Abstract

Our objective was to review the research on the effects of public clean air laws on smoking rates, compare these effects to those found in studies on the impact of private worksite restrictions and derive estimates of the potential reductions in smoking rates that might be expected from the implementation of the two types of policies. Data sources were computerized databases, references identified from pertinent peer‐reviewed journal articles and books, and suggestions by experts on tobacco control policy. Comprehensive public clean air laws have the potential to reduce prevalence and consumption rates of the entire population (including non‐working and non‐indoor working smokers) by about 10%. Studies on private worksite regulations also suggest that strong worksite restrictions have the potential to reduce the prevalence rate of the entire population by about 6% over the long‐term and the quantity smoked by continuing smokers by 2–8%, depending on the length of time after the ban. Further research is needed on the effects of the different types of public clean air policies on the entire smoking population and on different sociodemographic groups, how the effects of public clean indoor air laws depend on private restrictions already in place, and how the effect of private restrictions depend on whether or not they are supported by public clean air laws.

Introduction

Clean air laws have been enacted to reduce the harmful effects of environmental tobacco smoke (ETS) on non‐smokers by restricting or banning smoking in designated public areas (Centers for Disease Control and Prevention, Office on Smoking and Health, 1986; Environmental Protection Agency, 1992; National Cancer Institute, 1999). While clean air laws are often justified in terms of reducing exposure to second‐hand smoke, they may also reduce smoking rates. Restrictions on smoking may limit smoking by reducing opportunities to smoke and by changing attitudes towards smoking (Department of Health and Human Services, 2000; Levy and Friend 2001). By reducing opportunities to smoke, smoking restrictions directly reduce the quantity of cigarettes smoked in the restricted areas, which may improve the chances of quitting. Smoking restrictions may also change norms regarding the social acceptability of smoking. As social attitudes change, smokers may be induced to attempt to quit or not initiate, thereby reducing the number of smokers.

Recent attention has centered on getting clean indoor laws passed as part of a comprehensive tobacco control strategy (Centers for Disease Control and Prevention, 1999; Department of Health and Human Services, 2000). Smoking may be restricted at worksites, restaurants and other public places, such as grocery stores, shopping malls and public transit. Clean air laws in the US are mostly implemented at the state or local level. In 2001, two states did not allow any smoking in restaurants, one state required that work sites and restaurants allow smoking only in designated smoking areas with separate ventilation, 28 states required or allowed smoking in restaurants in designated areas, and 20 states required or allowed smoking in worksites in designated areas (www.cdc.gov/nccdphp/osh/state/rpt_map). The numbers have not changed since 1996. By 1998, there were about 850 communities with major smoke‐free worksite and/or restaurant ordinances. Most of the growth occurred in several states in the years 1993 and 1994 (National Cancer Institute, 2000).

Many private businesses in the US have adopted smoking restrictions in states that have not enacted clean air laws (National Cancer Institute, 2000). The percent of indoor workers who reported that their worksites were smoke‐free was 47% in 1993 (Gerlach et al., 1997), 64% in 1996 (Burns et al., 2001) and 69% in 1999 (Shopland et al., 2001). While the percentage of indoor air workers who work in places that have smoking bans is correlated (Shopland et al., 2001), many of the private worksite bans were implemented in states and cities that did not have public laws requiring such extensive restrictions. Further, many workers in states with worksite laws have reported that there were no smoking restrictions at their worksites.

Prior reviews have mostly examined studies of the effects of smoking restrictions implemented by individual private worksites [e.g. (Brownson, 1997; Eriksen and Gottlieb, 1998; Chapman et al., 1999; Burns et al., 2001; Hopkins et al., 2001; Fichtenburg and Glantz, 2002)]. Moreover, these studies have generally been limited to firms in particular industries, which may not be representative of firms in the rest of the country. These reviews are sometimes used to justify the passage of public clean air laws, without explicitly considering differences between public laws and private restrictions, and the issues involved in translating laws into practice.

In this paper, we consider a broader literature than previous reviews. In addition to considering studies of individual firms or firms in particular industries, we extensively consider population studies of the effect of private work restrictions on smoking behaviors. In addition, we review an extensive literature on the effects of public clean air laws. Within and across the different literatures, we compare the results found for effects of smoking restrictions on smoking behaviors. The results of these different types of investigations are also compared in terms of their respective strengths and limitations. Interestingly, although the literatures generally employ different methodologies, they provide complementary results. Limitations of the studies are then considered to guide future research.

Methods

Studies were collected using the Centers for Disease Control and Prevention’s Office of Smoking and Health’s website, the Medline website, Science Citation Index, Social Sciences Citation Index and the Tobacco Control website. In addition, we examined references identified from pertinent articles and books, including recent reviews (Brownson et al., 1997; Eriksen and Gottlieb, 1998; Chapman et al., 1999; Centers for Disease Control and Prevention, 2000; Burns et al., 2001; Hopkins et al., 2001). We also consulted with a panel of experts on tobacco control policy regarding available studies. The panel helped ensure that we included all relevant studies and helped us integrate the research to derive best estimates of the effects of clean air policies and smoking restrictions.

The review is limited to studies conducted in the US since 1985. We limit the review to one country and a limited time period so that the results from the different studies can be more reliably compared, but we address studies from other nations in the Discussion section. We only include studies that examined the effect of smoking restrictions on at least one type of smoking behavior, such as changes in quantity smoked per smoker, prevalence and cessation rates. In attempting to provide a broad overview of all studies, we include any study that either made comparisons over time (i.e. before and after smoking restrictions were implemented) or between populations exposed and those not exposed to smoking restrictions. We include published articles and books, and refereed reports (e.g. by the National Bureau of Economic Research).

For purposes of comparing the different studies, differences in smoking behaviors between restricted and unrestricted areas were measured relative to their levels in the absence of restrictions (i.e. the absolute change relative to the base rate without restrictions) wherever possible. Comparing results across a divergent set of studies is a challenging task. We considered conducting meta‐analysis to address this issue. However, because investigations varied widely in their designs, methodologies, samples and outcomes, we chose not to employ this technique (Hedges and Olkin, 1985). Instead, we decided to integrate study results by comparing the range of results and characteristics of the sample (time period and population), and by considering study quality. We gave greater emphasis to the better‐conducted studies, such as those that included comparison groups and followed cohorts over time, in developing overall estimates. We also enlisted the assistance of tobacco control experts to ensure that the derived estimates seemed reasonable in deriving a reasonable range of the possible impact of laws and private smoking restrictions.

The effects of private worksite restrictions

Studies of individual worksites

Numerous studies have examined the effects of smoking restrictions in an individual firm or selected worksites within a particular industry. This research has generally compared smoking‐related behaviors at a point or points in time after workplace restrictions were implemented to baseline rates, using both prospective and retrospective approaches (see Table I). The more well‐designed studies included control sites and followed a cohort of workers over time.

Because comprehensive reviews of this literature have already been conducted [e.g. (Brownson et al., 1997; Eriksen and Gottlieb, 1998; Chapman et al., 1999; Hopkins et al., 2001; Fichtenburg and Glantz 2002)], we summarized the overall findings (see Table III). In general, retrospective studies of the impact of individual worksite restrictions reported that quantity smoked among continuing smokers was reduced in the range of 10–20% after 6–13 months. Prospective cohort studies also reported similar reductions in prevalence rates, ranging from 7 to 20% reductions, after the restrictions had been in place 1 year or more. The effects on quit rates were less consistent.

As important as the magnitude of the effects on quantity and prevalence is how these effects appear to unfold over time. Reductions in quantity smoked appeared to show their greatest decline within the first 6 months and then decreased over time [e.g. (Centers for Disease Control and Prevention, 1990; Hudzinski and Frohlich, 1990; Stave and Jackson, 1991; Jeffery et al., 1994; Olive and Ballard, 1996]. By contrast, prevalence and quit rates generally showed no or little immediate effect [e.g. (Rosenstock et al., 1986; Becker et al., 1989; Mullooly et al., 1990)] and subsequent greater increases over time [e.g. (Centers for Disease Control and Prevention, 1990; Stillman et al., 1990)] although the evidence for this trend is less consistent than for quantity smoked (Jeffrey et al., 1994). Longo et al. (Longo et al., 1996, 1999) found that quit rates of workers in hospitals with bans increase over the 6 years following a ban, and that relapse rates appear to be similar to those who quit and are not in firms with bans.

Thus, evidence suggests that the effects of restrictions on quantity erode, while the effects on quit rates increase and are maintained. Smokers who reduce their consumption may have a greater likelihood of quitting, but successful cessation may require more than one attempt (Hughes, 2000). The return to higher levels of quantity smoked found in more long‐term studies may be due to smokers with lower levels of consumption quitting, resulting in heavier smokers assuming a disproportionate number in the sample of remaining smokers.

Like previous reviews (Brownson et al., 1997; Chapman et al., 1999; Eriksen and Gottlieb, 1998; Hopkins et al., 2001), we found that worksite policies are generally associated with reductions in the quantity of cigarettes smoked per smoker, but less clear patterns emerge for cessation and prevalence rates. Overall, there is considerable variation in the amount of reduction in smoking behaviors observed for each of the different measures. Some of the variations can be attributed to differences in sample size, time of follow‐up, type of industry, the way in which smoking is measured (e.g. whether a smoker who quits is included in follow‐up consumption measures), the extent of smoking restrictions and interventions implemented in conjunction with the ban (e.g. cessation treatment programs). In particular, the type of workplace in which a ban is implemented may influence the impact of the policy. For example, many of the studies examined health care facilities. If those sites were more likely to attract non‐smokers as employees, restrictions may be more enforceable. Because the investigations of individual worksites only examined specific firms, their results may not generalize to other types of businesses. Moreover, because surveys were often conducted at the workplace, there is the potential for bias in reporting due to the social unacceptability of smoking.

Population‐based studies of worksites

Nine different population‐based studies are shown in Table II. Population‐based studies typically employ a random selection of workers in a large area, such as a state or an entire country. Population‐based worksite studies may be more representative of the effects of workplace bans than the studies of individual firms discussed above. While businesses choosing to implement workplace bans may still be different in some way from those that do not, specific firms are not singled out. Consequently, these studies are likely to be more representative of the population impact of smoking bans than studies that select one or a limited number of firms. In addition, the surveys were not conducted at the workplace, thereby reducing the likelihood of workers only providing socially desirable responses.

Apart from having a narrower sample, most of the better firm level case studies discussed in the previous section examined smoking behaviors over a particular period of time, such as a year after a ban was implemented. By contrast, most population studies compared smoking behavior at a point in time in those firms with smoking restrictions to those without restrictions, without distinguishing when the restrictions were adopted. The latter investigations did not control for how long the smoking restrictions has been in effect, but tended to show more clearly the long‐term impact of smoking restrictions, since the bans have been in effect in most firms for more than one year.

Population‐based worksite studies usually employed multivariate regression analysis to control for other factors affecting the relationship between the bans and changes in smoking behaviors. Nevertheless, because they did not directly examine changes in behavior over time, their reliability in isolating the role of smoking restrictions depends on the ability to control for confounding factors that may explain differences across firms.

We chose to present general results of population‐based worksite studies, with greater detail provided in Table II. Consistent with the findings reported for individual worksite restrictions, population‐based bans were also associated with lower daily quantity smoked among smokers who continued to smoke. Various authors (Woodruff et al., 1993; Glasgow et al., 1997; Evans et al., 1999; Farkas et al., 1999) reported between 7 and 15% fewer cigarettes smoked per smoker. Patten et al. (Patten et al., 1995) obtained decreases of similar magnitudes in quantity smoked, but the results were not statistically significant. Larger effects (34%) were indicated by Kinne et al. (Kinne et al., 1993), but they only examined mean differences in smoking rates and did not control for other confounding factors.

Workers in firms with worksite bans were also found to have lower prevalence rates than workers in firms without bans. Evans et al. (Evans et al., 1999) and Farrelly et al. (Farrelly et al., 1999) found 15–20% lower prevalence rates, with absolute differences of 5 percentage points. Greater differences were indicated by Woodruff et al. (Woodruff et al., 1993) and Kinne et al. (Kinne et al., 1993), but these studies did not control for confounding factors that may have also contributed to smoking reductions.

Workers in firms with worksite bans have also generally found higher rates of quit attempts and successful quits lasting at least 3 months. Burns et al. (Burns et al., 2001), Farkas et al. (Farkas et al., 1999) and Glasgow et al. (Glasgow et al., 1997) reported that cessation rates were about 10–15% higher in firms with bans. Some studies, however, reported little or no significant effects (Burns et al., 2001; Patten et al., 1995). Biener and Nyman (Biener and Nyman, 1999) found relatively large, but insignificant, effects after 3 years.

Population‐based studies also considered the impact of different levels of smoking restrictions. In general, partial restrictions appeared to have little or no effect on smoking behaviors, whereas larger effects were observed for more extensive restrictions (Glasgow et al., 1997; Farrelly et al., 1999). These findings suggest that the ability to circumvent a smoking ban will be an important consideration in determining a ban’s impact.

In summary, population‐based studies have generally found that firms with strong smoking restrictions reported a 10–15% reduction in quantity consumed and 15–20% decrease in smoking prevalence relative to businesses with minimal or no restrictions. Most studies also reported at least 10–15% higher cessation rates in firms that implemented total smoking bans. In addition, larger effects were found for those who worked longer hours, suggesting a dose–response relationship (Evans et al., 1999).

Discrepancies among study results may be accounted for in part by the fact that some studies did not control for confounding factors, such as smoker sociodemographic characteristics [e.g. (Kinne et al., 1993; Woodruff et al., 1993)]. In addition, studies may need to consider the health habits of the workers being examined to account for the possibility that those firms with smoking bans may be those where the workers were generally more health conscious (Evans et al., 1999). Nevertheless, the results of the population‐based studies are generally consistent with and complement the studies of individual firms. Moreover, by providing estimates of population impacts over longer periods of time, they increase our confidence in our overall estimates of effects. Interestingly, the effects on prevalence rates are somewhat higher and for quantity consumed per continuing smoker are lower for the population‐based studies than the studies of individual firms.

The effects of public laws on smoking behaviors

Eighteen studies were found that directly examined the effects of public clean air laws on population smoking rates. Table III includes studies that combine all ages, whereas Table IV includes studies for particular age groups. All of the studies in Tables III and IV employed some form of multivariate analysis, in which they controlled for the effects of other tobacco control policies, such as taxes, and for other factors, such as smokers’ sociodemographic characteristics.

Studies of public clean air laws either individually categorized different clean air policies or combined the laws into a single index. In the latter case, an index of restrictiveness was used, based on a classification scheme proposed by the Surgeon General (Centers for Disease Control and Prevention, Office on Smoking and Health, 1986, 1989). Typically, the index was calibrated on a 0–1 scale, in which 0 indicated no restrictions, 0.25 represented basic restrictions (e.g. one to three designated smoking areas locations, excluding private worksites and restaurants), 0.5 represented nominal restrictions (e.g. four or more designated areas, excluding private worksites and restaurants), 0.75 represented a moderate policy (e.g. restaurant but no private worksite restrictions) and 1.0 indicated extensive clean air laws (e.g. laws that included private worksite and other restrictions).

Most studies examined the effect of clean air laws measured in terms of cigarettes smoked per capita (including non‐smokers). Results from Emont et al. (Emont et al., 1992) indicated that states with extensive clean air laws had 12% lower per capita consumption rates than other states. After accounting for smoking sentiment, Chaloupka and Saffer (Chaloupka and Saffer, 1992) found about a 20% reduction in per capita consumption from public clean air laws, compared to 4–8% effects without such controls. They found, however, that the effect of worksite laws fell from an 8% reduction to statistically non‐significant effects when social attitudes regarding smoking were taken into account. They also reported that basic and moderate levels of restrictions were sufficient to yield significant effects, whereas Emont et al. (Emont et al., 1992) found that moderate or even extensive restrictions were necessary. Wasserman et al. (Wasserman et al., 1991) estimated that expanding smoking restrictions from minimal to extensive policies would reduce overall per capita smoking by almost 6%. Yurekli and Zhang (Yurekli and Zhang, 2000) used a modified restrictiveness index, in which worksite laws and laws that had been in effect longer were more heavily weighted. They estimated a 5% reduction in smoking due to the clean air laws that were in effect in 1995. Upon extrapolating to all states implementing extensive laws, their results implied that consumption would be reduced by about 10% in changing from no laws to extensive laws. [These calculations were derived by multiplying the maximum value of the restrictiveness index (3) by the coefficient on the restrictiveness index.]

Few investigations, except for studies of youth smoking, examined the effects of clean air laws on smoking prevalence. Results from Emont et al. (Emont et al., 1992) indicated that states with extensive clean air laws had mean prevalence rates that were 14% lower than other states. Consistent with this result, Ohsfeldt et al. (Ohsfeldt et al., 1998) found that extensive laws, relative to minimal laws, were associated with a 13% lower smoking rate for all males, with a smaller effect for females. Among different age groups, the largest effects were found for those aged 25–44. Like the Chaloupka and Saffer (Chaloupka and Saffer, 1992) investigation described above, Ohsfeldt et al. (Ohsfeldt et al., 1998) also considered smoking sentiment, but generally found greater reductions in smoking when attitudes toward smoking sentiment were taken into account.

Two studies explored the effects of clean air laws on quit rates. Results from Emont et al. (Emont et al., 1992) indicated that states with extensive clean air laws had 12% higher mean quit (former current) rates than other states. Moskowitz et al. (Moskowitz et al., 2000) reported a 38% higher 6‐month cessation rate in areas with a strong local ordinance. Similar results were obtained for different sociodemographic groups and for 5‐year quit rates.

Impact on youth smoking

While restrictions by private firms affect the workers who are generally above age 18, clean air laws may also affect youth. Recent studies of youth and young adult cigarette demand have also shown that clean air laws reduced smoking among these groups. Wasserman et al. (Wasserman et al., 1991) found that increasing state restrictions to the most comprehensive level reduced youth per consumption per smoker by over 40%. However, their measure of clean air laws was highly correlated with price, rendering this result questionable. Chaloupka and Grossman (Chaloupka and Grossman, 1996) found that restrictions on smoking in schools resulted in decreased quantity smoked by young smokers, but the effects of other clean air laws were often insignificant. Using a broader range of smoking restrictions (e.g. self‐reported home restrictions) and enforcement variables, Wakefield et al. (Wakefield et al., 2000) found more consistent effects of clean air policies on youth quantity smoked, including relatively consistent effects for reducing progression to established smoking. Tauras and Chaloupka (Tauras and Chaloupka, 1999) also found relatively consistent effects of clean air laws for older youth and young adults.

Several studies have found that the effects of clean air laws on youth varied by demographic subgroup. Lewitt et al. (Lewitt et al., 1997) found that only males reduced their smoking rates in response to clean air policies, and Chaloupka and Pacula (Chaloupka and Pacula, 1999) reported that only males and whites were influenced. Chaloupka and Wechsler (Chaloupka and Wechsler, 1997) found almost significant effects for college aged youth.

Discrepant results have been reported for school restrictions. Chaloupka and Grossman (Chaloupka and Grossman, 1996) did not find effects on use, suggesting that school clean air policies alone might be insufficient to combat youth smoking. In contrast, Wakefield et al. (Wakefield et al., 2000), using refined statistical techniques and more recent data, obtained significant effects of school restrictions on smoking rates.

Prospective case studies

Unlike the studies cited above, which used multivariate regression analyses, Pentz et al. (Pentz et al., 1989) and Paulozzi et al. (Paulozzi et al., 1992) used a prospective case study approach. Pentz et al. compared smoking rates in schools with strict bans to schools with only smoking education programs. After 17 months, they found that those schools with comprehensive restrictions and smoking education programs were more effective in reducing smoking behavior, especially quantity smoked among seventh graders than schools with education programs only. Paulozzi et al. considered a Vermont worksite smoking law. Sixteen months after the ban, the number of cigarettes and the percent of adults smoking declined about 30% at work and about 20% at home.

Summary of the research on clean air laws

In summary, relatively extensive clean air laws were generally found to be associated with lower smoking consumption and prevalence rates, and higher cessation rates. Studies found that overall consumption was reduced in the 4–20% range. The results also suggested that workplace clean air laws may have their greatest impact on men, especially those between the ages of 25 and 44, who are more heavily represented in the workforce than women. In general, studies found that clean air laws yield reduced smoking behaviors among youth, but provided little guidance on the magnitude of effects.

In evaluating the studies of clean air laws, several methodological issues merit attention. First, these studies often did not control for the length of time that a law was in place. In addition, most studies used a single index for all laws, due to the difficulty in untangling the effects of different clean air laws. Related to the previous point, many studies did not consider the restrictiveness of the law (e.g. whether it applied to only designated areas). These methodological problems may limit the validity of our estimates of the laws’ impact. In addition, all of the studies except Yurekli and Zhang (Yurekli and Zhang, 2000) examined state laws in effect before 1994, when strict laws were unlikely to have yet been implemented (National Cancer Institute, 2000). The impact of the more extensive laws currently in place and the role of additional private restrictions remain to be determined.

A potential problem in determining the specific effect of clean air laws is that the states with strongest anti‐tobacco attitudes may be more likely to enact and enforce clean air policies. Thus, it is difficult to distinguish what part of the reduction in smoking rates is attributable to the laws and what is attributable to attitudinal changes. Anti‐smoking sentiment may result in increased compliance with the law [(Levy and Friend, 2001; Rigotti et al., 1992, 1993, 1994; Pierce, 1994), but see (Moskowitz et al., 1999)], suggesting that one could not exist without the other and that attempting to determine their individual unique role may be difficult.

If the government is held responsible for enforcement, the extent of compliance may depend on the specific government agency required to carry out this task. Local laws may be accompanied by greater community support, which may increase compliance with the laws and support norms against smoking. Enforcement efforts may, however, be more efficient at the state level. In particular, states laws reduce the ability of workers to avoid more stringent laws in neighboring communities by changing jobs. Almost all of the studies examined only state laws. Studies are needed to determine the effects of local versus state laws on compliance. In the absence of government enforcement, voluntary compliance may be high if there is sufficient public support for the law, as indicated by recent studies of compliance with non‐worksite laws (Jacobson and Wasserman, 1997; Goldstein and Sobel, 1998; Hyland et al., 1999).

Discussion

Overall summary

A multitude of studies using different methodologies have found that smoking restrictions, whether imposed by public laws or private firms, reduced the quantity smoked per smoker and smoking prevalence. Studies of clean air laws indicated reductions of between 5 and 20% in rates of smoking per capita over the entire population. A rough estimate of the impact of changing from minimal clean air laws (three or fewer areas, excluding restaurants or worksites) to extensive laws (including worksite and restaurant bans) is a 10% reduction in prevalence rates. Thus, this decrease implies that the current US prevalence estimate of 22.8% (Centers for Disease Control and Prevention 2002) would drop to 20.5% once the laws had been in place for several years. Similar effects were found for studies of total cigarettes consumed. Ohsfeldt et al. (Ohsfeldt et al., 1998) explicitly distinguished worksite laws from other clean air laws and found that they explained about half of the effect of clean air laws.

The research on individual firms found that private workplace restrictions reduced smoking rates, with the initial effects focusing mostly on quantity smoked per smoker and later effects on use rates. These studies obtained stronger evidence on quantity smoked than on prevalence rates or quitting, but cross‐sectional studies yielded quite consistence effects on smoking prevalence. Population‐based studies indicated effects closer to the 20% estimate for prevalence rates and 5–20% for quantity smoked per smoker.

The results from studies of private work restrictions are consistent with those of clean air laws. In comparing results, we take into account that that only 67% of workers worked indoors (Evans et al., 2000) and that 63% of the population worked (Department of Commerce, 1997). Thus, the effects are spread over 42% of the population. The 10–20% reductions in smoker prevalence by workers would imply between 4 and 8% reduction for the overall adult smoking population, which is consistent with the effects found in studies of clean air worksite laws. Reductions in quantity smoked per smoker would be between 2 and 8%, depending on the length of time after the ban was implemented. These effects appear to erode over time, as those who most reduce their quantity may quit and are no longer represented as smokers with reduced quantities smoked.

We have not included studies from other countries in our estimates of effects of smoking restrictions and clean air laws because other countries may differ in important ways from the US in their smoking policies and practices. Nonetheless, these studies generally provide similar results as those reported above. Three Australian studies, a prospective cohort study by Borland et al. (Borland et al., 1990), a prospective cross‐sectional study by Borland et al. (Borland et al., 1991) and a population study by Wakefield et al. (Wakefield et al., 1992), all found a 20–25% reduction in the number of cigarettes smoked per smoker. The two studies by Borland et al. (Borland et al., 1990, 1991) reported between 5 and 10% lower prevalence rate in worksites with a ban. Etter et al. (Etter et al., 1999) found increased numbers of quit attempts, but no difference in the number of cigarettes smoked or prevalence rates, after 4 months in a retrospective cohort study for a Swiss university. By contrast, a population study conducted by Brenner and Fleischle (Brenner and Fleischle, 1994) for Germany found 36% fewer cigarettes smoked per smoker and a 7% lower prevalence rate in firms with smoking bans. A Canadian cohort prospective study by Broder et al. (Broder et al., 1993) found an 11% reduction in the number of cigarettes smoked, but no change in smoking rates. Two studies of clean air laws were found for countries other than the US. In Canada, Stephens et al. (Stephens et al., 1997) found that the odds ratio for being a smoker was 1.21 where clean air laws covered a smaller percent of the population, compared to an odds ratio of 1.26 where cigarettes were relatively inexpensive, suggesting that stricter clean air laws are nearly as effective as large price differences in reducing smoking rates. Stephens et al. (Stephens et al., 2001) found that the restrictiveness of municipal bylaws limiting public smoking was positively associated with the odds of being a non‐smoker and negatively with amount smoked for women but not men. Following implementation of a national smoke‐free law in Finland, Heloma et al. (Heloma et al., 2001) found declines in smoking prevalence of 17% and a reduction in the number of cigarettes smoked per smoker of 16% in firms previously without bans.

Suggestions for future research

While there is a strong basis for claiming that smoking restrictions yielded reductions in smoking behaviors, our ability to determine which specific aspects of clean air laws were responsible for changes in smoking rates is subject to the limitations indicated above. In particular, the influence of social norms merits additional study. Further attention is also needed to determine how the effects of clean air laws unfold over time. While smoking restrictions implemented by private firms appear to first affect the quantity of cigarettes smoked per smoker and later lead to cessation, knowledge about the interaction between reductions in quantity smoked and future quits is limited (Hughes, 2000). Attention also needs to be directed at developing rigorous measures of quit rates and determining the effects of clean air laws on future cessation.

Another limitation of the literature is that studies with null findings may not be published (‘the file drawer problem’). The possibility of publication bias suggests that the actual effects are lower than that which would be indicated by the empirical literature. We have, however, found similar relatively similar results across a diverse set of studies and methodologies, and, in particular, for population studies where observations are randomly chosen over the population.

The effect of newly implemented clean air laws may depend on which firms already have strong workplace restrictions (Levy and Friend, 2001). Sorensen et al. (Sorensen et al., 1996) reported that non‐manufacturing firms and firms with more female workers were more likely to adopt worksite restrictions. In general, those workplaces with more smoking restrictions may be those that have the fewest smokers, the most health conscious workers or those that originally had the greatest ETS problem (Woodruff et al., 1993; Evans et al, 1999). Over time, smokers may be drawn to work in firms with the fewest restrictions.

The issue of compliance with worksite restrictions and clean air laws also merits further study. Compliance with new clean air laws may be more difficult in those firms that have not already voluntarily enacted bans, while the implementation of clean air laws may increase compliance rates in firms that had previously implemented worksite restrictions.

At the same time, studies of private workplace restrictions typically have not considered how the impact of those restrictions may depend on the state local laws governing their areas. Biener and Nyman (Biener and Nyman, 1999) found that many firms do not enforce their own policies. Smokers in firms that already have restrictions may reduce their smoking rates if new laws increase compliance with individual firm restrictions or the difficulty of workers switching jobs in order to work at a firm that allows smoking.

Thus, previous studies of clean air laws have not considered the effect of private restrictions already in place and population studies of private restrictions have not considered the effect of clean air laws. The rate of firms with smoking restrictions and the extent of clean air laws varies both in the US and across the globe. Additional investigations are needed that examine the inter‐relationship between clean air laws and private smoking restrictions already in place. In those countries without norms against smoking, voluntary firm restrictions may be less likely and clean air laws may be more difficult to enforce.

The effect of both private restrictions and clean air laws may be enhanced by other tobacco control policies already in place. While some studies of clean air laws control for taxes, none specifically consider interactive effects of other policies with clean air laws. For example, Farkas et al. (Farkas et al., 1999) found large effects of voluntary home smoking bans on smoking rates. These bans may reinforce the effect of clean air laws and workplace restrictions, although the ability to distinguish the effect of worksite bans from home smoking bans may be difficult if the imposition of home bans and firms bans are inter‐related. Other policies may intensify the effects of clean air laws if they reinforce anti‐smoking norms, such as media publicity to generate support for and continued compliance with the law.

Additional research is also needed on how smoking restrictions influence different sociodemographic and smoker groups. Clean air laws are likely to be particularly effective at reducing smoking by those in the workplace. With substantial interest directed at those below the age of 18, the effects of public policy on those aged 18–24 becomes more important as these individuals become both legal purchasers of cigarettes and full‐time workers. The effect of college bans also merits consideration. In addition, future work is warranted that explores the effects of laws on heavy versus light smokers.

In conclusion, clean air policies have a direct effect on smokers. While studies of private firms and of public laws both indicate that smoking restrictions affect smoking behavior, further information is needed on the interactions between private and publicly imposed restrictions and the time pattern of effects. Clean air laws have a well‐defined place in comprehensive tobacco control programs (Centers for Disease Control and Prevention, 1999; Department of Health and Human Services, 2000), but better information is needed on how their effects compare with other policies. Knowledge of the effects of clean air laws, alone and in combination with other tobacco control policies, and how they affect different sociodemographic groups, will be important in developing comprehensive strategies to reduce smoking rates.

Acknowledgements

This report was written as part of a contract from the Substance Abuse and Mental Health Services Administration. We would like to thank L. Biener, M. Carmona, F. Chaloupka, M. Cummings, J. DiFranza, W. Evans, M. Farrelly and J. Forster for serving on our expert panel and for their comments on previous drafts of this paper. Not for quotation or circulation without permission of Pacific Institute for Research and Evaluation.

Table I.

Effects of individual worksite restrictions on smoking rates

Study and year Sample and type of restriction Study methods Percent change in cigarettes smoked per smoker Percent change in smoking prevalence unless otherwise indicated 
Rosenstock et al. (1986) HMO; ban except in one room retrospective; no control –11.4% at 4 months NS at 4 months 
Peterson et al. (1988) insurance company; ban in work areas only retrospective; no control –29.5% at 3 months –6.3% at 3 months; +0.5% quit rate 
Biener et al. (1989) hospital; designated areas prospective with controls N/A NS (incl. quit attempts) 
Becker et al. (1989) children’s hospital; total ban prospective; no control NS at 6 months –6.6% at 6 months 
Scott et al. (1989) insurance company; designated areas Retrospective; one group, post‐ban only; no control heavy smokers: –47%; light smokers: –22%; OR = 3.08 +11% QR at 7 months 
CDC (1990) psychiatric hospital; designated areas (pre‐ban) to total ban retrospective cohort; no control –11% at 12 months –13.8% at13 months; –17.2% at 17 months 
Gottlieb et al. (1990) state government agency; designated areas prospective cohort; no control % smoking 15 or more cigarettes: –5.7% at 1 month; +2% at 6 months +13.8% at 1 month; –14.8% at 6 months; +1.6% QA at 1 month; +2.2% QA at 6 months 
Hudzinski and Frohlich (1990) hospital; total ban, except psychiatric inpatients prospective cohort; no control –20.1% at 12 months –9% at 6 months; –36% at 1 year 
Mullooly et al. (1990) non‐physician employees at large HMO; 11 sites; total ban prospective cross‐sectional, no control NS at 12 or 24 months NS (including quit attempts) at 12 or 24 months 
Stillman et al. (1990) hospital; total ban prospective cohort; no control –20% at 6 months –25% at 6 months; +25.6% quit rate at ≥8 months 
Baile et al. (1991) cancer center; total ban retrospective; no control –54.2% at 4 months too small (N = 5) to analyze 
Stave and Jackson (1991) two medical centers; total ban retrospective with controls –24.1% at intervention site at 1 year; –1.1% at control site self‐reported: +45% QR at 3 months; +69% at 9 months; CO validated: +85% QR at 3 months; +73% at 9 months 
Daughton et al. (1992) hospital; total ban prospective cohort; no control –18.6% at 5 months +14% QR rate at 12 months; +82% at 24 months 
Offard et al. (1992) hospital; total ban retrospective; no control 30.2% decreased, 7.4% increased and 62.4% no change at 2 years –17.4% at 30 months; +22.5% QR at 30 months 
Brigham et al. (1994) hospital; total ban prospective cohort with control N/A NS QR at 4 weeks 
Jeffery et al. (1994) 32 worksites; ban in most areas versus no ban or designated area retrospective and prospective cohort with control –11.2% at 2 years; NS in sites without ban NS prevalence, QA or QR 
Olive and Ballard (1996) two federal hospitals; designated area versus total ban prospective cross‐sectional; no control Hospital 1: –13.7% at 6 months; –14.3% at 12 months; Hospital 2: –7.4% at 6 months Hospital 1: –20.4% at 12 months; +43% QR at 6 months, +70.6% at 1 year; Hospital 2: –18.5% at 6 months; +30.1% QR at 6 months 
Longo et al. (1996) hospital; total ban retrospective; logistic analysis N/A QR = 0.35, OR = 2.1 at 6 months; QR = 0.066, OR = 1.8 at 1 year; QR = 0.112, OR = 1.9 at 1.5 years; QR = 0.268, OR = 2.3 at 3 years; QR = 0.506, OR = 1.7 at 5 years 
Longo et al. (1999) hospital; total ban versus other non‐smoke‐free workplaces prospective cohort, with hazard rate models NA post‐ban QR range of 44% higher at 5 years (OR = 1.6), 98% higher at 6 years (OR = 2.4). 
Study and year Sample and type of restriction Study methods Percent change in cigarettes smoked per smoker Percent change in smoking prevalence unless otherwise indicated 
Rosenstock et al. (1986) HMO; ban except in one room retrospective; no control –11.4% at 4 months NS at 4 months 
Peterson et al. (1988) insurance company; ban in work areas only retrospective; no control –29.5% at 3 months –6.3% at 3 months; +0.5% quit rate 
Biener et al. (1989) hospital; designated areas prospective with controls N/A NS (incl. quit attempts) 
Becker et al. (1989) children’s hospital; total ban prospective; no control NS at 6 months –6.6% at 6 months 
Scott et al. (1989) insurance company; designated areas Retrospective; one group, post‐ban only; no control heavy smokers: –47%; light smokers: –22%; OR = 3.08 +11% QR at 7 months 
CDC (1990) psychiatric hospital; designated areas (pre‐ban) to total ban retrospective cohort; no control –11% at 12 months –13.8% at13 months; –17.2% at 17 months 
Gottlieb et al. (1990) state government agency; designated areas prospective cohort; no control % smoking 15 or more cigarettes: –5.7% at 1 month; +2% at 6 months +13.8% at 1 month; –14.8% at 6 months; +1.6% QA at 1 month; +2.2% QA at 6 months 
Hudzinski and Frohlich (1990) hospital; total ban, except psychiatric inpatients prospective cohort; no control –20.1% at 12 months –9% at 6 months; –36% at 1 year 
Mullooly et al. (1990) non‐physician employees at large HMO; 11 sites; total ban prospective cross‐sectional, no control NS at 12 or 24 months NS (including quit attempts) at 12 or 24 months 
Stillman et al. (1990) hospital; total ban prospective cohort; no control –20% at 6 months –25% at 6 months; +25.6% quit rate at ≥8 months 
Baile et al. (1991) cancer center; total ban retrospective; no control –54.2% at 4 months too small (N = 5) to analyze 
Stave and Jackson (1991) two medical centers; total ban retrospective with controls –24.1% at intervention site at 1 year; –1.1% at control site self‐reported: +45% QR at 3 months; +69% at 9 months; CO validated: +85% QR at 3 months; +73% at 9 months 
Daughton et al. (1992) hospital; total ban prospective cohort; no control –18.6% at 5 months +14% QR rate at 12 months; +82% at 24 months 
Offard et al. (1992) hospital; total ban retrospective; no control 30.2% decreased, 7.4% increased and 62.4% no change at 2 years –17.4% at 30 months; +22.5% QR at 30 months 
Brigham et al. (1994) hospital; total ban prospective cohort with control N/A NS QR at 4 weeks 
Jeffery et al. (1994) 32 worksites; ban in most areas versus no ban or designated area retrospective and prospective cohort with control –11.2% at 2 years; NS in sites without ban NS prevalence, QA or QR 
Olive and Ballard (1996) two federal hospitals; designated area versus total ban prospective cross‐sectional; no control Hospital 1: –13.7% at 6 months; –14.3% at 12 months; Hospital 2: –7.4% at 6 months Hospital 1: –20.4% at 12 months; +43% QR at 6 months, +70.6% at 1 year; Hospital 2: –18.5% at 6 months; +30.1% QR at 6 months 
Longo et al. (1996) hospital; total ban retrospective; logistic analysis N/A QR = 0.35, OR = 2.1 at 6 months; QR = 0.066, OR = 1.8 at 1 year; QR = 0.112, OR = 1.9 at 1.5 years; QR = 0.268, OR = 2.3 at 3 years; QR = 0.506, OR = 1.7 at 5 years 
Longo et al. (1999) hospital; total ban versus other non‐smoke‐free workplaces prospective cohort, with hazard rate models NA post‐ban QR range of 44% higher at 5 years (OR = 1.6), 98% higher at 6 years (OR = 2.4). 

Percent changes measured relative to initial rates, QA = quit attempts, QR = quit rate, NS = not significant (other results are significant), OR = odds ratio.

Table II.

Population‐based studies of effects of worksite restrictions on smoking

Study and year Methods Survey, year and (sample size)a Smoking restrictions Percent change in quantity smoked Percent change in prevalence unless otherwise indicated 
Woodruff et al. (1993) means tests and logistic regression California Tobacco Survey, 1990 (11 704) smoke‐free (SF), work area (WA), lesser restrictions (LR) versus no restrictions (NR) –13% everyday: –33%; OR: SF = 1.0, WA = 1.15, LR 1.36, NR NS; someday: NS  
Kinne et al. (1993) means tests Washington, 1989–1990 (1228) compared smoke‐free versus no restrictions –34% –48.8% 
Patten et al. (1995) means tests and logistic regression CA Tobacco Survey, 1990 versus 1992 (1844) compared smoke‐free versus not smoke‐free NS NS 
Glasgow et al. (1997) bivariate analyses; logistic and multiple regressions; adjusted for worksite cessation programs COMMIT, 1988 baseline and 1993 follow‐up (8271) compared smoke‐free (SF), designated areas (DA) versus not smoke‐free (NSF) SF: –2.78 cig/smoker, DA: –1.17 cig/smoker; QA: SF: OR = 1.27, DA: OR = 1.16, NSF: OR = 1; QR (6+ months): SF: OR = 1.27, DA: OR = 1.0 (NS), NSF = 1  
Farkas et al. (1999) logistic regressions (weighted) with home/ workplace restrictions Current Population Survey, 1993 (48 584) compared smoke‐free (SF), work area ban (WAB) versus not smoke‐free (NSF) (more likely to smoke >15 cigarettes per day) SF: OR = 1.53 WAB: OR = 1.10 (NS) QA in last year: SF: OR = 1.04 (NS), WAB: OR = 1.14; QR 6+ month: SF: 13.4%, OR = 1.21; WAB: 9.7%, OR = 0.93 NS 
Biener and Nyman (1999) logistic and multiple regression Massachusetts Tobacco Survey, 1993 baseline and 1996 follow‐up (369) compared continuously smoke‐free, became smoke‐free and not smoke‐free NA QR at 3 years: NSF: OR = 1; NS, except in workplaces with low ETS exposure. 
Evans et al. (2000) normalized probit or OLS estimate Current Population Survey, 1993 (97 882) compared work area bans (WAB) versus not smoke‐free (NSF) WAB: –14% mean difference –2.0 cigarettes WAB: –15%, mean difference –0.048 
Evans et al. (2000) normalized probit and OLS and two‐stage least squares National Health Interview Survey, 1991 (9704) and 1993 (8386) compared work area ban (WAB) versus not smoke‐free (NSF) WAB: –12% mean difference –2.5 cigarettes WAB: –19% mean difference –0.057 
Farrelly et al. (1999) normalized probit and OLS Current Population Survey, 1993 (97 882) compared smoke‐free (SF), work area ban and restricted common area (WARCA), partial work area and common area (PWACA) versus no ban SF: –14%,‐ 2.67 cigarettes; WARCA: –8%, –1.48 cigarettes; PWACA: –3%, –0.57 cigarettes absolute change: SF: –0.057; WARCA: –0.026; PWACA: 0.005 
Burns et al. (2001) logistic regression, adjusted for average level of state restrictions Current Population Survey, 1993 (16 041) and 1996 (13 422); daily smokers 1 year before survey; current every or some day smokers aged 25–64 compared smoke‐free to lesser restrictions  1993: QA: NS someday; NS former smokers OR = 1.18 QR 3 months OR = 1.39; similar results for 1996 except QA: OR = 1.09 
Study and year Methods Survey, year and (sample size)a Smoking restrictions Percent change in quantity smoked Percent change in prevalence unless otherwise indicated 
Woodruff et al. (1993) means tests and logistic regression California Tobacco Survey, 1990 (11 704) smoke‐free (SF), work area (WA), lesser restrictions (LR) versus no restrictions (NR) –13% everyday: –33%; OR: SF = 1.0, WA = 1.15, LR 1.36, NR NS; someday: NS  
Kinne et al. (1993) means tests Washington, 1989–1990 (1228) compared smoke‐free versus no restrictions –34% –48.8% 
Patten et al. (1995) means tests and logistic regression CA Tobacco Survey, 1990 versus 1992 (1844) compared smoke‐free versus not smoke‐free NS NS 
Glasgow et al. (1997) bivariate analyses; logistic and multiple regressions; adjusted for worksite cessation programs COMMIT, 1988 baseline and 1993 follow‐up (8271) compared smoke‐free (SF), designated areas (DA) versus not smoke‐free (NSF) SF: –2.78 cig/smoker, DA: –1.17 cig/smoker; QA: SF: OR = 1.27, DA: OR = 1.16, NSF: OR = 1; QR (6+ months): SF: OR = 1.27, DA: OR = 1.0 (NS), NSF = 1  
Farkas et al. (1999) logistic regressions (weighted) with home/ workplace restrictions Current Population Survey, 1993 (48 584) compared smoke‐free (SF), work area ban (WAB) versus not smoke‐free (NSF) (more likely to smoke >15 cigarettes per day) SF: OR = 1.53 WAB: OR = 1.10 (NS) QA in last year: SF: OR = 1.04 (NS), WAB: OR = 1.14; QR 6+ month: SF: 13.4%, OR = 1.21; WAB: 9.7%, OR = 0.93 NS 
Biener and Nyman (1999) logistic and multiple regression Massachusetts Tobacco Survey, 1993 baseline and 1996 follow‐up (369) compared continuously smoke‐free, became smoke‐free and not smoke‐free NA QR at 3 years: NSF: OR = 1; NS, except in workplaces with low ETS exposure. 
Evans et al. (2000) normalized probit or OLS estimate Current Population Survey, 1993 (97 882) compared work area bans (WAB) versus not smoke‐free (NSF) WAB: –14% mean difference –2.0 cigarettes WAB: –15%, mean difference –0.048 
Evans et al. (2000) normalized probit and OLS and two‐stage least squares National Health Interview Survey, 1991 (9704) and 1993 (8386) compared work area ban (WAB) versus not smoke‐free (NSF) WAB: –12% mean difference –2.5 cigarettes WAB: –19% mean difference –0.057 
Farrelly et al. (1999) normalized probit and OLS Current Population Survey, 1993 (97 882) compared smoke‐free (SF), work area ban and restricted common area (WARCA), partial work area and common area (PWACA) versus no ban SF: –14%,‐ 2.67 cigarettes; WARCA: –8%, –1.48 cigarettes; PWACA: –3%, –0.57 cigarettes absolute change: SF: –0.057; WARCA: –0.026; PWACA: 0.005 
Burns et al. (2001) logistic regression, adjusted for average level of state restrictions Current Population Survey, 1993 (16 041) and 1996 (13 422); daily smokers 1 year before survey; current every or some day smokers aged 25–64 compared smoke‐free to lesser restrictions  1993: QA: NS someday; NS former smokers OR = 1.18 QR 3 months OR = 1.39; similar results for 1996 except QA: OR = 1.09 

aUnless otherwise indicated, sample is indoor workers who smoked in the last year or when sample began.

Percent changes measured relative to initial rates, NS = not significant (other results are significant), OR = odds ratio.

Table III.

Studies of effects of clean air laws on smoking behavior, combining all age groups

Study and year Methods and outcome variable Samplea Measure of clean air laws Resultsb 
Emont et al. (1992) means and ordered alternatives, regression analysis; consumption, prevalence and quits (ratio of former to ever smokers) 1989 Current Population Survey for prevalence and quits; 1989 Tobacco Institute for consumption by state state‐level restrictiveness index means (extensive laws versus others): –14% prevalence rates; –12% per capita consumption; +12% quits; significant effects for moderate and extensive laws on quits. 
Chaloupka and Saffer (1992) linear, single and simultaneous equation; consumption 1975–85, Tobacco Institute, by state and year two state‐level indicator variables: public use (PU: 4 or more places), worksite (WS) single equation: –4 to –8% PU and –8% WS; simultaneous equation: –20% PU; simultaneity bias found for WS laws 
Chaloupka (1992) rational addiction regression model; consumption 1976–1980 NHANES II, individual four 0–1 state‐level indicators; variables based on USDHHS significant effects at basic and moderate levels but no additional effect at extensive level; greater effect on men than women 
Keeler et al. (1993) time series regression; consumption 1980–90 California Tobacco Survey, monthly, by community restrictiveness index for local anti‐smoking ordinances only significant effect of smoking restrictions, except when time trend included 
Yurekli and Zhang (2000) multiple regression; consumption 1970–95 Tobacco Institute annual consumption data by state state‐level restrictiveness index, weighted by time in effect and type –4.7% in 1995 due to existing laws, similar effects to a 10% price increase  
Moskowitz et al. (2000) logistic regression; quit in the last 6 months 1990 California Tobacco Survey, by community local ordinance: compared no law, weak, moderate and strong odds ratios: weak laws: 0.99; NS moderate laws: 1.38; NS strong laws: 1.61; similar results over various sociodemographic groups 
Study and year Methods and outcome variable Samplea Measure of clean air laws Resultsb 
Emont et al. (1992) means and ordered alternatives, regression analysis; consumption, prevalence and quits (ratio of former to ever smokers) 1989 Current Population Survey for prevalence and quits; 1989 Tobacco Institute for consumption by state state‐level restrictiveness index means (extensive laws versus others): –14% prevalence rates; –12% per capita consumption; +12% quits; significant effects for moderate and extensive laws on quits. 
Chaloupka and Saffer (1992) linear, single and simultaneous equation; consumption 1975–85, Tobacco Institute, by state and year two state‐level indicator variables: public use (PU: 4 or more places), worksite (WS) single equation: –4 to –8% PU and –8% WS; simultaneous equation: –20% PU; simultaneity bias found for WS laws 
Chaloupka (1992) rational addiction regression model; consumption 1976–1980 NHANES II, individual four 0–1 state‐level indicators; variables based on USDHHS significant effects at basic and moderate levels but no additional effect at extensive level; greater effect on men than women 
Keeler et al. (1993) time series regression; consumption 1980–90 California Tobacco Survey, monthly, by community restrictiveness index for local anti‐smoking ordinances only significant effect of smoking restrictions, except when time trend included 
Yurekli and Zhang (2000) multiple regression; consumption 1970–95 Tobacco Institute annual consumption data by state state‐level restrictiveness index, weighted by time in effect and type –4.7% in 1995 due to existing laws, similar effects to a 10% price increase  
Moskowitz et al. (2000) logistic regression; quit in the last 6 months 1990 California Tobacco Survey, by community local ordinance: compared no law, weak, moderate and strong odds ratios: weak laws: 0.99; NS moderate laws: 1.38; NS strong laws: 1.61; similar results over various sociodemographic groups 

aSample is national unless otherwise indicated.

bResults are in terms of percent change where able to determine.

NS = not significant (other results are significant), Restrictiveness index: 1.0 if private workplaces, 0.75 if restaurant restrictions (75%), 0.5 if four or more areas, 0.25 if one to three 3 areas. WP = workplace. PU = public use laws. Note: consumption refers to cigarettes per capita, quantity refers to cigarettes per smoker and prevalence refers to smoker status.

Table IV.

Studies of effects of clean air laws on smoking behavior, by age groups

Study and year Methods and outcome variable Samplea and age group Clean air measure Resultsb 
Wasserman (1991) pseudo‐poisson generalized linear consumption model; two‐part logistic prevalence and quantity models 1970–85 NHANES II (selected years); adults and teens state level RI change in consumption going from minimal to extensive laws: –6% for adults, –41% for teens; in two‐part models, only adult consumption and youth prevalence were significant 
Chaloupka and Grossman (1996) two‐part logistic models; 30‐day prevalence and quantity 1992, 1993 and 1994 Monitoring the Future; 8th, 10th and 12th graders state and local RI, and five category indicators, including school workplace and strong laws decrease prevalence; school laws decrease quantity smoked 
Chaloupka and Wechsler (1997) individual probit model for prevalence; ordered probit quantile and conditional demand for quantity 1993 Harvard College Alcohol Survey; college students state and local RI, and five category indicators, including school NS, except for other places affecting quantity smoked 
Lewitt et al. (1997) logistic regression; 30‐day prevalence and intention to smoke 1992 COMMIT; 9th graders state and adjusted local RI NS 
Chaloupka and Pacula (1999) probit and log linear models; daily and last 30‐day prevalence 1992, 1993 and 1994 Monitoring the Future; 8th, 10th and 12th graders state and local index; five category indicators, including school significant effects found only for men and whites  
Tauras and Chaloupka (1999) probit model for last 30‐day prevalence; log‐linear model for daily smoking 1976–93 Monitoring the Future; 12th graders, with follow‐ups through 1996 state and local RI; five category indicators, including school effects mostly negative or NS; stronger laws had larger effects 
Wakefield et al. (2000) probit model for last 30‐day prevalence; log‐linear model for daily smoking and stages of smoking uptake 1996 Robert Wood Johnson High School Students Special Survey state and local PU, and school; school and home bans based on questionnaire PU laws and home bans reduce smoking uptake and prevalence; school restrictions decrease prevalence and consumption 
Study and year Methods and outcome variable Samplea and age group Clean air measure Resultsb 
Wasserman (1991) pseudo‐poisson generalized linear consumption model; two‐part logistic prevalence and quantity models 1970–85 NHANES II (selected years); adults and teens state level RI change in consumption going from minimal to extensive laws: –6% for adults, –41% for teens; in two‐part models, only adult consumption and youth prevalence were significant 
Chaloupka and Grossman (1996) two‐part logistic models; 30‐day prevalence and quantity 1992, 1993 and 1994 Monitoring the Future; 8th, 10th and 12th graders state and local RI, and five category indicators, including school workplace and strong laws decrease prevalence; school laws decrease quantity smoked 
Chaloupka and Wechsler (1997) individual probit model for prevalence; ordered probit quantile and conditional demand for quantity 1993 Harvard College Alcohol Survey; college students state and local RI, and five category indicators, including school NS, except for other places affecting quantity smoked 
Lewitt et al. (1997) logistic regression; 30‐day prevalence and intention to smoke 1992 COMMIT; 9th graders state and adjusted local RI NS 
Chaloupka and Pacula (1999) probit and log linear models; daily and last 30‐day prevalence 1992, 1993 and 1994 Monitoring the Future; 8th, 10th and 12th graders state and local index; five category indicators, including school significant effects found only for men and whites  
Tauras and Chaloupka (1999) probit model for last 30‐day prevalence; log‐linear model for daily smoking 1976–93 Monitoring the Future; 12th graders, with follow‐ups through 1996 state and local RI; five category indicators, including school effects mostly negative or NS; stronger laws had larger effects 
Wakefield et al. (2000) probit model for last 30‐day prevalence; log‐linear model for daily smoking and stages of smoking uptake 1996 Robert Wood Johnson High School Students Special Survey state and local PU, and school; school and home bans based on questionnaire PU laws and home bans reduce smoking uptake and prevalence; school restrictions decrease prevalence and consumption 

aSample is by individual and national unless otherwise indicated.

bResults in terms of percent change where able to determine.

NS = not significant (other results are significant), RI = restrictiveness index: 1.0 if private workplaces, 0.75 if restaurant restrictions, 0.5 if four or more other public areas, 0.25 if one to three other public areas. WP = workplace. PU = public use laws. Note: consumption refers to cigarettes per capita, quantity refers to cigarettes per smoker and prevalence refers to smoker status.

References

Baile, W.F., Gibertini, M., Ulschak F., Snow‐Antle, S. and Hann, D. (
1991
) Impact of a hospital smoking ban: changes in tobacco use and employee attitudes.
Addictive Behaviors
 ,
16
,
419
–426.
Becker, D.M., Conner, H.F., Waranch, H.R., Stillman, F., Pennington, L., Lees, P.S.J. and Oski, F. (
1989
) The impact of a total ban on smoking in the Johns Hopkins Children’s Center.
Journal of the American Medical Association
 ,
262
,
799
–802.
Biener, L. and Nyman, A.L. (
1999
) Effect of workplace smoking policies on smoking cessation: a longitudinal study.
Journal of Occupational and Environmental Medicine
 ,
41
,
1121
–1127.
Biener, L., Abrams, D.B., Follick, M.J. and Dean, L. (
1989
) A comparative evaluation of a restrictive smoking policy in a general hospital.
American Journal of Public Health
 ,
79
,
192
–195.
Borland, R., Chapman, S., Owen, N. and Hill, D. (
1990
) Effects of workplace smoking bans on cigarette consumption.
American Journal of Public Health
 ,
80
,
178
–180.
Borland, R., Owen, N. and Hocking, B. (
1991
) Changes in smoking behaviour after a total workplace smoking ban.
Australian Journal of Public Health
 ,
15
,
130
–134.
Brenner, H. and Fleischle, B. (
1994
) Smoking regulations at the workplace and smoking behavior: a study from southern Germany.
Preventative Medicine
 ,
23
,
230
–234.
Brigham, J., Gross, J., Stitzer, M.L. and Felch, L.J. (
1994
) Effects of a restricted work‐site smoking policy on employees who smoke.
American Journal of Public Health
 ,
84
,
773
–778.
Broder, I., Pilger, C. and Corey, P. (
1993
) Environment and well‐being before and following smoking ban in office buildings.
Canadian Journal of Public Health
 ,
84
,
254
–258.
Brownson, R.C., Davis, J.R., Jackson‐Thompson, J. and Wilkerson, J.C. (
1995
) Environmental tobacco smoke awareness and exposure: impact of a statewide clean indoor air law and the report of the US Environmental Protection Agency.
Tobacco Control
 ,
4
,
132
–138.
Brownson, R.C., Eriksen, M.P., Davis, R.M. and Warner, K.E. (
1997
) Environmental tobacco smoke: health effects and policies to reduce exposure.
Annual Review of Public Health
 ,
18
,
163
–185.
Burns, D., Shanks, T., Major, J., Gower, K. and Shopland, D. (
2001
) Restrictions on smoking in the workplace. In Population Based Smoking Cessation Monograph 12. NIH publ. no. 00‐4804, 2000.39. National Institutes of Health, National Cancer Institute, Washington, DC.
Centers for Disease Control and Prevention, Office on Smoking and Health (
1986
) The Health Consequences of Involuntary Smoking: A Report of the Surgeon General. Government Printing Office, Washington, DC.
Centers for Disease Control and Prevention, Office on Smoking and Health (
1989
) Reducing the Health Consequences of Smoking: 25 Years of Progress. A Report of the Surgeon General, Executive Summary. Government Printing Office, Washington, DC.
Centers for Disease Control and Prevention (
1990
) Evaluation of an employee smoking policy—Pueblo, Colorado, 1989–1990.
Mortality and Morbidity Weekly Reports,
 
39
,
673
–676.
Centers for Disease Control and Prevention (
1999
) Best Practices for Comprehensive Tobacco Control Programs. Department of Health and Human Services, Atlanta, GA.
Centers for Disease Control and Prevention (
2000
) Intervention: Smoking Restrictions to Reduce Exposure to Environmental Tobacco Smoke. CDC, Atlanta, GA.
Centers for Disease Control and Prevention (
2002
) Cigarette smoking among adults—United States, 2000.
Morbidity and Mortality Weekly Report
 ,
51
,
642
–645.
Chapman, S., Borland, R., Scollo, M., Brownson, R.C., Dominello, A. and Woodward, S. (
1999
) The impact of smoke‐free workplaces on declining cigarette consumption in Australia and the United States.
American Journal of Public Health
 ,
89
,
1018
–1023.
Chaloupka, F.J. (
1992
) Clean indoor air laws, addiction, and cigarette smoking.
Applied Economics
 ,
24
,
193
–205.
Chaloupka, F.J. and Pacula, R.L. (
1999
) Sex and race differences in young people’s responsiveness to price and tobacco control policies.
Tobacco Control
 ,
8
,
373
–377.
Chaloupka, F.J. and Saffer, H. (
1992
) Clean indoor air laws and the demand for cigarettes.
Contemporary Policy Issues
 ,
10
,
72
–83.
Chaloupka, F.J. and Wechsler, H. (
1997
) Price, tobacco control policies, and smoking among young adults.
Journal of Health Economics
 ,
16
,
359
–373.
Daughton, D.M., Andrews, C.E., Orona, C.P., Patil, K.D. and Rennard, S.I. (
1992
) Total indoor smoking ban and smoker behavior.
Preventive Medicine
 ,
21
,
670
–676.
Department of Commerce (
1997
) Statistical Abstract of the United States 1997. Department of Commerce, Washington, DC.
Department of Health and Human Services (
2000
) Reducing Tobacco Use: A Report of the Surgeon General. Department of Health and Human Services, Centers for Disease Control and Prevention, Office of Smoking and Health, Atlanta, GA.
Emont, S.L., Choi, W.S., Novotny, T.E. and Giovino, G.A. (
1992
) Clean indoor air legislation, taxation and smoking behavior in the United States: an ecological analysis.
Tobacco Control
 ,
2
,
13
–17.
Environmental Protection Agency (
1992
) Respiratory Health Effects of Passive Smoking: Lung Cancer and other Disorders. EPA publ. no. EPA/600/6‐90/006F. Environmental Protection Agency, Office of Research and Development, Office of Air and Radiation, Washington, DC.
Eriksen, M.P. and Gottlieb, N.H. (
1998
) A review of the health impact of smoking control at the workplace.
American Journal of Health Promotion
 ,
13
,
83
–104.
Etter, J.F., Ronchi, A. and Perneger, T. (
1999
) Short‐term impact of a university based smoke free campaign.
Journal of Epidemiology and Community Health
 ,
53
,
710
–715.
Evans, W.N., Farrelly, M.C. and Montgomery, E. (
1999
) Do workplace smoking bans reduce smoking?
The American Economic Review
 ,
89
,
728
–747.
Farkas, A., Gilpin, E., Distefan, J. and Pierce, J.P. (
1999
) The effects of household and workplace smoking restrictions on quitting behaviors.
Tobacco Control
 ,
8
,
261
–265.
Farrelly, M.C., Evans, W.N. and Sfekas, A. (
1999
) The impact of workplace smoking bans: results from a national survey.
Tobacco Control
 ,
8
,
272
–277.
Fichtenburg, C.M. and Glantz, S.A. (
2002
) Effect of smoke‐free workplaces on smoking behaviour: systematic review.
British Medical Journal
 ,
325
, 188.
Gerlach, K.K., Shopland, D.R., Hartman, A.M., Gibson, J.T. and Pechacek, T.F. (
1997
) Workplace smoking policies in the United States: results from a national survey of more than 100000 workers.
Tobacco Control
 ,
6
,
199
–206.
Glantz, S.A. and Smith, L.R.A. (
1994
) The effect of ordinances requiring smoke‐free restaurants on restaurant sales.
American Journal of Public Health
 ,
84
,
1081
–1085.
Glantz, S.A. and Smith, L.R.A. (
1997
) The effect of ordinances requiring smoking‐free restaurants and bars on revenues: a follow‐up.
American Journal of Public Health
 ,
87
,
1687
–1692.
Glasgow, R.E., Cummings, K.M. and Hyland, A. (
1997
) Relationship of worksite policies to changes in employee tobacco use: findings from COMMIT.
Tobacco Control
 ,
S2
,
44
–48.
Goldstein, A.O. and Sobel, R.A. (
1998
) Environmental tobacco smoke regulations have not hurt restaurant sales in North Carolina.
North Carolina Medical Journal
 ,
59
,
284
–287.
Gottlieb, N.H., Eriksen, M.P., Lovato, C.Y., Weinstein, R.P. and Green, L.W. (
1990
) Impact of a restrictive work site smoking policy on smoking behavior, attitudes, and norms.
Journal of Occupational Medicine
 ,
32
,
16
–23.
Hedges, L.V. and Olkin, I. (
1985
) Statistical Methods for Meta‐Analysis. Academic Press, New York.
Heloma, A., Jaakkola, M.S., Kahkonen, E. and Reijula, K. (
2001
) The short‐term impact of national smoke‐free workplace legislation on passive smoking and tobacco use.
American Journal of Public Health
 ,
91
,
1416
–1418.
Hopkins, D.P., Briss, P.A., Ricard, C.J., Husten, C.G., Carande‐Kulis, V.G., Fielding, J.E., Alao, M.O., McKenna, J.W., Sharp, D.J., Harris, J.R., Woollery, T.A. and Harris, K.W. (
2001
) Reviews of evidence regarding interventions to reduce tobacco use and exposure to environmental tobacco smoke.
American Journal of Preventative Medicine
 ,
20
(2 Suppl.),
16
–66.
Hudzinski, L.G. and Frohlich, E.D. (
1990
) One‐year longitudinal study of a no‐smoking policy in a medical institution.
Chest
 ,
97
,
1198
–1202.
Hughes, J.R. (
2000
) Reduced smoking: an introduction and review of the evidence.
Addiction
 ,
95
(Suppl. 1),
S3
–S8.
Hyland, A., Cummings, K.M. and Wilson, M. (
1999
) Compliance with the New York City Smoke‐free Air Act.
Journal of Public Health Management and Practice
 ,
5
,
43
–52.
Jacobson, P.D. and Wasserman J. (
1997
) Tobacco Control Laws: Implementation and Enforcement. RAND, Santa Monica, CA.
Jeffery, R.W., Kelder, S.H., Forster, J.L., French, S.A., Lando, H.A. and Baxter, J.E. (
1994
) Restrictive smoking policies in the workplace: effects on smoking prevalence and cigarette consumption.
Preventive Medicine
 ,
23
,
78
–82.
Kinne, S., Kristal, A.R., White, E. and Hunt, J. (
1993
) Work‐site smoking policies: their population impact in Washington State.
American Journal of Public Health
 ,
83
,
1031
–1033.
Levy, D. and Friend, K. (
2001
) A framework for developing and evaluating clean indoor air laws.
Journal of Public Health Management and Practice
 ,
7
,
87
–97.
Lewitt, E.M., Hyland, A., Kerrebock, N. and Cummings, K.M. (
1997
) Price, public policy, and smoking in young people.
Tobacco Control
 ,
6
(Suppl. 2),
S17
–S30.
Lightwood, J.M. and Glantz, S.A. (
1997
) Short‐term economic and health benefits of smoking cessation: myocardial infarction and stroke.
Circulation
 ,
96
,
1089
–1096.
Longo, D.R., Brownson, R.C. and Johnson, J.C. (
1996
) Hospital smoking bans and employee smoking behavior: results of a national survey.
Journal of the American Medical Association
 ,
235
,
1252
–1257.
Longo, D.R., Johnson, J.C., Kruse, R.L., Brownson, R.C. and Hewett, J.E. (
2001
) A prospective investigation of the impact of smoking bans on tobacco cessation and relapse.
Tobacco Control
 ,
10
,
267
–272.
Moskowitz, J., Lin, Z. and Hudes, E. (
1999
) The impact of California’s smoking ordinances on worksite smoking policy and exposure to environmental tobacco smoke.
American Journal of Health Promotion
 ,
13
,
278
–281.
Moskowitz, J., Lin, Z. and Hudes, E. (
2000
) The impact of California’s smoking ordinances on smoking cessation.
American Journal of Public Health
 ,
90
,
57
–62.
Mullooly, J.P., Schuman, K.L., Stevens, V.J., Glasgow, R.E. and Vogt, T.M. (
1990
) Smoking behavior and attitudes of employees of a large HMO before and after a work site ban on cigarette smoking.
Public Health Reports
 ,
105
,
623
–628.
National Cancer Institute (
1997
) Changes in Cigarette‐Related Disease Risks and Their Implication for Prevention and Control. NIH publ. no. 97‐4213. National Cancer Institute, National Institutes of Health, Washington, DC.
National Cancer Institute (
1999
) Health Effects of Exposure to Environmental Tobacco Smoke: The Report of the California Environmental Protection Agency. NIH publ. no. 99‐4645. National Cancer Institute, National Institutes of Health, Washington, DC.
National Cancer Institute (
2000
) State and Local Legislative Action to Reduce Tobacco Use. Smoking and Tobacco Control Monograph 11. NIH publ. no. 00‐4804. National Cancer Institute, National Institutes of Health, Washington, DC.
Offord, K.P., Hurt, R.D., Berge, K.G., Frusti, D.K. and Schmidt, L. (
1992
) Effects of the implementation of a smoke‐free policy in a medical center.
Chest
 ,
102
,
1531
–1536.
Ohsfeldt, R.L., Boyle, R.G. and Capilouto, E.I. (
1998
) Tobacco Taxes, Smoking Restrictions and Tobacco Use. NBER Working Paper 6486. National Bureau of Economic Research, Cambridge, MA.
Olive, K.E. and Ballard, J.A. (
1996
) Changes in employee smoking behavior after implementation of restrictive smoking policies.
Southern Medical Journal
 ,
89
,
699
–706.
Patten, C.A., Gilpin, E., Cavin, S.W. and Pierce, J.P. (
1995
) Workplace smoking policy and changes in smoking behaviour in California: a suggested association.
Tobacco Control
 ,
4
,
36
–41.
Paulozzi, L.J., Spengler, R.F. and Gower, M.A. (
1992
) An evaluation of the Vermont worksite smoking law.
Public Health Reports
 ,
107
,
724
–726.
Pentz, M.A., Brannon, B.R., Charlin, V.L., Barrett, E.J., MacKinnon, D.P. and Flay, B.R. (
1989
) The power of policy: The relationship of smoking policy to adolescent smoking.
American Journal of Public Health
 ,
79
,
857
–862.
Petersen, L.R., Helgerson, S.D., Gibbons, C.M., Calhoun, C.R., Ciacco, K.H. and Pitchford, K.C. (
1998
) Employee smoking behavior changes and attitudes following a restrictive policy on worksite smoking in a large company.
Public Health Reports
 ,
103
,
115
–120.
Pierce, J.P., Shanks, T.G., Pertschuk, M., Gilpin, E., Shopland, D., Johnson, M. and Bal, D. (
1994
) Do smoking ordinances protect non‐smokers from environmental tobacco smoke at work?
Tobacco Control
 ,
3
,
15
–20.
Rigotti, N.A., Bourne, D., Rosen, A., Locke, J.A. and Schelling, T.C. (
1992
) Workplace compliance with a no‐smoking law: a randomized community intervention trial
American Journal of Public Health
 ,
82
,
229
–235.
Rigotti, N.A., Stoto, M.A., Bierer, M.F., Rosen, A., Schelling, T. (
1993
) Retail stores’ compliance with a city no‐smoking law.
American Journal of Public Health
 ,
83
,
227
–232.
Rigotti, N.A., Stoto, M.A. and Schelling, T.C. (
1994
) Do businesses comply with a no‐smoking law? Assessing the self‐enforcement approach.
Preventive Medicine
 ,
23
,
223
–229.
Rosenstock, I.M., Stergachis, A. and Heaney, C. (
1986
) Smoking regulations at the workplace and smoking behavior: a study from southern Germany.
American Journal of Public Health
 ,
76
,
1014
–1015.
Sciacca, J.P. (
1996
) A mandatory smoking ban in restaurants: concerns versus experiences.
Journal of Community Health
 ,
21
,
133
–150.
Sciacca, J. and Eckrem, M. (
1993
) Effects of a city ordinance regulation smoking in restaurants and retail stores.
Journal of Community Health
 ,
18
,
175
–182.
Scott, C.J. and Gerberich, S.G. (
1989
) Analysis of a smoking policy in the workplace.
AAOHN Journal
 ,
37
,
265
–273.
Shopland, D.R., Gerlach, K.K., Burns, D.M., Hartman, A.M. and Gibson, J.T. (
2001
) State‐specific trends in smoke‐free workplace policy coverage: the current population survey tobacco use supplement, 1993 to 1999.
Journal of Occupational and Environmental Medicine
 ,
43
,
680
–686.
Sorensen, G., Rigotti, N, Rosen, A., Pinney, J. and Prible, R. (
1991
) Effects of a worksite nonsmoking policy: evidence for increased cessation.
American Journal of Public Health
 ,
81
,
202
–204.
Sorensen, G., Lando, H. and Pechacek, T.F. (
1993
) Promoting smoking cessation at the workplace. Results of a randomized controlled intervention study.
Journal of Occupational Medicine
 ,
35
,
121
–126.
Sorenson, G., Beder, B., Prible, C.R. and Pinney, J. (
1995
) Reducing smoking at the workplace: implementing a smoking ban and hypnotherapy.
Journal of Occupational and Environmental Medicine
 ,
37
,
453
–460.
Sorenson, G., Glasgow, R.E., Topor, M. and Corbett, K. (
1997
) Worksite characteristics and changes in worksite tobacco‐control initiatives.
Journal of Occupational and Environmental Medicine
 ,
39
,
520
–526.
Stave, G.M. and Jackson, G.W. (
1991
) Effect of a total work‐site smoking ban on employee smoking and attitudes.
Journal of Occupational Medicine
 ,
33
,
884
–890.
Stephens, T., Pederson, L.L., Koval, J.J. and Kim, C. (
1997
) The relationship of cigarette prices and no‐smoking bylaws to the prevalence of smoking in Canada.
American Journal of Public Health
 ,
87
,
1519
–1521.
Stephens, T., Pederson, L.L., Koval, J.J. and Macnab, J. (
2001
) Comprehensive tobacco control policies and the smoking behaviour of Canadian adults.
Tobacco Control
 ,
10
,
317
–322.
Stillman, F.A., Becker, D.M., Swank, R.T., Hantula, D., Moses, H., Glantz, S. and Waranch, R. (
1990
) Ending smoking at the Johns Hopkins Medical Institutions. An evaluation of smoking prevalence and indoor air pollution.
Journal of the American Medical Association
 ,
264
,
1565
–1569.
Tauras, J.A. and Chaloupka, F.J. (
1999
) Price, Clean Indoor Air Laws, and Cigarette Smoking: Evidence from Longitudinal Data for Young Adults. NBER Working Paper 6937. National Bureau of Economic Research, Cambridge, MA.
Wakefield, M.A., Wilson, D., Owen, N., Esterman, A. and Roberts, L. (
1992
) Workplace smoking restrictions, occupational status, and reduced cigarette consumption.
Journal of Occupational Medicine
 ,
34
,
693
–697.
Wakefield, M.A., Chaloupka, F.J., Kaufman, N.J., Orleans, C.T., Barker, D.C. and Ruel, E.E. (
2000
) Effect of restrictions on smoking at home, at school and in public places on teenage smoking: cross sectional study.
British Medical Journal
 ,
725
,
333
–336.
Wasserman, J., Manning, W.G., Newhouse, J.P. and Winkler, J.D. (
1991
) The effects of excise taxes and regulations on cigarette smoking.
Journal of Health Economics
 ,
10
,
43
–64.
Woodruff, T.J., Rosbrook B., Pierce, J. and Glantz, S.A. (
1993
) Lower levels of cigarette consumption found in smoke‐free workplaces in California.
Archives of Internal Medicine
 ,
153
,
1485
–1493.
Yurekli, A.A. and Zhang, P. (
2000
) The impact of clean indoor‐air laws and cigarette smuggling on demand for cigarettes: an empirical model.
Health Economics
 ,
9
,
159
–170.

Author notes

Pacific Institute for Research and Evaluation, University of Baltimore, 11710 Beltsville Drive, Calverton, MD 20705, and 1Center for Alcohol and Addiction Studies, Brown University, Box G‐BH, Providence, RI 02912, USA

Comments

0 Comments