Tobacco outlet density, retailer cigarette sales without ID checks and enforcement of underage tobacco laws: associations with youths’ cigarette smoking and beliefs
Abstract
Aims
To estimate the relationships of tobacco outlet density, cigarette sales without ID checks and local enforcement of underage tobacco laws with youth's life-time cigarette smoking, perceived availability of tobacco and perceived enforcement of underage tobacco laws and changes over time.
Design
The study involved: (a) three annual telephone surveys, (b) two annual purchase surveys in 2000 tobacco outlets and (c) interviews with key informants from local law enforcement agencies. Analyses were multi-level models (city, individual, time).
Setting
A sample of 50 mid-sized non-contiguous cities in California, USA.
Participants
A total of 1478 youths (aged 13–16 at wave 1, 52.2% male); 1061 participated in all waves.
Measurements
Measures at the individual level included life-time cigarette smoking, perceived availability and perceived enforcement. City-level measures included tobacco outlet density, cigarette sales without ID checks and compliance checks.
Findings
Outlet density was associated positively with life-time smoking [OR = 1.12, P < 0.01]. An interaction between outlet density and wave (OR = 0.96, P < 0.05) suggested that higher density was associated more closely with life-time smoking at the earlier waves when respondents were younger. Greater density was associated positively with perceived availability (β = 0.02, P < 0.05) and negatively with perceived enforcement (β = −0.02, P < 0.01). Sales rate without checking IDs was related to greater perceived availability (β = 0.01, P < 0.01) and less perceived enforcement (β = −0.01, P < 0.01). Enforcement of underage tobacco laws was related positively to perceived enforcement (β = 0.06, P < 0.05).
Conclusions
Higher tobacco outlet density may contribute to life-time smoking among youths. Density, sales without ID checks and enforcement levels may influence beliefs about access to cigarettes and enforcement of underage tobacco sales laws.
Introduction
Smoking in adolescence is related to a range of short- and long-term health problems, including a higher risk for tobacco dependence in adulthood, increased number and severity of respiratory illnesses, decreased fitness and potential retardation of lung growth 1. In the United States in 2013, 16.3% of high school seniors (aged ≈ 17–18 years), 9.1% of 10th graders (aged ≈ 15–16 years) and 4.5% of 8th graders (aged ≈ 13–14 years) reported smoking cigarettes in the past 30 days 2. Although these rates are well below the peak smoking levels reported in the 1990s, smoking by young people remains a serious public health issue. Because nearly two-thirds of adult smokers begin daily smoking before age 18 3-5, it is imperative to develop effective interventions to prevent and reduce smoking by young people.
Environmental approaches to reduce youths’ access to tobacco, such as controlling outlet density, improving retailer compliance with underage sales laws and enforcing underage tobacco laws, may be important strategies for preventing and reducing smoking by young people 6. The overall goal of this paper is to investigate the longitudinal relationships of these community-level factors with youths’ cigarette smoking and beliefs. A number of cross-sectional studies have examined the associations between tobacco outlet density and youths’ cigarette smoking. Results of these studies, however, are mixed. Some studies have shown positive associations 7-12, whereas others have found no or small relationships 13-17. In addition, no longitudinal studies have investigated if tobacco outlet density is related to youths’ smoking over time.
Reducing sales of tobacco products to underage youths is a common approach of numerous community interventions targeting adolescents’ smoking. Many studies have found that enforcement of underage sales laws through compliance checks is associated with decreased rates of tobacco sales to youths 18, 19. Similarly, programs that reinforce retailers for not selling tobacco products to minors have been shown to be effective in increasing ID checking rates 20. The effect of reduced underage sales on youths’ cigarette smoking, however, remains inconclusive. Some studies have found that higher compliance rates were related to lower levels of youths’ smoking 21-24. Others have reported no relationship 25-27.
Although a key purpose of policies targeting youths’ access to tobacco is to reduce smoking, it is important to also consider if such policies are associated with adolescents’ tobacco-related beliefs. A central assumption of cognitive social-learning approaches to understanding behavior is that human actions, including smoking, are the result of an individual's beliefs about the relative costs and benefits associated with them 28, 29. Within this framework, costs include not only anticipated social and physical consequences (e.g. getting into trouble with police or parents) and economic costs (i.e. price), but also the opportunity costs (e.g. effort) associated with a behavior. From this perspective, adolescents’ beliefs relating to the ease of access to tobacco products and to enforcement of underage tobacco laws may be particularly important in their decision to smoke. To date, there are conflicting study results linking tobacco compliance checks to decreases in perceived access to tobacco 30. Moreover, few studies, if any, have investigated whether outlet density and levels of enforcement are related to such beliefs.
Going beyond previous research, the goals of this longitudinal study are to examine the relations of retail access measures, including tobacco outlet density, cigarette sales without ID checks and local enforcement of underage tobacco laws with youths’: (a) life-time cigarette smoking; (b) perceived ease of access to tobacco; and (c) perceived enforcement of underage tobacco laws. We examined interactions between the retail access measures and time to assess changes in these relationships over time.
Methods
Overview
The study involved three data collection activities undertaken in a geographically representative sample of 50 mid-sized non-contiguous California communities: (a) a three-wave annual longitudinal telephone survey of adolescents (aged 13–16 at wave 1); (b) purchase surveys to assess rates of cigarette sales without ID checks; and (c) interviews with representatives from local law enforcement agencies to determine enforcement of underage tobacco sales laws. To determine tobacco outlet density, we obtained data about the number of licensed tobacco retail outlets in each city. Institutional review board approval was obtained prior to study implementation.
Sample of cities
A geographical sampling method was used to select non-contiguous 50 California cities with populations between 50 000 and 500 000. Specifically, a list of all such cities in California was generated. A city was then selected randomly for inclusion and all contiguous cities and cities adjacent to contiguous cites were eliminated. A second city from among those remaining on the list was then selected randomly and contiguous and adjacent cities were eliminated. This process was repeated until a geographically diverse sample of 50 mid-sized cities was obtained. This sampling procedure is described elsewhere in more detail 9, 31.
Survey
Survey sample
The survey targeted youths aged 13–16 at wave 1 in the 50 communities. Households within each city were sampled from a purchased list of telephone numbers and addresses. At each wave, an invitation letter describing the study and inviting participation was mailed to sampled households, followed by a telephone contact. Interviewers obtained parental consent for the interviews followed by assent from the youth respondents. Respondents received $25 at waves 1 and 2 and $35 at wave 3 as compensation for their participation in the study.
Of 3062 sampled households with eligible respondents, 1543 (50.4%) participated in the first interview (wave 1). Of these, 1312 participated in the second telephone interview (wave 2) 1 year later (85% follow-up) and 1121 also participated in the third telephone interview (wave 3) 2 years after wave 1 (72% follow-up from wave 1).
Survey methods
Youths were surveyed using a computer-assisted telephone interview (CATI). The interviews were given in either English or Spanish at the respondent's request and lasted approximately 30–40 minutes. Wave 1 took place in 2009. The current study is based on data from 1478 youths (52.2% male, mean age at wave 1 = 14.6 years, standard deviation (SD) = 1.0) who (a) participated in at least one wave of data collection; (b) lived in the same city across the study years; and (c) provided complete data for all non-varying demographic measures (i.e. age at wave 1, gender, race and ethnicity). Of the 1478 youth participants, 1061 (71.8%) responded to all three waves, 187 (12.7%) responded to two waves and 230 (15.6%) responded to one wave only. An average of 29.6 youths (range: 20–47, SD = 5.97) in each city provided data used for the analyses reported here.
Survey measures
Life-time cigarette smoking
Respondents to the survey were asked if they had ever smoked a whole cigarette in their life, more than just a few puffs (‘no’/‘yes’).
Perceived availability of cigarettes
Perceived availability of cigarettes was ascertained by asking: ‘If you wanted to get cigarettes how easy or difficult would that be for you?’. Possible response options were ‘very easy’, ‘somewhat easy’, ‘somewhat difficult’ and ‘very difficult’. This item was coded so a higher value represented greater perceived availability of cigarettes.
Perceived enforcement of underage tobacco laws
Three questions were used to measure perceived enforcement of underage tobacco laws. Respondents were asked to indicate how likely they thought it was that a person their age would: (a) get into trouble with the police for having cigarettes; (b) get into trouble with the police if he or she were smoking in public; and (c) get into trouble if he or she tried to buy cigarettes at a store. Response options were: ‘very unlikely’, ‘somewhat unlikely’, ‘somewhat likely’ and ‘very likely’. A mean score was computed, with a higher score indicating greater perceived enforcement. The internal reliability (Cronbach's α) for this scale ranged from 0.76 to 0.90 across the three survey waves.
Demographics
Youth respondents were asked their gender, age, race and ethnicity. Because too few participants identified with a specific racial group other than white to allow for meaningful analysis, race and ethnicity were treated as dichotomous variables (white versus non-white and Hispanic versus non-Hispanic). Age at wave 1 was treated as a continuous variable. Because respondents who were aged 16 years at wave 1 were of legal age to purchase tobacco in California at wave 3 (aged 18 years), we included a dummy variable indicating whether the respondent was aged 18 years at each wave. All respondents had a zero value at waves 1 and 2 and approximately 20% of the sample were aged 18 years at wave 3 and were given a value of 1 for that wave.
Tobacco purchase surveys
Rates of cigarette sales without ID checks were assessed with purchase surveys conducted in 2011 and 2012 by eight confederate buyers (four males and four females), who were aged more than 18, but judged to appear younger by independent panels. At each outlet, a single confederate buyer attempted to purchase a pack of cigarettes. If asked about their age, they stated that they were more 18 years old and if asked for age identification they indicated they had none. If a sale was refused, the buyers left without attempting to pressure the clerk. After leaving the outlet, the buyer documented whether the sale was made. Purchase attempts were made at 1000 randomly selected tobacco outlets, 20 per community, each year of the purchase survey for a total of 2000 attempts 31. To obtain a more reliable estimate of sales rates without checking IDs, the overall percentage of outlets that sold a pack of cigarettes in each city across the two waves was calculated 32
Enforcement interviews
Between 2010 and 2012, three annual telephone interviews were conducted with key informants from police departments in each of the 50 cities. The interviews asked about enforcement activities focusing on reducing youths’ access to tobacco. Representatives from police departments in 45 of the 50 communities (90%) participated in the first interview, 47 (94%) participated in the second interview and 39 (78%) in the third interview. A representative from all the 50 police departments participated in at least one interview. Respondents were asked to estimate the number of retail compliance checks their agency conducted in the past year to enforce underage tobacco sales laws. The mean number of compliance checks the agency conducted served as a measure of local enforcement of underage tobacco sales laws. Because this variable was skewed, with 44% of the cities not conducting any compliance checks, it was coded as a dichotomous variable (none versus any compliance checks).
Tobacco outlet density
The total number of licensed tobacco retail outlets in each city was obtained from State of California Board of Equalization data files for September 2011. These data files included the total number of licensed tobacco outlets by city and ZIP-code, but not outlet names or addresses. Outlet density in each city was calculated as the number of outlets per 10000 people.
City demographics
Measures of city demographics were obtained from 2010 GeoLytics data 33. A community-level socio-economic status (SES) indicator was derived from median household income, percentage of the population with a college education and percentage of the population unemployed. The correlations among these measures ranged from 0.52 to 0.79. A principal components analysis yielded a single-factor solution, accounting for 75.1% of the variance (factor loadings range: 0.78–0.91). A factor score was used to represent community SES. Additional city demographics included population density (i.e. population per square mile), percentage of population aged less than 18 years, percentage white, percentage African American and percentage Hispanic. These measures were transformed to Z-scores.
Data analysis
We conducted three-level multi-level random-effects logistic and linear regression analyses with HLM version 7.0 software 34. Importantly, HLM accommodates missing data points in longitudinal data, allowing the inclusion of all cases in the analyses. Spatial autocorrelation was not a concern in analyses, as the 50 cities were chosen specifically to be non-contiguous. We utilized three models to investigate associations between the retail access measures and each of the outcomes: (1) life-time cigarette smoking; (2) perceived availability of cigarettes; and (3) perceived enforcement of underage tobacco laws. The intraclass correlation coefficients (ICCs) indicated that 0.05, 0.08 and 1.03% in the variability in life-time cigarette smoking, perceived availability of cigarettes and perceived enforcement of underage tobacco policies, respectively, lies between cities. These city-level effects are relatively small, but indicate the importance of including cities as a random third-level unit.
In each model, outlet density, rates of tobacco sales without an ID check, enforcement levels and city demographics were included as city-level variables (level 3). Gender, age at wave 1, race and ethnicity were included as individual-level variables (level 2). A time variable (i.e. survey wave) and the dummy variable indicating whether a respondent was aged 18 years were included at the observation level (level 1). Intercept coefficients at levels 2 and 3 were allowed to vary and a random-effects term for residual variance was included at level 1. To examine whether the retail access measures were associated with life-time smoking and beliefs over time, cross-level interactions between the retail access measures and time were included. The interaction terms were dropped from the model if not statistically significant. Variables in each model were entered simultaneously.
The number of communities and the survey sample size were based on power analyses which took into account the nested design assuming an ICC of 0.10 and an 82% retention rate between each survey wave. The sampling frame provided 80% power to detect small to moderate effects at the city (δ = 0.31) and individual (δ = 0.12) levels.
Results
Descriptive statistics
Descriptive statistics for the study variables are shown in Table 1. The average city-level rate of cigarette sales without ID checks was 10.2% (SD = 7.8). This rate is somewhat higher than the Synar non-compliance rates reported by the California Department of Public Health in 2011 (5.6%) and 2012 (8.7%) 35. The average number of compliance checks the local enforcement agencies conducted each year ranged from 0 to 8, with 44% of city agencies reporting that they did not conduct any compliance checks. The number of tobacco outlets per 10 000 people in the 50 cities ranged from 3.32 to 19.18, with a mean of 10.3 (SD = 3.3). No statistically significant association was found between outlet density and rates of sales without ID checks (r = 0.10).
| Variables | % or mean (SD) | Range |
|---|---|---|
| City level (n = 50) | ||
| Retail access | ||
| Sales without ID checks | 10.2 (7.8) | 0.0–32.5 |
| Compliance checks (yes/no) | 56 | 0.0–8.0 |
| Tobacco outlet density (per 10 000 people) | 10.3 (3.3) | 3.3–19.2 |
| City demographics | ||
| Population density | 4870.1 (3347.5) | 1337.2–22 330.2 |
| Percentage age < 18 years | 23.7 (3.2) | 17.0–30.0 |
| Socio-economic statusa | 0.0 (1.0) | −1.7 to 1.7 |
| Percentage white | 79.2 (14.5) | 33.5–97.9 |
| Percentage African American | 5.3 (6.3) | 0.6–33.5 |
| Percentage Hispanic | 34.2 (20.2) | 8.2–97.4 |
| Individual level (n = 1478) | ||
| Age | 14.6 (1.0) | 13.0–16.0 |
| Male | 52.2 | |
| Hispanic | 20.7 | |
| White | 57.8 | |
| Observation level | ||
| Wave 1 (n = 1478) | ||
| Life-time cigarette smoking | 8.9 | |
| Perceived availability of cigarettesb (n = 1466) | 2.3 (1.1) | 1.0–4.0 |
| Perceived enforcement of underage tobacco lawsb (n = 1477) | 3.8 (0.8) | 1.0–4.0 |
| Wave 2 (n = 1248) | ||
| Life-time cigarette smoking | 12.7 | |
| Perceived availability of cigarettesb (n = 1240) | 2.6 (1.1) | 1.0–4.0 |
| Perceived enforcement of underage tobacco lawsb | 2.9 (0.8) | 1.0–4.0 |
| Wave 3 (n = 1061) | ||
| Life-time cigarette smoking | 13.7 | |
| Perceived availability of cigarettesb (n = 1054) | 2.9 (1.1) | 1.0–4.0 |
| Perceived enforcement of underage tobacco lawsb | 2.5 (1.0) | 1.0–4.0 |
- a Socio-economic status measured as a factor score derived from: median household income, percentage of population with a college education and percentage of population unemployed.
- b Mean (SD) of 1–4-point response scale.
Prevalence rates for life-time cigarette smoking increased as the sample aged (8.9% at wave 1, 12.7% at wave 2 and 13.7% at wave 3). Perceived availability of cigarettes also increased over time and perceived enforcement of underage tobacco laws decreased.
Life-time cigarette smoking
Results of the multi-level analyses are shown in Table 2. Overall, controlling for individual- and city-level demographics, greater tobacco outlet density was associated with increased likelihood of life-time smoking. In addition, a significant cross-level interaction was found between outlet density and time. Specifically, at wave 1, the likelihood of life-time smoking among youths who lived in cities with greater density of tobacco outlets was higher than that of youths who lived in cities with lower density of tobacco outlets. By wave 3, however, the likelihood of life-time smoking of youths in cities with lower tobacco outlet density converged with that of youths who lived in cities with higher tobacco outlet density. Figure 1 shows the regression slopes of wave and life-time cigarette smoking for high and low levels of tobacco outlet density (25th and 75th percentiles). Neither sales rates nor local enforcement of underage tobacco sales laws were associated with youths’ life-time smoking. Sensitivity analyses were conducted excluding data from those who were of legal tobacco purchase age in California at wave 3 (i.e. aged 18 years). Results were very similar to those for the complete sample, but because of reduced power the interaction between time and outlet density only approached significance (P = 0.09).
| Predictors | Life-time cigarette smoking OR (95% CIs) | Perceived cigarette availability Beta (95% CIs) | Perceived enforcement Beta (95% CIs) |
|---|---|---|---|
| City level (n = 50) | |||
| Sales without ID checks | 0.99 (0.97, 1.02) | 0.01 (0.00, 0.02)** | −0.01 (−0.01, −0.01)** |
| Compliance checks | 1.26 (0.83, 1.53) | −0.07 (−0.17, 0.03) | 0.06 (−0.00, 0.12)* |
| Tobacco outlet density | 1.12 (1.04, 1.22)** | 0.02 (−0.00, 0.04)* | −0.02 (−0.04, 0.00)** |
| Time × sales without ID checks | – | – | – |
| Time × compliance checks | – | – | – |
| Time × outlet density | 0.96 (0.93, 0.99)* | – | – |
| Population density | 0.81 (0.69, 0.96)* | −0.15 (−0.25, −0.05)** | 0.10 (0.06, 0.14)** |
| Percentage age < 18 years | 0.96 (0.75, 1.23) | 0.01 (−0.08, 0.09) | −0.00 (−0.06, 0.06) |
| Socio-economic status | 1.13 (0.91, 1.41) | 0.06 (−0.02, 0.14) | −0.06 (−0.12, 0.00)* |
| Percentage white | 1.12 (0.85, 1.48) | 0.01 (−0.07, 0.09) | 0.05 (−0.01, 0.11) |
| Percentage African American | 1.13 (0.88, 1.45) | −0.04 (−0.10, 0.02) | 0.07 (0.01, 0.13)* |
| Percentage Hispanic | 1.12 (0.82, 1.53) | 0.05 (−0.07, 0.17) | −0.05 (−0.13, 0.031) |
| Individual level | n = 1478 | n = 1475 | n = 1478 |
| Age (years) | 1.70 (1.44, 2.01)** | 0.33 (0.29, 0.37)** | −0.27 (−0.29,-0.25)** |
| Male | 1.76 (1.31, 2.36)** | 0.09 (0.01, 0.17)* | −0.01 (−0.069, 0.049) |
| Hispanic | 1.38 (0.75, 2.53) | 0.07 (−0.07, 0.21) | 0.02 (−0.10, 0.14) |
| White | 0.79 (0.54, 1.16) | −0.07 (−0.19, 0.05) | −0.02 (−0.10, 0.06) |
| Observation level | n = 3787 | n = 3760 | n = 3786 |
| Survey wave | 3.04 (2.09, 4.46)** | 0.27 (0.25, 0.29)** | −0.23 (−0.27, −0.19)** |
| Age 18 years | 0.86 (0.60, 1.21) | 0.36 (0.26, 0.46)** | −0.80 (−0.92, −0.68)** |
- * P < 0.05;
- ** P < 0.01. CI = confidence interval; OR = odds ratio.
Tobacco beliefs
Controlling for individual- and city-level demographics, greater tobacco outlet density and rates of sales without ID checks were associated with greater perceived availability of cigarettes and with less perceived enforcement of underage tobacco laws (Table 2). Local enforcement of underage tobacco sales laws was associated with increased perceived enforcement of underage tobacco policies.
Discussion
This study examined the associations of tobacco outlet density, retailer sales rates without ID checks and enforcement of underage tobacco laws with youths’ life-time cigarette smoking and tobacco beliefs over 3 years. Consistent with some other studies 7-11, results showed that tobacco outlet density was associated with smoking among youths. That is, life-time smoking by adolescents was greater in cities with higher outlet density. This is also consistent with results of another study, which found that adolescents who visited tobacco outlets two or more times per week had greater odds of smoking initiation 36. Our results also showed that tobacco outlet density was related to youths’ tobacco beliefs, such that greater density was associated with greater perceived cigarette availability and with lower perceived enforcement of underage tobacco policies. These results suggest that increased exposure to tobacco outlets may affect youth's beliefs about ease of access to cigarettes and about the likelihood of getting into trouble for smoking or buying cigarettes, as well as their smoking behavior.
Considering tobacco outlet density and life-time cigarette smoking over time, we found that the likelihood of smoking at wave 1, when respondents were relatively young (aged 13–16 years), was greater in communities with higher outlet density compared with youths who lived in cities with lower density of tobacco outlets. Two years later, however, life-time smoking among youths in cities with lower density of tobacco outlets converged with that of youths who lived in cities with higher tobacco outlet density. These results suggest that higher tobacco outlet density may be associated with smoking among youths at earlier ages. The implications of these results should be considered when implementing interventions that target younger youths, as early smoking initiation is associated with progression to daily smoking, heavier smoking, nicotine dependence in adulthood and to greater difficulty with smoking cessation 1.
No main effects were found for levels of local enforcement or for retailer sales of cigarettes without ID checks on youths’ life-time smoking. Rate of sales without ID checks, however, was related positively to youths’ beliefs about the ease of obtaining cigarettes and negatively to beliefs about enforcement of underage tobacco laws. Conversely, local enforcement was associated with beliefs about enforcement of underage tobacco policies. These findings are important, given that these beliefs are associated with, and may be predisposing factors for, youths’ cigarette smoking 37, 38. These findings thus suggest that compliance checks and other programs to increase retailer compliance with underage tobacco sales laws may also be of value.
The findings from this study have implications for prevention. Living in communities with higher outlet densities was associated with smoking by youths, greater perceived access to cigarettes and lower levels of perceived enforcement of underage tobacco laws. The relationship between density and smoking was most pronounced at the initial waves of the survey when the participants were younger. Overall, these findings suggest that controls on the numbers and locations of tobacco outlets is a key element in a comprehensive strategy for delaying smoking, especially among younger adolescents. High densities of tobacco outlets may also serve a normative function, affecting adolescents’ beliefs about the ease of obtaining cigarettes and the likelihood of getting into trouble with authorities for smoking.
The results of this study should be considered in light of some limitations. The study included only mid-sized cities in California. A different pattern of relations might obtain in rural or more urban communities. Also, data about local enforcement were obtained from a single key informant in each community, and therefore may not be accurate or may be subject to social desirability biases. The positive association, however, between the reports obtained from the key informants and youths’ perceived enforcement of underage tobacco laws suggests that the single source of enforcement information may not have jeopardized study validity. It should be noted further that this study focused on the associations of community-level outlet density, sales without ID checks and enforcement with youths’ smoking behaviors and beliefs. Considering these factors at smaller geographic units such as the neighborhood level or within youths’ everyday activity spaces may clarify further how they relate to smoking beliefs and behaviors 39. Finally, this study did not consider youth's access to cigarettes outside their city of residence. As youths grow older they experience increased mobility, through driving or increased use of public transportation, which may moderate any relationship between retail access to cigarettes within city boundaries and youths’ smoking behaviors and beliefs. Future studies should consider youths’ mobility and retail access to cigarettes in the broader environment where they spend their time. Despite these limitations, the current study makes a significant contribution by providing additional evidence regarding the roles that outlet density, enforcement of underage sales laws, and retailer non-compliance may play in youths’ life-time cigarette smoking and beliefs.
Declaration of Interests
None.
Acknowledgements
This research and preparation of this manuscript were made possible by grants CA138956 from the National Cancer Institute (NCI) and AA006282 from the National Institute on Alcohol Abuse and Alcoholism (NIAAA) of the National Institutes of Health (NIH) and grant 19CA-016 from the Tobacco-Related Disease Research Program (TRDRP; http://www.trdrp.org). The content is solely the responsibility of the authors and does not necessarily represent the official view of NIH, NCI, NIAAA or TRDRP.
