Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jul 1.
Published in final edited form as: Addict Behav. 2017 Feb 10;70:83–89. doi: 10.1016/j.addbeh.2017.02.004

Identifying “Social Smoking” U.S. Young Adults Using an Empirically-Driven Approach

Andrea C Villanti 1,2, Amanda L Johnson 1, Jessica M Rath 2,3, Valerie Williams 3, Donna M Vallone 2,3,4, David B Abrams 1,2,5, Donald Hedeker 6, Robin J Mermelstein 7
PMCID: PMC5390897  NIHMSID: NIHMS853427  PMID: 28214741

Abstract

The phenomenon of “social smoking” emerged in the past decade as an important area of research, largely due to its high prevalence in young adults. The purpose of this study was to identify classes of young adult ever smokers based on measures of social and contextual influences on tobacco use. Latent class models were developed using social smoking measures, and not the frequency or quantity of tobacco use. Data come from a national sample of young adult ever smokers aged 18–24 (Truth Initiative Young Adult Cohort Study, N = 1,564). The optimal models identified three latent classes: Class 1 – Nonsmokers (52%); Class 2 – Social smokers (18%); and Class 3 – Smokers (30%). Nearly 60% of the “social smoker” class self-identified as a social smoker, 30% as an ex-smoker/tried smoking, and 12% as a non-smoker. The “social smoker” class was most likely to report using tobacco mainly or only with others. Past 30-day cigarette use was highest in the “smoker” class. Hookah use was highest in the “social smoker” class. Other tobacco and e-cigarette use was similar in the “social smoker” and “smoker” classes. Past 30-day tobacco and e-cigarette use was present for all products in the “non-smoker” class. Young adult social smokers emerge empirically as a sizable, distinct class from other smokers, even without accounting for tobacco use frequency or intensity. The prevalence of hookah use in “social smokers” indicates a group for which the social aspect of tobacco use could drive experimentation and progression to regular use.

Keywords: Models, Statistical, Population Surveillance, Smoking/epidemiology, Tobacco, Young Adult

1. Introduction

The phenomenon of “social smoking,” or smoking primarily in social contexts (Schane, Glantz, & Ling, 2009b), emerged in the past decade as an important area of research, largely due to the high prevalence of this behavior in college students, a group of young adults who often engage in a variety of health-risk behaviors at high rates (Schane, Glantz, & Ling, 2009a). More broadly, social smoking young adults comprise a subset of nondaily (occasional) smokers who report cigarette use in social situations or when drinking alcohol with others (Berg et al., 2012; Berg et al., 2009; Hoek, Maubach, Stevenson, Gendall, & Edwards, 2013; Moran, Wechsler, & Rigotti, 2004; Schane et al., 2009a; Shiffman, Kirchner, Ferguson, & Scharf, 2009) and national surveys estimate the prevalence of social smoking between 51% and 80% in young adult (college and non-college) smokers (Jiang, Lee, & Ling, 2014; Lisha, Delucchi, Ling, & Ramo, 2015; Moran et al., 2004; Song & Ling, 2011; Villanti, Rath, & Vallone, 2012). There are two main concerns with the long-term tobacco use behaviors of social smoking young adults: (1) that, like light and intermittent young adult smokers, approximately half will become heavy smokers in a short period (H. R. White, Bray, Fleming, & Catalano, 2009); or (2) that they will maintain a stable low-level smoking pattern throughout adulthood (Schane et al., 2009b). Both smoking patterns can lead to tobacco-related disease or death (Schane, Ling, & Glantz, 2010; U.S. Department of Health and Human Services, 2014). To the extent that they resist identifying themselves as “smokers” (Berg et al., 2009; Choi, Choi, & Rifon, 2011; Hoek et al., 2013; Levinson et al., 2007; Ridner, Walker, Hart, & Myers, 2010), young adult “social smokers” will continue to elude traditional tobacco prevention or cessation interventions.

Primary socialization theory (PST) posits that individuals learn normative and deviant behavior from a small number of social influences that change dynamically with development from childhood to adolescence and young adulthood (Oetting, 1999). In line with this theory, recent studies have identified a number of social influences on smoking behavior in young adults, including: living with a smoker (Klein, Forster, & Erickson, 2013), having close friends who smoke (Klein et al., 2013), and the number of one’s friends who smoke (Dietz, Sly, Lee, Arheart, & McClure, 2013). A study of smoking trajectories among college student smokers also showed positive associations between perceived close-friend approval of smoking, exposure to a social smoking environment, and large increases in number of cigarettes smoked (Colder, Flay, Segawa, & Hedeker, 2008). Similarly, a systematic review of longitudinal population-based studies of adolescent and young adult smokers showed that not having friends who smoke and being able to resist peer pressure to smoke were robust predictors of quitting in these age groups (Cengelli, O’Loughlin, Lauzon, & Cornuz, 2012).

The social context is another important influence on smoking in young adults. Studies on cue reactivity in adult and young adult smokers indicate that non-daily smokers are more highly influenced to smoke by situational contexts, including smoking by others and drinking alcohol, compared to daily smokers (Ferguson, Shiffman, Dunbar, & Schuz, 2016; Shiffman, Dunbar, & Ferguson, 2015; Shiffman et al., 2014; Shiffman, Li, et al., 2015; Thrul, Buhler, & Ferguson, 2014). Bar/club attendance and alcohol use in young adults have also been associated with current smoking (Dietz et al., 2013) and reduced cessation behavior among very light smokers (Jiang & Ling, 2013).

Several studies have identified latent subgroups of cigarette smokers (Rose et al., 2007; Sutfin, Reboussin, McCoy, & Wolfson, 2009) or polysubstance users, including tobacco users (Conway et al., 2013; Quek et al., 2013; A. White et al., 2013) in adolescent and young adult samples. Other studies have focused on multiple tobacco product use in a national sample of adolescents (Nasim, Blank, Cobb, & Eissenberg, 2012) and in a Midwestern sample of young adults (Erickson, Lenk, & Forster, 2014). Our previous work has identified latent classes of combustible tobacco users, including a class of non-daily light social smokers (Villanti, Pearson, Cantrell, Vallone, & Rath, 2015), but none of these studies have examined whether social smoking emerges as a distinct class when looking at social and contextual influences on tobacco use in young adults.

The extent to which social influences drive the development – or deterrence – of smoking behavior in young adulthood is key to understanding the phenomenon of “social smoking.” The purpose of this study was to identify clusters of ever smokers in a recent cohort of young adults based on measures of social and contextual influences on tobacco use and to determine common characteristics of these clusters, particularly social smoking young adults. Findings from this study were expected to identify the measures most relevant to identifying “social smokers” and inform intervention approaches to mitigate or harness social and contextual influences to reduce smoking in young adults.

2. Material and Methods

The current study leverages data from seven consecutive, bi-annual waves of the Truth Initiative Young Adult Cohort Study (YA Cohort; July 2011–October 2014), a large contemporary cohort of U.S. young adults that includes information on trajectories of smoking behavior, social influences on smoking, and social and contextual influences on smoking behavior among young adults. The detailed methods of this study have been described elsewhere (Rath, Villanti, Abrams, & Vallone, 2012). The cohort 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. African American and Hispanic young adults were oversampled to ensure sufficient sample sizes for subgroup analyses and the survey was administered online in English and Spanish. This methodology has been reported previously (Chang & Krosnick, 2009; Yeager et al., 2011), and it has been used broadly in the peer-reviewed medical literature (Fowler, Gerstein, & Barry, 2013; Grande, Mitra, Shah, Wan, & Asch, 2013; Kumar, Quinn, Kim, Daniel, & Freimuth, 2012; Rhodes, Radecki Breitkopf, Ziegenfuss, Jenkins, & Vachon, 2015). The cohort is refreshed at each wave to retain the initial sample size.

The panel recruitment rate (RECR; The American Association for Public Opinion Research, 2015) ranged from 13.9%–14.9% across the seven waves. In 64 to 66% of the identified households, one member completed a core profile survey in which key demographic information was collected (profile rate—PROR). At each wave, only 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 (COMR) ranged from 46.2%–68.4% across waves. The cumulative response rate (CUMRR1) (a product of these three rates) ranged from 4.4%–6.6% (Callegaro & DiSogra, 2008). This study was approved by the Independent Investigational Review Board, Inc. for Waves 1–3 (Protocol #20036-007) and Chesapeake Institutional Review Board, Inc. for Waves 4–7 (Protocol #20036020). Online consent was collected from participants before survey self-administration. The present analysis focused on a subset of participants 18–24 at study entry (N = 1,564) who entered at any of the seven waves of data collection, reported ever smoking a cigarette, and provided complete responses on social smoking-related measures.

Among young adults aged 18–24 at study entry who had ever smoked a cigarette (n = 1,564), 69% were aged 21 or older and 61% were female (Table 3). The majority of respondents were White, non-Hispanic (60%). Twelve percent were Black, non-Hispanic, 2% were Hispanic, and 7% reported another race. Most participants had completed some college education (59%), with 29% reporting high school education, and 12% less than high school education. Twenty percent of the sample reported depressive symptoms and 22% reported anxiety symptoms in the past two weeks.

2.1. Measures

2.1.1. Variables in the latent class model

In line with PST, six social smoking measures comprised the latent class model: self-identified smoking status; confidence in resisting smoking in social situations; smoking at parties or social events; smoking at bars or restaurants; social smoking behavior; and alcohol use. All social smoking measures were assessed at study entry, regardless of wave of entry. Subgroups of participants based on self-identified smoking status were identified using 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.” For analyses, self-identified “social smoker” and “occasional smoker” were collapsed into one category. Additionally, “ex-smoker” and “tried smoking” were collapsed into one category. Confidence in resisting smoking in social situations was assessed using a 4-point scale (“Not at all confident”, “somewhat confident”, “moderately confident”, and “very confident”). Participants who selected “very confident” were defined as having the ability to resist social smoking. Participants were also asked if they now use any tobacco products “not at all”, “some days”, or “every day.” Those who reported “some days” or “every day” use were defined as current tobacco users. Current tobacco users were asked where they used tobacco products, including parties/social events and bars/restaurants (yes/no). Participants who reported using tobacco “not at all” were included with those who reported “no” to using tobacco at parties/social events or at bars/restaurants. As a result, 66% of participants were classified as not smoking at parties/social events and of those, 73% had not used a tobacco product in the past 30 days. Similarly, 79% of participants classified as not smoking at bars/restaurants and of those, 63% had not used a tobacco product in the past 30 days. Current tobacco users were asked about the frequency of using tobacco products alone relative to using tobacco with others (“alone”, “equally alone and with others”, “mainly with others”, “only with others”). For analyses, “mainly with others” and “only with others” were collapsed into a single category. Participants who did not report current tobacco use were coded in a new category as “not a current smoker.” Current alcohol use was defined as “every day” or “some days” use for Waves 1–6. In Wave 7, current alcohol use was defined as past month use of alcohol among past year drinkers. Those who were not current users could either be never users or not have used in the past 30 days; participants with missing data on alcohol use were also classified as non-current users.

2.1.2. Other correlates

As part of the KnowledgePanel® routine data collection, participants provided information at study entry on age, gender, race/ethnicity (White, non-Hispanic; Black, non-Hispanic; Other, non-Hispanic; and Hispanic) and educational attainment (less than high school, high school, and some college or greater). Current depressive and anxiety symptoms were assessed using two-item Patient Health Questionnaire (PHQ-2; Kroenke et al 2003) or the two-item Generalized Anxiety Disorder (GAD-2; Kroenke et al 2007) scale, respectively. Current marijuana and other drug use (cocaine, ecstasy, meth, etc.) were defined as “some days” or “every day” use; participants who reported “not at all” or had missing data were defined as non-current users. Participants were asked about past 30-day use (yes/no) of seven tobacco products including cigarettes, cigars (traditional cigars, little cigars, cigarillos, and bidis), electronic cigarettes (e-cigarettes), smokeless tobacco (chew, dip/snuff), and snus.

2.2 Data analysis

Latent class analyses were conducted in 2015 using Mplus 7.0 (www.statmodel.com) to identify subgroups based on social smoking measures in both samples of young adults. Selection of variables for the latent class model was informed by available measures in each study and by consensus agreement among all study authors. Full information maximum likelihood estimation (FIML) was used based on the assumption that the probability of missingness may depend on data that are observed (Enders & Bandalos, 2001). The optimal number of classes was determined by running models with a successive number of classes from two to five and comparing model fit indices, the odds of correct classification (OCC), entropy, and interpretability. Model fit indices included the log likelihood (−2LL), the Akaike Information Criterion, the Bayesian Information Criterion (BIC), and the sample size adjusted BIC, as well as Pearson and likelihood ratio chi-square statistics, and the Lo-Mendell-Rubin and Vuong-Lo-Mendell-Rubin likelihood ratio tests (Collins & Lanza, 2010). The optimal model was selected with the number of classes that minimized BIC, based on evidence showing that BIC outperformed other model fit indices in a simulation study (Nylund, Asparouhov, & Muthén, 2007), demonstrated odds of correct classification greater than five across all classes (Nagin, 2005), and was interpretable. Sensitivity analysis examined the extent to which the self-identified smoking status measure included in the model drove the formation of the latent classes.

Bivariate analyses were conducted in Mplus (version 7.4) to estimate the prevalence of the social smoking measures in the final latent class model, as well as the relationship between other correlates and latent class assignment. Chi-square tests were employed to identify differences in characteristics across latent class assignment with significance at p<0.05.

3. Results

3.1 Selection of the latent class model

Table 1 presents model fit statistics for all latent class models. In the YA Cohort data, the three-class solution was chosen as the best model. The two- through five-class models yielded entropy greater than 0.85 and significant LMR-LRT and VLMR-LRT test statistics at p<0.0001. Since the four- and five-class models did not report adequate latent class separation (OCC < 5), these models were not considered for the final model. The three-class model was selected over the two-class model due to the minimized BIC and retained odds of correct classification greater than five across all classes.

Table 1.

Optimal number of latent classes: assessment statisticsa

Model fit
Latent classes Number of free parameters LL BIC Sample size adjusted BIC LMR p-value VLMR LRT p-value Entropy
1 10 −7882.5 15838.6 15806.9 - - -
2 21 −6536.0 13226.5 13159.8 <0.0001 <0.0001 0.93
3 32 −6374.1 12983.5 12881.9 <0.0001 <0.0001 0.86
4 47 −6551.0 13447.8 13298.5 <0.0001 <0.0001 0.87
5 59 −6481.9 13397.8 13210.4 <0.0001 <0.0001 0.87
Odds of correct classificationb
Class 1 Class 2 Class 3 Class 4 Class 5
1 ∞
2 51.6 54.6
3 7.1 36.0 9.5
4 4.6 14.6 7.7 24.6
5 37.5 6.1 30.3 3.7 7.3
Open in a new tab

LL Log likelihood; BIC Bayesian Information Criteria; LMR Lo-Mendell-Rubin; VLMR Vuong-Lo-Mendell-Rubin; LRT likelihood ratio test for k (H0) versus k-1 classes

a

Three class model highlighted as best fit

b

Odds of correct classification (OCC) > 5 indicates a model with good latent class separation (Collins & Lanza, p. 74); ∞ Indicates perfect classification

3.2 Identification of latent classes

Response to the social smoking measures by latent class, along with the unweighted probability of class membership, and class size are presented in Table 2. In this sample, the three latent classes are described as: nonsmokers (Class 1, n=820, 52%); social smokers (Class 2, n=275, 18%); and smokers (Class 3, n=469, 30%). Average latent class probabilities for most likely latent class membership were 97% (Class 1), 88% (Class 2), and 91% (Class 3).

Table 2.

Response to the social smoking measures by latent class, Truth Initiative Young Adult Cohort Study (N = 1,564)

Class 1 Class 2 Class 3
Non smokers Social smokers Smokers Full sample
% % % %
Self-identified smoking status
 Smoker 1% 1% 74% 22%
 Social or occasional smoker 12% 57% 22% 24%
 Ex-smoker/tried smoking 43% 30% 4% 29%
 Non-smoker 44% 12% 0% 25%
Confident can resist smoking in social situationsa 82% 56% 11% 57%
Smoke at parties or social events 0% 71% 71% 34%
Smoke at bars or restaurants 0% 41% 47% 21%
Social smoking behaviora
 Not current smoker 85% 0% 0% 44%
 Smoke alone 6% 5% 12% 8%
 Smoke equally alone and with others 4% 29% 82% 31%
 Smoke mainly or only with others 5% 66% 6% 17%
Alcohol use, every day or some daysb 67% 89% 71% 72%
Pr(Class membership), unweighted 52% 18% 30% 100%
Class size, unweighted 820 275 469 1564
Open in a new tab
a

Missingness: Confidence to resist smoking in social situations (n=2, 0.1%), social smoking behavior (n=12, 0.8%).

b

Individuals with missing alcohol use frequency (n = 3, 0.2%) were defined as non-current users.

In the full sample, self-identified smoking status was roughly evenly distributed among the four categories, though the distribution of responses varied by latent class. The “non-smoker” class largely identified as non-smokers (44%) or ex-smokers/tried smoking (43%), while 74% of the “smoker” class self-identified as smokers and 57% of the “social smoker” class self-identified as “social or occasional smokers.” Confidence to resist smoking in social situations was greatest in the “non-smoker” class and lowest in the “smoker” class. The “non-smoker” class did not endorse smoking at parties or social events (0%), bars or restaurants (0%), but prevalence of these behaviors was similar in the “smoker” and “social smoker” classes. The “social smoker” class was most likely to endorse that they smoked “mainly or only with others” (66%), while the smoker class largely reported that they smoked “equally alone and with others” (82%). However, 29% of respondents in the “social smoker” class reported smoking “equally alone and with others” and 5% “alone.” The majority of the full sample had used alcohol every day or some days (72%), with the non-smoker class least likely to endorse alcohol use (67%) and the social smoker class most likely (89%).

3.3 Correlates of latent class

Table 3 presents the bivariate analyses of the sociodemographic, mental health, and tobacco use correlates of latent class. There were significant differences in gender and education across the latent classes, with nonsmokers being the most likely to be female and college-educated and social smokers reporting similarly high levels of education. Social smokers were the most likely to be under age 21 compared to the other two classes and were similar to nonsmokers with respect to white vs. non-white racial/ethnic composition. Smokers and social smokers were equally likely to be female and report a similar prevalence of depressive and anxiety symptoms, marijuana use, and other drug use, which were higher than the nonsmoker class. Social smokers were equally likely as smokers to report smoking at parties or social events and smoking at bars or restaurants. With respect to tobacco, the prevalence of past 30-day cigarette use was highest in the smoker class, followed by the social smoker class. The social smoker class had a higher prevalence of hookah use compared to the smoker class, with a similar prevalence of all other product use in these two groups. Past 30-day tobacco and e-cigarette use was present (i.e., non-zero) for all products in the “non-smoker” class.

Table 3.

Characteristics of the latent classes, unweighted

Total Class 1 Nonsmokers Class 2 Social smokers Class 3 Smokers
N = 1,564 N = 820 (52%) N = 275 (18%) N = 469 (30%) p*
Age 21.72 (0.05) 21.73 (0.06) 21.47 (0.11) 21.87 (0.08)
 18–20 31.3% 31.2% 39.3% 26.7% 0.25
 21+ 68.7% 68.8% 60.7% 73.4%
Gender
 Male 38.9% 35.7% 41.8% 42.6% 0.01
 Female 61.1% 64.3% 58.2% 57.4%
Race
 White, non-Hispanic 59.5% 57.0% 57.8% 64.8% <0.001
 Black, non-Hispanic 12.0% 11.1% 9.1% 15.1%
 Other, non-Hispanic 6.5% 6.0% 8.4% 6.2%
 Hispanic 2.2% 26.0% 24.7% 13.9%
Education
 Less than high school 12.3% 9.1% 9.1% 19.6% <0.001
 High school 28.5% 24.0% 26.9% 37.1%
 Some college or greater 59.3% 66.8% 64.0% 43.3%
Mental health/substance use
 Depressiona 20.3% 17.8% 21.5% 23.9% 0.01
 Anxietyb 22.1% 19.0% 22.5% 27.3% 0.001
 Marijuanac 25.3% 17.3% 35.3% 33.3% <0.001
 Other drug used 4.3% 3.0% 5.8% 5.8% 0.02
Past 30-day tobacco use
 Cigarette use 41.2% 4.9% 58.9% 94.5% <0.001
 Cigar usee 11.8% 2.7% 22.5% 21.5% <0.001
 Hookah use 6.6% 2.2% 16.4% 8.5% <0.001
 E-cigarette use 4.9% 1.8% 8.4% 8.1% <0.001
 Smokeless usef 4.0% 1.2% 8.4% 6.4% <0.001
 Snus use 1.6% 0.4% 2.9% 3.0% <0.001
Open in a new tab

P-values are from the chi-square test for difference across groups.

a

Depressive symptomatology is defined as a score of at least 3 on the PHQ-2.

b

Anxiety symptoms are defined as a score of at least 3 on the GAD-2.

c

Marijuana use is assessed in the past 30 days. Participants with missing data on marijuana use (n=9, 0.6%) were defined as non-current users.

d

Other drug use is assessed in the past 30 days. Participants with missing data on other drug use (n = 6, 0.4%) were defined as non-current users.

e

Cigar use includes use of little cigars, cigarillos, and large cigars.

f

Smokeless use includes dip, snuff, and chewing tobacco.

3.4 Sensitivity analysis

Sensitivity analyses removing self-identified smoking status from the latent class model highlighted that the classes were largely defined by this variable. When latent class models were run without the self-identified smoking status measure, the best fitting model was a 2-class model with “smokers” and “non-smokers” (Supplemental Table 1). However, when comparing the distribution of responses to the social smoking measures in the final three-class model (Table 2) versus using the self-identified smoking status variable alone, the latent class model was better at differentiating patterns of social smoking behavior, with “social smokers” (Class 2) most likely to report smoking “mainly or only with others” (66%) and “smokers” (Class 3) most likely to report smoking “equally alone and with others” (82%). Using only the self-identified smoking status variable, those who identified as “social smokers” were more likely to report smoking “equally alone and with others” (39%) than “mainly or only with others” (33%; p <0.001).

4. Discussion

Using a recent national sample of U.S. young adults, this study highlights three latent classes of ever cigarette smokers: nonsmokers, social smokers, and smokers. In line with primary socialization theory, these classes were characterized by the social context in which they smoked cigarettes, confidence resisting smoking in social situations, self-identified smoking status, and a common element in young adult smoking contexts – alcohol use. When examining past 30-day tobacco use as correlates of these three latent classes, past 30-day cigarette use was highest in the “smoker” class and past 30-day non-cigarette tobacco and e-cigarette use was higher or similar for most products in the “social smoker” class compared to the “smoker” class. These findings indicate that there is a group of young adults for which the social aspect of tobacco use could drive experimentation and progression to regular use. Consistent with previous research (Hassmiller, Warner, Mendez, Levy, & Romano, 2003; Sutfin et al., 2012; Wortley, Husten, Trosclair, Chrismon, & Pederson, 2003), social smokers were less likely to be white and more likely to be educated compared to smokers.

Previous studies have defined social smoking based on measures of behavior (Shiffman, Li, et al., 2015), identity (Moran et al., 2004), or both (Lisha et al., 2015; Song & Ling, 2011). In our study, a single measure of self-identified smoking status drove the formation of latent classes, with the “social smoker” class having the largest percentage of self-identified “social smokers.” Interestingly, 12% of respondents in the “social smoker” class identified as being a “non-smoker” and 30% as being an “ex-smoker” or having “tried smoking.” Nearly 60% of respondents in the “social smoker” class reported past 30-day cigarette use. Additionally, use of tobacco products and e-cigarettes was present in the “non-smoker” class. Recently, the term “phantom smoker” has emerged to characterize young adults who report smoking cigarettes, but do not self-identify as a smoker (Choi et al., 2011; Guillory, Lisha, Lee, & Ling, 2016). Among past 30-day young adult cigarette smokers recruited from 2012–2014 in U.S. bars and nightclubs, phantom smokers were 60% more likely to identify as social smokers than regular smokers (Guillory et al., 2016). Identifying as a social smoker seems to have utility for young adult smokers in dissociating their smoking behavior from the known harms of smoking. Since they do not see themselves as “smokers” or their smoking as problematic (Choi et al., 2011; Hoek et al., 2013), social smoking young adults may experience little pressure from others to quit smoking or feel reluctant to change their behavior. Identification of intervention strategies to prevent smoking escalation or increase cessation in this group is greatly needed.

The social nature of today’s young adults (Solomon, 2014) may partially explain the growth of “social smoking” in this group. This is reflected in the tobacco products used by “social smoker” young adults and the social nature of tobacco use in this group. Findings from our study highlight that past 30-day hookah use was nearly twice as high in the “social smoker” class as in the “smoker” class in line with studies noting the social nature of hookah use in young adults (Castaneda, Barnett, Soule, & Young, 2016; Sharma, Beck, & Clark, 2013; Sharma, Clark, & Sharp, 2014) and the depiction of social aspects of hookah use on social media (Carroll, Shensa, & Primack, 2013; Primack, Carroll, Shensa, Davis, & Levine, 2016). Of note, recent qualitative research supports the role of social modeling in hookah initiation among young adults (Castaneda et al., 2016). Respondents in the “social smoker” class endorsed smoking or using tobacco “mainly or only with others,” reinforcing the social utility of tobacco use in this group.

One strength of the current study is use of a contemporary cohort of young adults who came of age following the Master Settlement Agreement and increases in state-level tobacco control efforts. One limitation of the study was that latent class models run without the self-identified smoking status item did not produce the same number of classes, nor the distinct classes identified when the item was included. Another is the low cumulative response rate across survey waves.

5. Conclusions

The current study demonstrates that “social smokers” emerge as a distinct class from “smokers” based on contextual influences, even without accounting for tobacco use frequency or intensity. In this national sample of U.S. young adults, the” social smoker” class is sizable, accounting for 18% of young adults who have ever tried cigarettes. “Social smokers” look remarkably similar to “smokers” with respect to past 30-day cigar, e-cigarette, smokeless tobacco, and snus use, though they had a higher prevalence of hookah use. Findings from this study highlight the importance of the self-identified smoking status measure in identifying social smokers. They also indicate that interventions targeting hookah are likely to have a greater impact on “social smokers” than smoking. This study suggests that future interventions addressing contextual influences on smoking, including building confidence to resist smoking in social situations and decoupling smoking from alcohol use, may be particularly effective in reducing tobacco use among “social smokers.”

Supplementary Material

supplement

Highlights.

  • Three latent classes were identified from six social smoking measures.

  • Classes included young adult Nonsmokers, Social smokers, and Smokers.

  • Social smokers were most likely to report using tobacco mainly or only with others.

  • Smokers and Social smokers had similar prevalence of tobacco and e-cigarette use.

  • Social factors may drive progression to regular tobacco use in Social smokers.

Acknowledgments

Role of Funding Sources

This work was supported by Truth Initiative and the National Cancer Institute of the National Institutes of Health [grant numbers R03CA187756, 5P01CA098262]. The funders had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

AUTHOR DISCLOSURE STATEMENTS

Contributors

AV designed the study. AJ conducted the statistical analysis, with feedback from all study authors. AV and AJ wrote the first draft of the manuscript and all authors contributed to and have approved the final manuscript.

Conflict of Interest

The authors declare that they have no conflicts of interest.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Berg CJ, Ling PM, Hayes RB, Berg E, Nollen N, Nehl E, … Ahluwalia JS. Smoking frequency among current college student smokers: distinguishing characteristics and factors related to readiness to quit smoking. Health Education Research. 2012;27(1):141–150. doi: 10.1093/her/cyr106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Berg CJ, Lust KA, Sanem JR, Kirch MA, Rudie M, Ehlinger E, … An LC. Smoker self-identification versus recent smoking among college students. Am J Prev Med. 2009;36(4):333–336. doi: 10.1016/j.amepre.2008.11.010. [DOI] [PubMed] [Google Scholar]
  3. Callegaro M, DiSogra C. Computing response metrics for online panels. Public Opinion Quarterly. 2008;72(5):1008–1032. [Google Scholar]
  4. Carroll MV, Shensa A, Primack BA. A comparison of cigarette- and hookah-related videos on YouTube. Tobacco Control. 2013;22(5):319–323. doi: 10.1136/tobaccocontrol-2011-050253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Castaneda G, Barnett TE, Soule EK, Young ME. Hookah smoking behavior initiation in the context of Millennials. Public Health. 2016 doi: 10.1016/j.puhe.2016.02.013. Epub ahead of print, Mar 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cengelli S, O’Loughlin J, Lauzon B, Cornuz J. A systematic review of longitudinal population-based studies on the predictors of smoking cessation in adolescent and young adult smokers. Tob Control. 2012;21(3):355–362. doi: 10.1136/tc.2011.044149. [DOI] [PubMed] [Google Scholar]
  7. Chang L, Krosnick JA. National surveys via RDD telephone interviewing versus the internet comparing sample representativeness and response quality. Public Opinion Quarterly. 2009;73(4):641–678. [Google Scholar]
  8. Choi Y, Choi SM, Rifon N. “I smoke but I am not a smoker”: phantom smokers and the discrepancy between self-identity and behavior. Journal of American College Health. 2011;59(2):117–125. doi: 10.1080/07448481.2010.483704. [DOI] [PubMed] [Google Scholar]
  9. Colder CR, Flay BR, Segawa E, Hedeker D. 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. 2008;103(9):1534–1543. doi: 10.1111/j.1360-0443.2008.02280.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Collins LM, Lanza ST. Latent Class and Latent Transition Analysis with Applications in the Social, Behavioral, and Health Sciences. Hoboken, NJ: John Wiley & Sons, Inc; 2010. [Google Scholar]
  11. Conway KP, Vullo GC, Nichter B, Wang J, Compton WM, Iannotti RJ, Simons-Morton B. Prevalence and patterns of polysubstance use in a nationally representative sample of 10th graders in the United States. Journal of Adolescent Health. 2013;52(6):716–723. doi: 10.1016/j.jadohealth.2012.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Dietz NA, Sly DF, Lee DJ, Arheart KL, McClure LA. Correlates of smoking among young adults: the role of lifestyle, attitudes/beliefs, demographics, and exposure to anti-tobacco media messaging. Drug and Alcohol Dependence. 2013;130(1–3):115–121. doi: 10.1016/j.drugalcdep.2012.10.019. [DOI] [PubMed] [Google Scholar]
  13. Enders CK, Bandalos DL. The relative performance of full information maximum likelihood estimation for missing data in structural equation models. Structural Equation Modeling. 2001;8(3):430–457. [Google Scholar]
  14. Erickson DJ, Lenk KM, Forster JL. Latent classes of young adults based on use of multiple types of tobacco and nicotine products. Nicotine & Tobacco Research. 2014;16(8):1056–1062. doi: 10.1093/ntr/ntu024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ferguson SG, Shiffman S, Dunbar M, Schuz N. Higher stimulus control is associated with less cigarette intake in daily smokers. Psychology of Addictive Behaviors. 2016;30(2):229–237. doi: 10.1037/adb0000149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fowler FJ, Jr, Gerstein BS, Barry MJ. How patient centered are medical decisions?: Results of a national survey. JAMA Internal Medicine. 2013;173(13):1215–1221. doi: 10.1001/jamainternmed.2013.6172. [DOI] [PubMed] [Google Scholar]
  17. Grande D, Mitra N, Shah A, Wan F, Asch DA. Public preferences about secondary uses of electronic health information. JAMA Internal Medicine. 2013;173(19):1798–1806. doi: 10.1001/jamainternmed.2013.9166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Guillory J, Lisha N, Lee YO, Ling PM. Phantom smoking among young adult bar patrons. Tobacco Control. 2016 doi: 10.1136/tobaccocontrol-2015-052821. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hassmiller KM, Warner KE, Mendez D, Levy DT, Romano E. Nondaily smokers: who are they? American Journal of Public Health. 2003;93(8):1321–1327. doi: 10.2105/ajph.93.8.1321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hoek J, Maubach N, Stevenson R, Gendall P, Edwards R. Social smokers’ management of conflicted identities. Tob Control. 2013;22(4):261–265. doi: 10.1136/tobaccocontrol-2011-050176. [DOI] [PubMed] [Google Scholar]
  21. Jiang N, Lee YO, Ling PM. Young adult social smokers: their co-use of tobacco and alcohol, tobacco-related attitudes, and quitting efforts. Prev Med. 2014;69:166–171. doi: 10.1016/j.ypmed.2014.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jiang N, Ling PM. Impact of alcohol use and bar attendance on smoking and quit attempts among young adult bar patrons. Am J Public Health. 2013;103(5):e53–61. doi: 10.2105/AJPH.2012.301014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Klein EG, Forster JL, Erickson DJ. Longitudinal predictors of stopping smoking in young adulthood. Journal of Adolescent Health. 2013;53(3):363–367. doi: 10.1016/j.jadohealth.2013.04.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kumar S, Quinn SC, Kim KH, Daniel LH, Freimuth VS. 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. 2012;102(1):134–140. doi: 10.2105/AJPH.2011.300307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Levinson AH, Campo S, Gascoigne J, Jolly O, Zakharyan A, Tran ZV. Smoking, but not smokers: identity among college students who smoke cigarettes. Nicotine & Tobacco Research. 2007;9(8):845–852. doi: 10.1080/14622200701484987. [DOI] [PubMed] [Google Scholar]
  26. Lisha NE, Delucchi KL, Ling PM, Ramo DE. Prevalence and Correlates of Social Smoking in Young Adults: Comparisons of Behavioral and Self-Identified Definitions. Nicotine & Tobacco Research. 2015;17(9):1076–1084. doi: 10.1093/ntr/ntu242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Moran S, Wechsler H, Rigotti NA. Social smoking among US college students. Pediatrics. 2004;114(4):1028–1034. doi: 10.1542/peds.2003-0558-L. [DOI] [PubMed] [Google Scholar]
  28. Nagin D. Group-Based Modeling of Development. Cambridge, MA: Harvard University Press; 2005. [Google Scholar]
  29. Nasim A, Blank MD, Cobb CO, Eissenberg T. Patterns of alternative tobacco use among adolescent cigarette smokers. Drug and Alcohol Dependence. 2012;124(1–2):26–33. doi: 10.1016/j.drugalcdep.2011.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling. 2007;14(4):535–569. [Google Scholar]
  31. Oetting ER. Primary socialization theory. Developmental stages, spirituality, government institutions, sensation seeking and theoretical implications. V. Substance Use & Misuse. 1999;34(7):947–982. doi: 10.3109/10826089909039389. [DOI] [PubMed] [Google Scholar]
  32. Primack BA, Carroll MV, Shensa A, Davis W, Levine MD. Sex Differences in Hookah-Related Images Posted on Tumblr: A Content Analysis. J Health Comm. 2016;21(3):366–375. doi: 10.1080/10810730.2015.1095814. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Quek LH, Chan GC, White A, Connor JP, Baker PJ, Saunders JB, Kelly AB. Concurrent and simultaneous polydrug use: latent class analysis of an Australian nationally representative sample of young adults. Frontiers in Public Health. 2013;1:61. doi: 10.3389/fpubh.2013.00061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rath JM, Villanti AC, Abrams DB, Vallone DM. 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;2012:679134. doi: 10.1155/2012/679134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rhodes DJ, Radecki Breitkopf C, Ziegenfuss JY, Jenkins SM, Vachon CM. Awareness of breast density and its impact on breast cancer detection and risk. Journal of Clinical Oncology. 2015;33(10):1143–1150. doi: 10.1200/JCO.2014.57.0325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ridner SL, Walker KL, Hart JL, Myers JA. Smoking identities and behavior: evidence of discrepancies, issues for measurement and intervention. Western Journal of Nursing Research. 2010;32(4):434–446. doi: 10.1177/0193945909354904. [DOI] [PubMed] [Google Scholar]
  37. Rose JS, Chassin L, Presson C, Sherman SJ, Stein MD, Col N. A latent class typology of young women smokers. Addiction. 2007;102(8):1310–1319. doi: 10.1111/j.1360-0443.2007.01889.x. [DOI] [PubMed] [Google Scholar]
  38. Schane RE, Glantz SA, Ling PM. Nondaily and social smoking: an increasingly prevalent pattern. Archives of Internal Medicine. 2009a;169(19):1742–1744. doi: 10.1001/archinternmed.2009.315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Schane RE, Glantz SA, Ling PM. Social smoking implications for public health, clinical practice, and intervention research. Am J Prev Med. 2009b;37(2):124–131. doi: 10.1016/j.amepre.2009.03.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Schane RE, Ling PM, Glantz SA. Health effects of light and intermittent smoking: a review. Circulation. 2010;121(13):1518–1522. doi: 10.1161/CIRCULATIONAHA.109.904235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Sharma E, Beck KH, Clark PI. Social context of smoking hookah among college students: scale development and validation. Journal of American College Health. 2013;61(4):204–211. doi: 10.1080/07448481.2013.787621. [DOI] [PubMed] [Google Scholar]
  42. Sharma E, Clark PI, Sharp KE. Understanding psychosocial aspects of waterpipe smoking among college students. American Journal of Health Behavior. 2014;38(3):440–447. doi: 10.5993/ajhb.38.3.13. org/10.5993/AJHB.38.3.13. [DOI] [PubMed] [Google Scholar]
  43. Shiffman S, Dunbar MS, Ferguson SG. Stimulus control in intermittent and daily smokers. Psychology of Addictive Behaviors. 2015;29(4):847–855. doi: 10.1037/adb0000052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Shiffman S, Dunbar MS, Li X, Scholl SM, Tindle HA, Anderson SJ, Ferguson SG. Smoking patterns and stimulus control in intermittent and daily smokers. PLoS One. 2014;9(3):e89911. doi: 10.1371/journal.pone.0089911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Shiffman S, Kirchner TR, Ferguson SG, Scharf DM. Patterns of intermittent smoking: An analysis using Ecological Momentary Assessment. Addictive Behaviors. 2009;34(6–7):514–519. doi: 10.1016/j.addbeh.2009.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Shiffman S, Li X, Dunbar MS, Ferguson SG, Tindle HA, Scholl SM. Social smoking among intermittent smokers. Drug and Alcohol Dependence. 2015;154:184–191. doi: 10.1016/j.drugalcdep.2015.06.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Solomon M. 2015 Is The Year Of The Millennial Customer: 5 Key Traits These 80 Million Consumers Share. 2014 Dec 29; Retrieved from http://www.forbes.com/sites/micahsolomon/2014/12/29/5-traits-that-define-the-80-million-millennial-customers-coming-your-way/#47a929622a81.
  48. Song AV, Ling PM. Social smoking among young adults: investigation of intentions and attempts to quit. Am J Public Health. 2011;101(7):1291–1296. doi: 10.2105/AJPH.2010.300012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Sutfin EL, McCoy TP, Berg CJ, Champion H, Helme DW, O’Brien MC, Wolfson M. Tobacco use by college students: a comparison of daily and nondaily smokers. American Journal of Health Behavior. 2012;36(2):218–229. doi: 10.5993/AJHB.36.2.7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Sutfin EL, Reboussin BA, McCoy TP, Wolfson M. Are college student smokers really a homogeneous group? a latent class analysis of college student smokers. Nicotine & Tobacco Research. 2009;11(4):444–454. doi: 10.1093/ntr/ntp006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. The American Association for Public Opinion Research. Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys. 8 2015. [Google Scholar]
  52. Thrul J, Buhler A, Ferguson SG. Situational and mood factors associated with smoking in young adult light and heavy smokers. Drug and Alcohol Review. 2014;33(4):420–427. doi: 10.1111/dar.12164. [DOI] [PubMed] [Google Scholar]
  53. U.S. Department of Health and Human Services. The Health Consequences of Smoking—50 Years of Progress. A Report of the Surgeon General. Atlanta, GA: 2014. [Google Scholar]
  54. Villanti AC, Pearson JL, Cantrell J, Vallone DM, Rath JM. Patterns of combustible tobacco use in U.S. young adults and potential response to graphic cigarette health warning labels. Addictive Behaviors. 2015;42:119–125. doi: 10.1016/j.addbeh.2014.11.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Villanti AC, Rath JM, Vallone DM. Quit Behaviors in Young Adults: Comparing “smokers” and “social smokers”. Paper presented at the 2012 SRNT 18th Annual Meeting; Houston, TX. 2012. [Google Scholar]
  56. White A, Chan GC, Quek LH, Connor JP, Saunders JB, Baker P, … Kelly AB. The topography of multiple drug use among adolescent Australians: findings from the National Drug Strategy Household Survey. Addictive Behaviors. 2013;38(4):2068–2073. doi: 10.1016/j.addbeh.2013.01.001. [DOI] [PubMed] [Google Scholar]
  57. White HR, Bray BC, Fleming CB, Catalano RF. Transitions into and out of light and intermittent smoking during emerging adulthood. Nicotine & Tobacco Research. 2009;11(2):211–219. doi: 10.1093/ntr/ntn017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wortley PM, Husten CG, Trosclair A, Chrismon J, Pederson LL. Nondaily smokers: a descriptive analysis. Nicotine & Tobacco Research. 2003;5(5):755–759. doi: 10.1080/1462220031000158753. [DOI] [PubMed] [Google Scholar]
  59. Yeager DS, Krosnick JA, Chang L, Javitz HS, Levendusky MS, Simpser A, Wang R. Comparing the accuracy of RDD telephone surveys and internet surveys conducted with probability and non-probability samples. Public Opinion Quarterly. 2011:nfr020. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS853427-supplement.docx (13.9KB, docx)

RESOURCES