What Influences Demand for Cigars among African American Adult Cigar Smokers? Results from a hypothetical purchase task

Abstract

African Americans (AA) have historically been targeted by the tobacco industry and have the highest rates of current cigar use among racial/ethnic groups in the US. Yet, there is limited evidence on other factors influencing cigar use. Amongst a sample of 78 AA current cigar (any type) smokers, log-linear regression models examined correlates of cigar demand obtained from a validated behavioral economic purchase task. Mean intensity, or cigar demand when free, was 6.68 cigars (standard deviation [SD]: 8.17), while mean breakpoint, or the highest price a participant was willing to pay, was $4.62 (SD: 3.88). Mean maximum daily expenditure, Omax was $15.20 (SD: 25.73) and Pmax, the price at Omax was $5.25 (SD: 3.95). Participants aged 21 to 30 years compared to those aged 18 to 20 years, those with higher levels of dependence, and females compared to males, had a significantly higher intensity. Participants with cannabis use above the sample median in the last 30 days (4+ days) had significantly higher intensity and Omax than those below the median. Further, participants with a high school education or more had a significantly lower intensity, breakpoint, and Omax than those with less than high school education. Individuals with income below the federal poverty line also had a significantly lower breakpoint and Omax than those above. Finally, tobacco harm perceptions were inversely associated with Pmax. Stricter policies on cigar products, such as higher taxes and product-specific harm messaging, may have an immediate and sustained impact on health disparities related to cigar use.

Introduction

Cigars are rolled tobacco wrapped in tobacco leaves or other tobacco-containing material (Center for Disease Control and Prevention [CDC], 2020). An estimated 12 million adults in the US are current cigar smokers, a figure that has remained steady in recent years (Substance Abuse and Menatal Health Service [SAMHSA], 2015, 2017, 2018a). There are three major types of cigars sold in the US: large cigars, cigarillos (short, narrow cigars, generally without filters and frequently flavored), and little cigars (about the same size as a cigarette and generally have a filter; CDC, 2020). African Americans (AA) have the highest rates of current cigar use compared to other racial/ethnic groups in the US, with cigarillos the most common type of cigars used by AA cigar smokers (Chen-Sankey et al., 2021; SAMHSA, 2018a). Approximately 8% of AA adults report current cigar use compared to only 4.8% Non-Hispanic (NH) Whites and a national average of 4.9% (SAMHSA, 2018a, 2018b). The misperception that cigars are less harmful than cigarettes has contributed to their prevalence (Chen-Sankey et al., 2019; Kong et al., 2020). Yet, cigar use has the same significant health risks as cigarette smoking (National Cancer Institute, 1998), and some cigar products contain more nicotine and tobacco-related toxins than cigarettes (Chen-Sankey et al., 2019; Pickworth, Rosenberry, & Koszowski, 2017).

Despite having lower rates of overall tobacco use, AA have higher rates of cigar use and are more likely to die from tobacco-related diseases, such as heart disease, stroke and diabetes, when compared to NH-Whites (Carroll et al., 2011; CDC, 2019; Rosario & Harris, 2020; SAMHSA, 2018b). Further, AA are more likely to be poly-tobacco users (Kong et al., 2020; Mead et al., 2018) and co-use cigarillos and cannabis more frequently than using tobacco alone (Montgomery, 2015). These co-use patterns may compound the negative consequences of cigar smoking among AA and contribute to disproportionate disease burden (Cohn, Cobb, Niaura, & Richardson, 2015; Kypriotakis, Robinson, Green, & Cinciripini, 2018).

Cigars are available in a variety of flavors, such as menthol, fruit, or candy, which increase their appeal among AA cigar smokers (Nasim et al., 2016; Sterling, Fryer, Pagano, Jones, & Fagan, 2016). One study found that 38% of young adult AA believed that plain cigars were more addictive than flavored cigars (Chen-Sankey et al., 2019). Areas with large racial/ethnic minority populations usually have more tobacco retail outlets (CDC, 2019) and tend to have greater availability of flavored cigars (Ribisl et al., 2017). Other cigar-specific features that increase their appeal include their availability in “singles” and lower price point, particularly for cigarillos, compared to many other tobacco products (Delnevo, Hrywna, Giovenco, & Lo, 2017; Nonnemaker, Rostron, Hall, MacMonegle, & Apelberg, 2014; Stephens, Ogunsanya, Ford, Bamgbade, & Liang, 2015).

The public health inequities of cigar smoking, and the evolving regulatory environment, suggest an urgent need for scientific investigation into the use and appeal of cigar products among AA. Understanding which user characteristics are associated with greater abuse liability for cigars is essential for developing policies, regulations, and product standards that promote public health. Abuse liability refers to the likelihood of non-medical drug use leading to dependence and being influenced by characteristics of the drug or drug delivery system itself and that of the individual, social and policy environment (Bono et al., 2020). Demand indices from hypothetical purchase tasks are associated with abuse liability (Zvorsky et al., 2019). However, no studies to our knowledge have examined which individual and product characteristics are associated with cigar demand amongst a sample of AA cigar smokers. The current study addresses this critical gap by examining demand for cigars among a community sample of adult AA current cigar smokers using a validated hypothetical purchase task and assessing the role of socioeconomic status, perceptions of tobacco products, cannabis use, flavored tobacco product use, and other factors associated with demand.

Method

Data were collected from a community survey of adults (18 to 55 years) who were current cigar smokers (any type of cigar product in the past 30 days) or non-tobacco users (no use of any tobacco/nicotine product in the past 6 months) within and nearby low-income communities in Richmond, Virginia. Exclusion criteria included inability to read/write in English, serious medical condition in the past 12 months, eye, thoracic or abdominal surgery in the past 6 weeks, and any current symptoms/illness that might interfere with physiological measures assessment. Following in-person informed consent and screening measures, eligible participants were asked to perform a lung function test, provide a saliva sample, and complete additional measures via an online survey regarding their sociodemographic, tobacco and other substance use history. Compensation was $50 cash and choice of a $30-valued health-related item. A hypothetical cigar purchase task (CPT) was administered to cigar users only. Purchase tasks are used to understand how an individual values a tobacco product e.g. Reed et al. (2020) and have been validated across a number of studies for cigarette use (Grace, Kivell, & Laugesen, 2015; Mackillop et al., 2016; Wilson, Franck, Koffarnus, & Bickel, 2016).

Similar to other purchase tasks, e.g. Reed et al. (2020), participants were presented with the following scenario: “Imagine a typical day during which you smoke cigars. The following questions ask how many cigars of your favorite brand you would buy if they cost various amounts of money. The only available cigars are your favorite brand. Assume that you have the same income/savings that you have now and no access to any cigars or nicotine products other than those offered at these prices. In addition, assume that you would smoke the cigars that you request on that day; that is, you cannot save or stockpile cigars for a later date. Please respond to these questions honestly.” The first price a participant is asked has a price of $0: “If your favorite brand cost $0 (free) each. How many cigars would you smoke in one day?” The subsequent questions asked the number of cigars smoked at a price of $0.01, $0.02, $0.04, $0.08, $0.16, $0.32, $0.64, $1.28, $2.56, $3.84, $5.12, $6.40, $7.68, $8.96, and $10.24.

As is typical of the literature using purchase tasks, several data quality checks were applied (Barnes, Bono, Lester, Eissenberg, & Cobb, 2017; Stein, Koffarnus, Snider, Quisenberry, & Bickel, 2015). The theory of demand postulates that demand for a product decreases as its price increases (Rice, Thomas; Unruh, 2009). Therefore, the overarching criterion to ensure CPT responses were valid was whether cigar consumption decreased as prices increased. The sample was first restricted to AA current cigar smokers (111 out of 213 study completers). We identified nonsystematic data using the criteria that there was an increase in consumption from the lowest to the highest price (Barnes et al., 2017; Stein et al., 2015). Four participants were excluded as they either had all missing data for the CPT or did not self-identify as AA. Among those remaining, responses for participants with three or fewer instances of nonsystematic data (n=24) were imputed (see Appendix). Participants with four or more nonsystematic data (n=29) were excluded from the main analyses, but their potential influence was examined in a sensitivity analysis (see Appendix). Multiple imputation was used for missing covariates resulting in an analytic sample of 78 AA current cigar smokers. Our sample comprised 19% current traditional cigar smokers (smoked on average 5.6 days (standard deviation [SD]: 7.13) and 3.0 cigars/day on days they smoked (SD: 4.99)) and 95% current cigarillos/filtered cigar smokers (smoked on 14.4 days (SD: 11.4) and 5.8 cigars/day on days they smoked (SD:9.01)). This study was approved by the Institutional Review Board at the principal investigators’ university (HM20011338).

Outcomes

Four continuous cigar demand indices were recovered from the CPT (Mackillop et al., 2016; Strickland, Campbell, Lile, & Stoops, 2020) including: i) intensity - a measure of cigar demand when the price is $0 (free); ii) breakpoint - a measure of the highest price participants were willing to pay for cigars; iii) observed Omax - a measure for the maximum daily expenditure on cigars; and iv) observed Pmax - a measure for the price associated with Omax. Observed Omax and observed Pmax are herein referred to as “Omax” and “Pmax” respectively.

Covariates

Covariates previously found to be correlated with cigar use were included in our analytic models. Demographic characteristics included age (Mojtaba, Ahmed, & Peeran, 2016; Montgomery, 2015) (18–20 years, less than 21 years, 21–30 years, and 31 and more years) and sex (Cohn et al., 2015; Nonnemaker et al., 2014) (female vs. male). Socioeconomic status (SES) included education (high school diploma/GED or more vs. less than high school diploma/GED) (Montgomery, 2015; Stephens et al., 2015), employment status (employed vs. not) (Cohn et al., 2015; Montgomery, 2015), marital status (married/cohabiting vs. not) (Montgomery, 2015) and poverty status (below the 2019 federal poverty line (FPL) vs. not) (ASPE, 2020; Cohn et al., 2015; Mojtaba et al., 2016).

We included a measure for participants’ belief that tobacco products are harmful (Cohn et al., 2015; Stephens et al., 2015) based on responses to how harmful they thought non-menthol cigarettes, menthol cigarettes, traditional cigars, cigarillos/filtered cigars, and electronic cigarettes were. Participants rated how harmful each of these products were to health as “not at all harmful” (1) to “extremely harmful” (5). We added the responses across these questions to create a continuous scale of 0–20 (higher numbers represent a higher belief that tobacco products are harmful).

Substance use measures included alcohol use (Cohn et al., 2015; Mojtaba et al., 2016) and cannabis use (Montgomery, 2015). Both alcohol and cannabis use were included in the analysis as a binary variable for whether the frequency of past 30-day use was at or above vs. below the median number of days (+4 days) of use for that product. Finally, a binary measure (yes, no) of flavored past 30-day cigar use (mint/wintergreen/menthol, fruit, coffee drinks, candy, desserts, spices, alcohol drinks, or any other non-tobacco flavor) was included (Chen-Sankey et al., 2019). We include a measure to approximate cigar dependence (Zvorsky et al., 2019). The survey included one item adapted from the Fagerstrom Test for Nicotine Dependence (FTND) dependence measure (National Institute for Drug Abuse [NIDA], 2014). Participants were asked “how soon after you wake up do you smoke your first cigar?” Based on their response as either within 5 minutes, 6–30 minutes, 31–60 minutes or after 60 minutes, we created proxy measure of dependence ranging from 0–3 (higher numbers represent higher dependence)

Analysis

Sample characteristics were first described. Due to skewness, we assigned a small value (0.01) to the demand indices and log-transformed them to better approximate a normal distribution. We ran bivariate analysis of each of the four demand indices and each covariate using Pearson’s Correlation. We ran linear regression models on the log-transformed outcomes to examine the factors associated with demand for cigars among AA cigar smokers.

As the CPT has never been used with a community sample of AA cigar smokers, we conducted sensitivity analyses to better understand the participant characteristics that were associated with purchase task comprehension and/or attention to the task. Specifically, using the full sample before excluding the nonsystematic data, we used linear probability models to determine the characteristics associated with having nonsystematic data. We then ran a linear regression on the sample of participants with nonsystematic data only to determine the factors associated with the number of times there was nonsystematic in the CPT. Significance of all analyses was determined at p<0.05. We conducted all analyses in Stata/IC 15.1.

Results

The average intensity, or number of cigars purchased if they were free, was 6.68 cigars (SD: 8.17) and the average breakpoint, or highest price at which cigars were no longer purchased, was $4.62 (SD: $3.88, Table 1). The average Omax, or the maximum daily expenditure, was approximately $15.25 (SD: $25.73), and average Pmax, or price at which demand changes from inelastic to elastic, was $5.25 (SD: $3.95). The sample was predominantly female (63%) and 36% of participants reported having a high school/GED or greater education. The majority of participants were employed (54%), and 71% reported income below the FPL. On average, 40% of participants consumed alcohol above the median in the last 30 days, while about 42% had cannabis use above the median the last 30 days.

Table 1:

Sample Characteristics (n=78)

Variable	Mean (SD) (%)
Intensity (cigars)	6.68 (8.17)
Breakpoint ($)	4.62 (3.88)
Omax ($)	15.20 (25.73)
Pmax ($)	5.25 (3.95)
Age (years)
18 –20	7.7%
21–30	37.2%
31 and more	55.1%
Female	62.8%
High school/GED or more	35.9%
Married/Cohabiting	16.7%
Employed	53.8%
Below federal poverty line	70.5%
Belief tobacco products are harmful	13.38 (4.95)
Flavored cigar use	53.8%
Alcohol current use^	39.7%
Cannabis current use^	42.3%
Dependence&	1.51 (1.20)

Notes:

^{^}Alcohol and cannabis current use measures were defined as above or below the sample median of 4 or more days in the past 30 days for each product.

^&Dependence was proxied for using the following item adapted from the Fagerstrom Test for Nicotine Dependence (FTND) dependence measure, “how soon after you wake up do you smoke your first cigar?” Responses included: within 5 minutes, 6–30 minutes, 31–60 minutes or after 60 minutes, representing a proxy measure of dependence ranging from 03 (higher numbers represent higher dependence).

Unadjusted Correlates of Cigar Demand Indices

Prior to adjustment, there was a significant inverse association between intensity of cigar demand and participants’ belief tobacco products are harmful, but a significant positive association with cannabis use and our proxy measure of cigar dependence (p<0.05 each). Breakpoint and Omax were inversely correlated with educational attainment and belief that tobacco products are harmful, before adjustment for covariates (p<0.05 each). Additionally, Omax was positively associated with cannabis use and cigar dependence (p<0.05 each). Finally, Pmax was inversely associated with beliefs about tobacco product harm and alcohol use (p<0.05 each).

Adjusted Correlates of Intensity

Compared to participants aged 21 years or less, those aged 21–30 years reported they would purchase more cigars when free (intensity) (β==2.27, p<0.05, Table 2). Female participants had a higher intensity than male participants (β=1.08, p<0.05). Participants with high school/GED or more had a lower intensity (β=−1.25, p<0.05) compared to those with less than high school education. Further, participants with cannabis use above the median in the last 30 days had a higher intensity (β=1.34, p<0.01) than those who had past 30-day cannabis use below the median. Participants with a higher cigar dependence had a higher intensity (β=0.30, p<0.05).

Table 2:

Adjusted Correlates of Cigar Abuse Liability among African American Cigar Smokers

	Intensity ln(cigars)#	Breakpoint ln($)	Omax ln($)	Pmax ln($)
Ref: Age (years): 18–20
21–30	2.27 (1.06)*	1.73 (1.11)	1.71(1.27)	−0.53 (0.45)
31 and more	1.67 (1.09)	1.09 (1.15)	1.10 (1.29)	−0.46 (0.53)
Female	1.08 (0.47)*	0.69 (0.52)	0.70 (0.59)	−0.27 (0.29)
High school/GED or more	−1.25 (0.52)*	−1.68 (0.80)*	−2.05 (0.89)*	0.11 (0.41)
Married	0.45 (0.54)	0.26 (0.63)	0.39 (0.68)	−0.59 (0.54)
Employed	0.15 (0.35)	0.14 (0.55)	−0.14 (0.61)	0.42 (0.32)
Below federal poverty line	−0.25 (0.45)	−1.45 (0.63)*	−1.66 (0.70)*	−0.63 (0.42)
Belief tobacco products are harmful	−0.01 (0.04)	−0.05 (0.06)	−0.07 (0.06)	−0.12 (0.04)**
Smoke flavored cigar	0.49 (0.45)	0.06 (0.61)	0.08 (0.66)	−0.49 (0.35)
Alcohol current use^	0.07 (0.39)	0.28 (0.58)	0.21(0.63)	−0.17 (0.33)
Cannabis current use^	1.34 (0.43)**	1.11 (0.58)	1.46 (0.64)*	−0.19 (0.38)
Dependence&	0.30 (0.13)*	0.01 (0.23)	0.13 (0.26)	−0.01 (0.14)
Constant	−2.12 (1.57)	0.29 (1.80)	1.31 (1.98)	4.08 (0.92)**
Observations	78	78	78	78
R-squared	0.36	0.19	0.24	0.23

Notes:

^#Cigar purchase task outcomes were log-transformed to better approximate a normal distribution and then linear regression models using log-transformed outcomes were used fit to the data. Standard errors are shown in parentheses.

^{^}Alcohol and cannabis current use measures were defined as above or below the sample median of 4 or more days in the past 30 days for each product.

^&Dependence was proxied for using the following item adapted from the Fagerstrom Test for Nicotine Dependence (FTND) dependence measure, “how soon after you wake up do you smoke your first cigar?” Responses included: within 5 minutes, 6–30 minutes, 31–60 minutes or after 60 minutes, representing a proxy measure of dependence ranging from 0–3 (higher numbers represent higher dependence).

^*p<0.05

^**p<0.01.

Adjusted Correlates of Breakpoint

Participants with high school/GED or more had a lower breakpoint when compared to those with less than high school education (β=−1.68, p<0.05, Table 2). Conversely, participants living below the FPL had a lower breakpoint than their counterparts who were at or above the FPL (β=1.45, p<0.01).

Adjusted Correlates of Omax

Participants whose education exceeded a high school diploma/GED had a lower Omax (β=−2.05, p<0.05) when compared to participants with less than high school education (Table 2). We found that participants with income below the FPL had a lower Omax (β=−1.66, p<0.01) than participants at or above the FPL. Participants with cannabis use above the median in the last 30 days had a higher Omax (β=1.213eq, p<0.05).

Adjusted Correlates of Pmax

The more participants believed that smoking was harmful the lower their Pmax was (β=−0.12, p<0.01, Table 2).

Sensitivity analyses

Our main adjusted results were not sensitive to including an indicator for other current tobacco product use and other tobacco product use was not significantly associated with any of the four cigar demand indices. Participants with high school/GED education or more were less likely (28 percentage points p<0.05) to have nonsystematic data compared to those with less than a high school education (Appendix, Table B). Those with cannabis use above the median were more likely (25 percentage points, p<0.05) to have nonsystematic data compared to those below.

Discussion

Regulation of cigar products has historically been weaker than cigarettes, with further variation in regulation across cigar product type (e.g., large cigars vs. cigarillos) and the populations who predominantly use them, creating the need to better understand abuse liability for cigars to inform more effective and equitable cigar policies. We examined the correlates of cigar demand among a sample of AA current cigar users, 95% of whom were cigarillo smokers. We found that educational attainment, income, age, sex, cannabis use, and perceived harm from tobacco product use were associated with cigar demand as measured by a validated behavioral economic task (Mackillop et al., 2016).

Our findings that education was associated with several indices of cigar demand is in line with prior research (Montgomery, 2015). Specifically, those with a high school/GED or higher, compared to those not completing high school, purchased fewer cigars when offered for free (intensity), reduced their purchasing to zero at lower prices (breakpoint) and were willing to spend less in total dollars on cigars (Omax). For participants whose income was below the FPL, similar associations were observed for breakpoint and Omax in that income was inversely related to these indices. Although our findings differ from prior work suggesting tobacco use is higher among AA with lower incomes (Montgomery, 2015), other research (Corral, Landrine, Simms, & Bess, 2013) suggests that AA with higher incomes are more likely to smoke cigars filled with cannabis (blunts).

We found that when the price was $0 (intensity), demand was higher among those aged 21–30 years compared to those 18–20 years. Previous research supports a link between age, race/ethnicity, and price sensitivity for tobacco products (Cantrell et al., 2013; Henriksen et al., 2017). Additionally, prior evidence supports that price sensitivity is higher, and brand loyalty is lower for current generations of young adults than in previous generations (Carufel, 2019).

Moreover, demand for cigars was higher among AA cigar smokers who used cannabis four or more days in the past month (our sample median), compared to those who used it less frequently or not at all in the past month. Given that the relationship between cigars and cannabis is widely known, particularly for AA who use cannabis with cigars to smoke blunts (Persoskie et al., 2019), our study examined cannabis use regardless of whether it was used when replacing the tobacco in cigars and found further support that cannabis use was positively associated with cigar demand (Collins, Vincent, Yu, Liu, & Epstein, 2014). The evidence of complementarity of demand for cigars and cannabis suggests that regulatory policies aimed at cigar product prices may also influence the patterns of cannabis use among AA.

Additionally, we found a negative association with cigar demand for current cigar users who believed tobacco products were more harmful, as evidenced by the lower Pmax they reported. Our findings were robust when we tested this association using a measure of perceived harm specific to cigars. Tobacco harm messaging interventions have been found to be effective in reducing the number of cigarettes smoked by adolescents (Mason et al., 2015). However, further research is needed to understand tobacco control messages that are effective in shaping tobacco-related harm perceptions among AA adults who use cigar products. Finally, we found that higher cigar dependence was associated with a higher intensity. This finding is aligned to prior literature that show strong association of product dependence and indices from purchase tasks (Zvorsky et al., 2019).

Continued investigation of cigar demand among AA may help address health disparities related to cigar use in the US. Our findings add to the literature on the value of hypothetical purchase tasks by being the first to examine the correlates of cigar demand among a population disproportionately harmed by cigar products. Nonetheless, the study is subject to limitations that may limit generalizability. Our sample was likely impacted by self-selection, as are all convenience samples, as participants may have been motivated to participate based on factors, such as exposure to study advertising, interest in the study topic, constraints on being able to participate, and monetary incentives offered. Participants were recruited from low-income communities who might be more sensitive to tobacco price change than other cigar smokers. Further, data used in this research came from a community sample of AA current cigar users in one mid-Atlantic metropolitan area. AA are not a monolith and the social construction of race may differ across the US. Thus, the extent to which specific characteristics of AA cigar smokers are associated with cigar demand is person- and context-specific. Nonetheless, our work provides initial evidence that should be repeated using larger samples and scope for a richer understanding of the demand for cigar products to better inform more equitable tobacco regulatory policies. Future studies could also use similar purchase tasks to compare cigar abuse liability amongst other racial/ethnic groups as well as assess the abuse liability of other cigar product characteristics among AA cigar smokers, notably flavors (Bono et al., 2020; Reed et al., 2020; Zvorsky et al., 2019).

Our findings suggest that stricter price (e.g., higher taxes, restrictions on price promotion) and non-price policies (e.g., product- and/or community-specific harm messaging) on cigar products, specifically cigarillos, may have an immediate and sustained impact on health disparities related to cigar use. Such regulations might also address the well-researched marketing and advertisement environment in which tobacco products are introduced to AA by the tobacco industry and that mold perceptions of product use (Cantrell et al., 2013; Garner, Brock, & Seth, 2019; Richardson, Ganz, & Vallone, 2014).

Public Health Significance Statement.

African Americans use cigars at nearly twice the rate of Non-Hispanic Whites and the national average. Cigar products have been the target of fewer regulations than those imposed on cigarettes, creating the need to better understand abuse liability for cigars to inform more effective and equitable cigar policies. Findings suggest that educational attainment, income, age, sex, cannabis use, dependence, and perceived harm from tobacco product use are associated with cigar demand as measured by a validated behavioral economic task.

Appendix

We logically imputed cigar purchase task data for participants with three or less instances of disordered data (n=24) as follows: (i) if a response for consumption in the CPT was missing for a particular price and the responses before and after the missing value were the same, then the same value was assigned where the missing value was, (ii) if the response missing in the CPT was between two responses that were observed but were not the same, the higher value was assigned, and (iii) if the missing response was the first price on the CPT, the second response on the CPT was used to replace the missing value.

Table A:

Summary statistics stratified by disordered data

	Not Disordered (n=46) Mean (SD) (%)	Disordered (n=60) Mean (SD) (%)
Age (years)
Less than 21	4.4%	18.3%
21–30	37.0%	41.7%
31 and more	58.7%	40.0%
Female	65.2%	60.0%
High school/GED or more	47.8%	21.7%
Marri ed/C ohabiting	19.6%	21.7%
Employed	52.2%	48.3%
Below federal poverty line	60.9%	83.3%
Belief tobacco products are harmful	13.96 (4.70)	12.18 (4.73)
Flavored cigar use	58.7%	40.0%
Alcohol current use^	54.3%	46.7%
Cannabis current use^	34.8%	58.3%
Dependence&	1.38 (1.25)	1.59 (1.11)

Notes:

^{^}Alcohol and cannabis current use measures were defined as above or below the sample median of 4 or more days in the past 30 days for each product.

^&Dependence was proxied for using the following item adapted from the Fagerstrom Test for Nicotine Dependence (FTND) dependence measure, “how soon after you wake up do you smoke your first cigar?” Responses included: within 5 minutes, 6–30 minutes, 31–60 minutes or after 60 minutes, representing a proxy measure of dependence ranging from 03 (higher numbers represent higher dependence).

^*p<0.05

^**p<0.01.

Table B:

Factors associated with having disordered data

	Any Disordered	Number of Disordered
Age (years): ref: 18–20
21–30	−0.17 (0.13)	−0.517 (0.85)
31 and more	−0.19 (0.15)	−0.74 (1.02)
Female	−0.01 (0.10)	0.20 (0.61)
High school/GED or more	−0.28 (0.14) *	−0.18 (0.66)
Marri ed/C ohabiting	0.16 (0.11)	1.21 (0.61) *
Employed	0.03 (0.09)	−0.65 (0.58)
Below federal poverty line	0.13 (0.14)	−0.03 (0.81)
Belief tobacco products are harmful	−0.01 (0.01)	−0.08 (0.06)
Flavored cigar use	−0.09 (0.10)	−0.51 (0.66)
Alcohol current use^	−0.00 (0.10)	−0.36 (0.60)
Cannabis current use^	0.25 (0.10) *	0.27 (0.68)
Dependence&	0.02 (0.05)	−0.26 (0.26)
Constant	0.64 (0.26) **	4.72 (1.22) **
Observations	106	60
R-squared	0.22	0.23

Notes: Standard errors in parentheses.

^{^}Alcohol and cannabis current use measures were defined as above or below the sample median of 4 or more days in the past 30 days for each product.

^&Dependence was proxied for using the following item adapted from the Fagerstrom Test for Nicotine Dependence (FTND) dependence measure, “how soon after you wake up do you smoke your first cigar?” Responses included: within 5 minutes, 6–30 minutes, 31–60 minutes or after 60 minutes, representing a proxy measure of dependence ranging from 0–3 (higher numbers represent higher dependence).

^*p<0.05;

^**p<0.01.