C 2.5 (0.1) 2.1 (0.1) 2.9 (0.4) 3.3 (0.1) 3.8 (0.1) 3.7 (0.2) *** 1.9 (0.0) 1.6 (0.0) 3.2 (0.1) 3.0 (0.2) 3.2 (0.1) 2.8 (0.2) ***
Partner smokes
 Don’t have one 34.6% 35.1% 40.0% 27.4% 35.2% 50.0% *** - — — — — —
 No 40.8% 44.7% 40.0% 37.7% 21.1% 15.0% 75.1% 80.7% 40.4% 52.0% 47.5% 74.3% ***
 Yes 24.6% 20.2% 20.0% 34.9% 43.7% 35.0% 24.9% 19.3% 59.6% 48.0% 52.5% 25.7%
Open in a new tab

Note. SECAP = Social Emotional Contexts of Adolescent and Young Adult Smoking Patterns; SE standard error.

a

Unless otherwise noted, p values are from Pearson χ2 tests.

b

p Values are from Fisher exact tests for small cell sizes (n < 5).

c

p Values are from analyses of variance.

*

p < .05.

**

p < .01.

***

p < .001.

In both cohorts, the highest perceived prevalence estimates of smoking among peers, the number of closest friends who smoke, and prevalence of partner smoking were seen in the declining smoking classes and the stable smoking classes. In SECAP, respondents in the high/stable class reported significantly more close friends who smoke compared to those in the moderate/stable class (p = .02). Respondents in the escalating smoking class in SECAP reported the second highest mean number of close friends who smoke (3.7 [0.2]) behind the high/stable smoking class (3.8 [0.1]). In the YA Cohort, respondents in the escalating smoking class reported the second lowest number of friends who smoke (2.8 friends; SE = 0.2).

In the YA Cohort, those in the stable class reported the highest perceived mean prevalence of peer smoking (6.4 people [SE = 0.3]). There was no difference in mean prevalence of peer smoking among respondents in the slowly declining smoking class (6.2 [0.4]) compared to those in the rapidly declining smoking class (5.5 [0.4]). Those in the slowly declining smoking class and the stable smoking class reported the greatest number of close friends who smoke in the YA Cohort (3.2 friends [SE = 0.1]). There was no significant difference in number of close friends who smoke between those in the slowly declining and rapidly declining (3.0 [0.2]) classes. Respondents in the slowly declining class reported the greatest prevalence of partner smoking across all classes in the YA Cohort.

Discussion

Using two large contemporary cohorts of U.S. young adults (aged 18–24), the present study found evidence of five trajectories of smoking behavior when self-identification as a social smoker was included in the model. These trajectories were defined by nonsmoking, stable smoking (two classes in SECAP, one class in YA Cohort), declining smoking (one class in SECAP, two classes in YA Cohort), and escalating smoking behavior. Importantly, none of the trajectories in either cohort achieved smoking cessation during young adulthood. Characteristics of trajectories differed by sample. In each cohort, the majority of the sample reported consistently low smoking rates over the time period. The moderate/stable and nonsmoking trajectory classes had the greatest proportions of young adults identifying as social smokers in SECAP, whereas the rapidly declining and nonsmoking classes tied for the greatest proportion of self-identified social smokers in the YA Cohort. Consistency of findings across the two samples of the high proportions of social smokers in the nonsmoking classes suggests that there is a subset of social smokers who smoke infrequently and for whom identifying as a social smoker indicates some acknowledgment of their smoking behavior. Interestingly, the moderate/stable (SECAP) and rapidly declining (YA Cohort) classes achieve a similar level of moderate stable smoking at the end of follow-up in line with Klein, Bernat, Lenk, and Forster (2013); as in the nonsmoking classes, identifying as a social smoker in these groups may actually serve a protective function in maintaining a low level of smoking. In both samples, however, approximately 20% of the escalating class self-identified as social smokers. In these classes, identifying as a social smoker may be harmful in normalizing an escalating pattern of smoking behavior.

In both data sets, social and contextual influences on smoking remained correlated with trajectory class after controlling for self-identification as a social smoker. Living with a smoker, having a greater number of friends who smoke, and having a partner who smokes were associated with smoking behavior in both cohorts, consistent with earlier trajectory studies in adolescents (Abroms, Simons-Morton, Haynie, & Chen, 2005; Bernat, Erickson, Widome, Perry, & Forster, 2008; Costello et al., 2008). Interestingly, the escalating class looked more similar to nonsmokers than other smoker classes with respect to smoking rules at home (SECAP) and living with a smoker or having a partner who smokes (YA Cohort). This was a small class in both data sets but highlights a group of late starters for whom adolescent smoking prevention interventions may have been effective, but who may also require a novel smoking prevention intervention in young adulthood.

The limitations of this study include the lack of information regarding smoking behaviors at each age of young adulthood (18–24 years). Due to the sparseness of the data and the 6-month intervals between survey waves, we were unable to successfully run age-based GMM models. Future research should explore age-based trajectories of cigarette smoking from early adolescence through adulthood. Second, the restricted sample of young adults aged 18–24 and the period of data collection truncates our trajectories earlier than in other studies; a class that achieves cessation may be present in later years, though not captured in our follow-up window. Third, we dichotomized race/ethnicity as White/non-White due to small cell sizes, and this limits our ability to draw inferences about differences in trajectories by key subgroups. Lastly, this study relied on survey collection of self-reported data, which is subject to bias.

This study is unique in focusing on smoking trajectories within young adulthood, not from adolescence to adulthood. As in earlier cross-sectional work (Villanti et al., 2017), it highlights self-identified social smoker status as an important predictor of smoking behavior during this developmental period, but social and contextual influences on smoking remain strongly correlated with trajectories of cigarette use. Producing smoking abstinence in the nonsmoker and declining smoker trajectories is also important for this age-group, given the potential impacts of low-level smoking on health (Schane, Ling, & Glantz, 2010). Identifying developmental smoking patterns, including potential modifiable leverage points such as social influence factors and self-identified smoking status, will inform targeted interventions for young adults to reduce tobacco use.

Self-identifying as a social smoker may serve as a protective factor in some trajectories, while portending increased harm for others. Recognizing occasional smoking as risky may limit frequency or intensity of smoking in nonsmokers and classes that achieve moderate/stable smoking over time. However, labeling one’s self as a social smoker may also facilitate escalating smoking behavior by dissociating one’s behavior from known risks of the behavior. The fact that none of the trajectory classes achieved total smoking abstinence may reflect that social smokers perceive less of a need to quit and therefore continue smoking at low levels rather than quitting. Understanding the nuances of how smoking identity is formed and used to limit or facilitate smoking behavior in young adults will allow for the development of more effective interventions to reduce tobacco use in young adults.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Cancer Institute of the National Institutes of Health under Award Numbers R03CA187756 and 5P01CA098262.

Biography

Amanda L. Johnson was the lead research analyst at the Truth Initiative Schroeder Institute® throughout the development of this publication. Her interests include using longitudinal methods to explore how tobacco attitudes and behaviors change over time.

Andrea C. Villanti is an associate professor of psychiatry at The Vermont Center on Behavior and Health at The University of Vermont. Her research focuses on young adult tobacco use and translational research to improve tobacco control policy and program decision-making.

Valerie Williams is a senior scientist with General Dynamics Health Solutions. Her areas of interest include evaluative research methods, tobacco control, mental health services, public health surveillance, and criminal and juvenile justice.

Jessica M. Rath is the managing director at the Truth Initiative Schroeder Institute®. She is a behavioral scientist specializing in the planning, implementation and evaluation of large-scale interventions such as mass media campaigns for tobacco prevention.

Donna M. Vallone is the chief research officer at Truth Initiative. She leads the Truth Initiative Schroeder Institute® and oversees the evaluation of the national youth smoking prevention campaign truth. Her research interests include the influence of media messages to reduce tobacco use, particularly among lower socioeconomic status and racial/ethnic minority groups.

David B. Abrams is a professor of social and behavioral sciences at the College of Global Public Health at New York University. His interests include using systems frameworks to inform population health enhancement. His current focus is on using harm reduced nicotine products to accelerate saving a billion preventable premature deaths globally that are overwhelmingly caused by cigarettes and other deadly smoked tobacco products.

Donald Hedeker is a professor of biostatistics at the University of Chicago. HIs main expertise is in the development and dissemination of advanced statistical methods for clustered and longitudinal data. His research interests include missing data in longitudinal studies, Ecological Momentary Assessment (EMA) data, and mHealth studies focused on tobacco and addiction.

Robin J. Mermelstein is a Distinguished Professor in the Psychology Department, the Director of the Institute for Health Research and Policy, and Co-Director of the Center for Clinical and Translational Sciences a the University of Illinois at Chicago. Her research interests include understanding and reducing adolescent and young adult tobacco use, and developing and evaluating novel interventions for reducing cigarette smoking among all age groups.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

  1. Abroms L, Simons-Morton B, Haynie DL, & Chen R (2005). Psychosocial predictors of smoking trajectories during middle and high school. Addiction, 100, 852–861. doi: 10.1111/j.1360-0443.2005.01090.x [DOI] [PubMed] [Google Scholar]
  2. Arnett JJ (2000). Emerging adulthood. A theory of development from the late teens through the twenties. American Psychologist, 55, 469–480. [PubMed] [Google Scholar]
  3. Bernat DH, Erickson DJ, Widome R, Perry CL, & Forster JL (2008). Adolescent smoking trajectories: Results from a population-based cohort study. Journal of Adolescent Health, 43, 334–340. doi: 10.1016/j.jadohealth.2008.02.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brook DW, Brook JS, Zhang C, Whiteman M, Cohen P, & Finch SJ (2008). Developmental trajectories of cigarette smoking from adolescence to the early thirties: Personality and behavioral risk factors. Nicotine & Tobacco Research, 10, 1283–1291. doi: 10.1080/14622200802238993 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brook JS, Pahl K, & Ning Y (2006). Peer and parental influences on longitudinal trajectories of smoking among African Americans and Puerto Ricans. Nicotine & Tobacco Research, 8, 639–651. doi: 10.1080/14622200600789627 [DOI] [PubMed] [Google Scholar]
  6. Brook JS, Zhang C, Burke L, & Brook DW (2014). Trajectories of cigarette smoking from adolescence to adulthood as predictors of unemployment status. Nicotine & Tobacco Research, 16, 1559–1566. doi: 10.1093/ntr/ntu107 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brown AE, Carpenter MJ, & Sutfin EL (2011). Occasional smoking in college: Who, what, when and why? Addictive Behaviors, 36, 1199–1204. doi: 10.1016/j.addbeh.2011.07.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Caldeira KM, O’Grady KE, Garnier-Dykstra LM, Vincent KB, Pickworth WB, & Arria AM (2012). Cigarette smoking among college students: Longitudinal trajectories and health outcomes. Nicotine & Tobacco Research, 14, 777–785. doi: 10.1093/ntr/nts131 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chang L, & Krosnick JA (2009). National surveys via RDD telephone interviewing versus the Internet comparing sample representativeness and response quality. Public Opinion Quarterly, 73, 641–678. [Google Scholar]
  10. Chassin L, Presson CC, Pitts SC, & Sherman SJ (2000). The natural history of cigarette smoking from adolescence to adulthood in a midwestern community sample: Multiple trajectories and their psychosocial correlates. Health Psychology, 19, 223–231. [PubMed] [Google Scholar]
  11. Chen P, & Jacobson KC (2012). Developmental trajectories of substance use from early adolescence to young adulthood: Gender and racial/ethnic differences. Journal of Adolescent Health, 50, 154–163. doi: 10.1016/j.jadohealth.2011.05.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Colder CR, Flay BR, Segawa E, & Hedeker D (2008). Trajectories of smoking among freshmen college students with prior smoking history and risk for future smoking: Data from the University Project Tobacco Etiology Research Network (UpTERN) study. Addiction, 103, 1534–1543. doi: 10.1111/j.1360-0443.2008.02280.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Collins LM, & Lanza ST (2010). Latent class and latent transition analysis with applications in the social, behavioral, and health sciences. Hoboken, NJ: John Wiley & Sons. [Google Scholar]
  14. Costello DM, Dierker LC, Jones BL, & Rose JS (2008). Trajectories of smoking from adolescence to early adulthood and their psychosocial risk factors. Health Psychology, 27, 811–818. doi: 10.1037/0278-6133.27.6.811 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Dutra LM, Glantz SA, Lisha NE, & Song AV (2017). Beyond experimentation: Five trajectories of cigarette smoking in a longitudinal sample of youth. PLoS One, 12, e0171808. doi: 10.1371/journal.pone.0171808 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Falomir JM, & Invernizzi F (1999). The role of social influence and smoker identity in resistance to smoking cessation. Swiss Journal of Psychology/Schweizerische Zeitschrift für Psychologie/ Revue Suisse de Psychologie, 58, 73. [Google Scholar]
  17. Foldes SS, An LC, Rode P, Schillo BA, Davern M, Alesci NL, … Manley MW. (2010). The prevalence of unrecognized tobacco use among young adults. American Journal of Health Behavior, 34, 309–321. doi: 10.5555/ajhb.2010.34.3.309 [DOI] [PubMed] [Google Scholar]
  18. Fowler FJ, Gerstein BS, & Barry MJ (2013). How patient centered are medical decisions? Results of a national survey. JAMA Internal Medicine, 173, 1215–1221. [DOI] [PubMed] [Google Scholar]
  19. Fuemmeler B, Lee CT, Ranby KW, Clark T, McClernon FJ, Yang C, & Kollins SH (2013). Individual- and community-level correlates of cigarette-smoking trajectories from age 13 to 32 in a U.S. population-based sample. Drug and Alcohol Dependence, 132, 301–308. doi: 10.1016/j.drugalcdep.2013.02.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Grande D, Mitra N, Shah A, Wan F, & Asch DA (2013). Public preferences about secondary uses of electronic health information. JAMA Internal Medicine, 173, 1798–1806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Guillory J, Lisha N, Lee YO, & Ling PM (2017). Phantom smoking among young adult bar patrons. Tobacco Control, 26, 153–157. doi: 10.1136/tobaccocontrol-2015-052821 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hair E, Bennett M, Williams V, Johnson A, Rath J, Cantrell J, … Vallone D. (2017). Progression to established patterns of cigarette smoking among young adults. Drug and Alcohol Dependence, 177, 77–83. doi: 10.1016/j.drugalcdep.2017.03.040 [DOI] [PubMed] [Google Scholar]
  23. Hertel AW, & Mermelstein RJ (2012). Smoker identity and smoking escalation among adolescents. Health Psychology, 31, 467–475. doi: 10.1037/a0028923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hertel AW, & Mermelstein RJ (2016). Smoker identity development among adolescents who smoke. Psychol Addictive Behaviors, 30, 475–483. doi: 10.1037/adb0000171 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hoek J, Maubach N, Stevenson R, Gendall P, & Edwards R (2013). Social smokers’ management of conflicted identities. Tobacco Control, 22, 261–265. doi: 10.1136/tobaccocontrol-2011-050176 [DOI] [PubMed] [Google Scholar]
  26. Høie M, Moan IS, & Rise J (2010). An extended version of the theory of planned behaviour: Prediction of intentions to quit smoking using past behaviour as moderator. Addiction Research & Theory, 18, 572–585. [Google Scholar]
  27. Huh J, Huang Z, Liao Y, Pentz M, & Chou CP (2013). Transitional life events and trajectories of cigarette and alcohol use during emerging adulthood: Latent class analysis and growth mixture modeling. Journal of Studies on Alcohol and Drugs, 74, 727–735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jackson KM, Sher KJ, & Schulenberg JE (2008). Conjoint developmental trajectories of young adult substance use. Alcoholism: Clinical &Experimental Research, 32, 723–737. doi: 10.1111/j.1530-0277.2008.00643.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Jessor R (1991). Risk behavior in adolescence: A psychosocial framework for understanding and action. Journal of Adolescent Health, 12, 597–605. [DOI] [PubMed] [Google Scholar]
  30. Jiang N, Lee YO, & Ling PM (2014). Young adult social smokers: Their co-use of tobacco and alcohol, tobacco-related attitudes, and quitting efforts. Preventive Medicine, 69, 166–171. doi: 10.1016/j.ypmed.2014.09.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jung T, & Wickrama K (2008). An introduction to latent class growth analysis and growth mixture modeling. Social and Personality Psychology Compass, 2, 302–317. [Google Scholar]
  32. Klein EG, Bernat DH, Lenk KM, & Forster JL (2013). Non-daily smoking patterns in young adulthood. Addictive Behaviors, 38, 2267–2272. doi: 10.1016/j.addbeh.2013.03.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kumar S, Quinn SC, Kim KH, Daniel LH, & Freimuth VS (2012. The impact of workplace policies and other social factors on self-reported influenza-like illness incidence during the 2009 H1N1 pandemic. American Journal of Public Health, 102, 134–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lee JY, Brook JS, Finch SJ, & Brook DW (2016). Trajectories of cigarette smoking beginning in adolescence predict insomnia in the mid thirties. Substance Use & Misuse, 51, 616–624. doi: 10.3109/10826084.2015.1126747 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lessov-Schlaggar CN, Kristjansson SD, Bucholz KK, Heath AC, & Madden PA (2012). Genetic influences on developmental smoking trajectories. Addiction, 107, 1696–1704. doi: 10.1111/j.1360-0443.2012.03871.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lisha NE, Delucchi KL, Ling PM, & Ramo DE (2015). Prevalence and correlates of social smoking in young adults: Comparisons of behavioral and self-identified definitions. Nicotine & Tobacco Research, 17, 1076–1084. doi: 10.1093/ntr/ntu242 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Little RJ, & Rubin DB (1987). Statistical analysis with missing data. New York, NY: John A. Wiley & Sons. [Google Scholar]
  38. Little RJ, & Rubin DB (1989). The analysis of social science data with missing values. Sociological Methods & Research, 18, 292–326. [Google Scholar]
  39. Mayhew KP, Flay BR, & Mott JA (2000). Stages in the development of adolescent smoking. Drug and Alcohol Dependence, 59, S61–S81. [DOI] [PubMed] [Google Scholar]
  40. Mays D, Gilman SE, Rende R, Luta G, Tercyak KP, & Niaura RS (2014). Parental smoking exposure and adolescent smoking trajectories. Pediatrics, 133, 983–991. doi: 10.1542/peds.2013-3003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Moan IS, & Rise J (2005). Quitting smoking: Applying an extended version of the theory of planned behavior to predict intention and behavior. Journal of Applied Biobehavioral Research, 10, 39–68. [Google Scholar]
  42. Moran S, Wechsler H, & Rigotti NA (2004). Social smoking among US college students. Pediatrics, 114, 1028–1034. doi: 10.1542/peds.2003-0558-L [DOI] [PubMed] [Google Scholar]
  43. Muthén BO (2001). Second-generation structural equation modeling with a combination of categorical and continuous latent variables: New opportunities for latent class–latent growth modeling. In Collins L & Sayer A (Eds.), New methods for the analysis of change (pp. 291–322). Washington, DC: American Psychological Association. [Google Scholar]
  44. Muthén BO (2002). Beyond SEM: General latent variable modeling. Behaviormetrika, 29, 81–117. [Google Scholar]
  45. Muthen BO, & Muthen LK (2002). Mplus short course: Traditional latent variable modeling using Mplus. Washington, DC. [Google Scholar]
  46. Muthén LK, & Muthén BO (1998). Mplus user’s guide: The comprehensive modeling program for applied researchers. Los Angeles, CA: Author. [Google Scholar]
  47. Nagin D (2005). Group-based modeling of development. Cambridge, MA: Harvard University Press. [Google Scholar]
  48. Nagin DS (1999). Analyzing developmental trajectories: A semi-parametric, group-based approach. Psychological Methods, 4, 139. [DOI] [PubMed] [Google Scholar]
  49. Nylund KL, Asparouhov T, & Muthén BO (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14, 535–569. [Google Scholar]
  50. Oetting ER (1999). Primary socialization theory. Developmental stages, spirituality, government institutions, sensation seeking and theoretical implications. V. Substance Use & Misuse, 34, 947–982. [DOI] [PubMed] [Google Scholar]
  51. Orlando M, Tucker JS, Ellickson PL, & Klein DJ (2004). Developmental trajectories of cigarette smoking and their correlates from early adolescence to young adulthood. Journal of Consulting and Clinical Psychology, 72, 400–410. doi: 10.1037/0022-006X.72.3.400 [DOI] [PubMed] [Google Scholar]
  52. Pierce JP, White MM, & Messer K (2009). Changing age-specific patterns of cigarette consumption in the United States, 1992–2002: Association with smoke-free homes and state-level tobacco control activity. Nicotine & Tobacco Research, 11, 171–177. doi: 10.1093/ntr/ntp014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Pollard MS, Tucker JS, Green HD, Kennedy D, & Go MH (2010). Friendship networks and trajectories of adolescent tobacco use. Addictive Behaviors, 35, 678–685. doi: 10.1016/j.addbeh.2010.02.013 [DOI] [PubMed] [Google Scholar]
  54. Rath JM, Villanti AC, Abrams DB, & Vallone DM (2012). Patterns of tobacco use and dual use in US young adults: The missing link between youth prevention and adult cessation. Journal of environmental and public health, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Rhodes DJ, Breitkopf CR, Ziegenfuss JY, Jenkins SM, & Vachon CM (2015). Awareness of breast density and its impact on breast cancer detection and risk. Journal of Clinical Oncology, 33, 1143–1150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Richardson A, Williams V, Rath J, Villanti AC, & Vallone D (2014). The next generation of users: Prevalence and longitudinal patterns of tobacco use among US young adults. American Journal of Public Health, 104, 1429–1436. doi: 10.2105/AJPH.2013.301802 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Riggs NR, Chou CP, Li C, & Pentz MA (2007). Adolescent to emerging adulthood smoking trajectories: When do smoking trajectories diverge, and do they predict early adulthood nicotine dependence? Nicotine & Tobacco Research, 9, 1147–1154. doi: 10.1080/14622200701648359 [DOI] [PubMed] [Google Scholar]
  58. Schane RE, Ling PM, & Glantz SA (2010). Health effects of light and intermittent smoking: A review. Circulation, 121, 1518–1522. doi: 10.1161/CIRCULATIONAHA.109.904235 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Selya AS, Dierker LC, Rose JS, Hedeker D, & Mermelstein RJ (2012). Risk factors for adolescent smoking: parental smoking and the mediating role of nicotine dependence. Drug and Alcohol Dependence, 124, 311–318. doi: 10.1016/j.drugalcdep.2012.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Song AV, & Ling PM (2011). Social smoking among young adults: Investigation of intentions and attempts to quit. American Journal of Public Health, 101, 1291–1296. doi: 10.2105/AJPH.2010.300012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Stanton WR, Flay BR, Colder CR, & Mehta P (2004). Identifying and predicting adolescent smokers’ developmental trajectories. Nicotine & Tobacco Research, 6, 843–852. [DOI] [PubMed] [Google Scholar]
  62. Tein JY, Coxe S, & Cham H (2013). Statistical power to detect the correct number of classes in latent profile analysis. Structural Equation Modeling, 20, 640–657. doi: 10.1080/10705511.2013.824781 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Thompson AB, Mowery PD, Tebes JK, & McKee SA (2017). Time trends in smoking onset by sex and race/ethnicity among adolescents and young adults: Findings from the 2006–2013 national survey on drug use and health. Nicotine & Tobacco Research. doi: 10.1093/ntr/ntx010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. U.S. Department of Health and Human Services. (2012). Preventing tobacco use among youth and young adults: A report of the surgeon general. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 3. [Google Scholar]
  65. U.S. Department of Health and Human Services. (2014). The health consequences of smoking—50 years of progress. A report of the surgeon general. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 17. [Google Scholar]
  66. Villanti AC, Johnson AL, Rath JM, Williams V, Vallone DM, Abrams DB, … Mermelstein RJ. (2017). Identifying “social smoking” U.S. young adults using an empirically-driven approach. Addictive Behaviors, 70, 83–89. doi: 10.1016/j.addbeh.2017.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Villanti AC, Rath JM, & Vallone DM (2012). Quit behaviors in young adults: Comparing “smokers” and “social smokers.” Paper presented at the 2012 SRNT 18th Annual Meeting, Houston, TX. [Google Scholar]
  68. White HR, Bray BC, Fleming CB, & Catalano RF (2009). Transitions into and out of light and intermittent smoking during emerging adulthood. Nicotine & Tobacco Research, 11, 211–219. doi: 10.1093/ntr/ntn017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. White HR, Pandina RJ, & Chen PH (2002). Developmental trajectories of cigarette use from early adolescence into young adulthood. Drug and Alcohol Dependence, 65, 167–178. [DOI] [PubMed] [Google Scholar]
  70. Yeager DS, Krosnick JA, Chang L, Javitz HS, Levendusky MS, Simpser A, & Wang R (2011). Comparing the accuracy of RDD telephone surveys and internet surveys conducted withprobabilityand non-probability samples. Public Opinion Quarterly, 75(4), 709–747. [Google Scholar]

RESOURCES