Objectives. In early 2000, the American Legacy Foundation launched the national “truth” campaign, the first national antismoking campaign to discourage tobacco use among youths. We studied the impact of the campaign on national smoking rates among US youths (students in grades 8, 10, and 12).

Methods. We used data from the Monitoring the Future survey in a pre/post quasi-experimental design to relate trends in youth smoking prevalence to varied doses of the “truth” campaign in a national sample of approximately 50000 students in grades 8, 10, and 12, surveyed each spring from 1997 through 2002.

Results. Findings indicate that the campaign accounted for a significant portion of the recent decline in youth smoking prevalence. We found that smoking prevalence among all students declined from 25.3% to 18.0% between 1999 and 2002 and that the campaign accounted for approximately 22% of this decline.

Conclusions. This study showed that the campaign was associated with substantial declines in youth smoking and has accelerated recent declines in youth smoking prevalence.

Mass media campaigns can be an effective public health strategy to prevent youth smoking.1–3 Antismoking television campaigns have emphasized diverse themes to discourage smoking, including highlighting short-and long-term health consequences, deglamorizing its social appeal through humorous and unflattering portrayals, and countering misperceptions that smoking is widespread among teens. A more recent theme, first used by California in the 1990s, focuses on exposing deceptive tobacco industry marketing practices and denials of tobacco’s health and addictive effects. In 1998, the Florida Department of Health launched a tobacco prevention program that featured a mass media campaign known as “truth” that countered industry influences with hard-hitting television advertisements that deglamorized smoking and portrayed youth confronting the tobacco industry. After 2 years, the prevalence of any past 30-day smoking among middle and high school students dropped by 40% and 18%, respectively.4 In a longitudinal study, Sly et al.5 linked exposure to the Florida “truth” campaign to declines in youth smoking prevalence.

As a result of the Master Settlement Agreement between tobacco companies and 46 states, the American Legacy Foundation (Legacy) initiated the national “truth” campaign in February 2000. From 2000 to 2002, annual funding for the campaign averaged $100 million per year. A national media purchase was employed by the campaign, as opposed to a randomized exposure design, for 2 primary reasons. First, Legacy could not ethically assign some media markets to low or zero exposure, given the documented successes of the Florida “truth” campaign. Second, a national media purchase was roughly 40% cheaper than a market-to-market purchase, which would have been necessary to randomize exposure. Although the “truth” campaign builds upon the experiences of Florida and other state campaigns, no similar large-scale national antismoking effort has occurred since the period of the Fairness Doctrine from 1967 to 1970, when TV networks were required to maintain a balance between anti-and prosmoking ads. The “truth” campaign ads are designed to avoid overt and directive messages that tell teens not to smoke and instead use graphic images depicting stark facts about death and disease caused by tobacco and exposés of manipulative marketing practices. For example, an early commercial, “Body Bags,” showed youths piling 1200 body bags outside a major tobacco company’s headquarters to highlight the daily death toll from tobacco use.

This is the first study to evaluate the behavioral outcomes of the campaign. Previous studies have shown that the campaign influenced campaign-related attitudes toward tobacco use and the tobacco industry and that negative attitudes about the tobacco industry are correlated with reduced risk of smoking.6–9 The current study assessed whether there was a dose–response relationship between the level of exposure to the campaign and youth smoking prevalence during the first 2 years of the campaign.

Study Design Overview

This study used a pre/post quasi-experimental design that related changes in youth smoking prevalence to varied exposure to the campaign over time and across media markets in the United States; secular trends in smoking prevalence and other confounding influences were controlled. The years 1997–1999 represent a precampaign study period, and although the campaign was launched nationally, the dose of campaign messages varied considerably across media markets and over time from 2000 to 2002. The primary source of variation in exposure is the presence or absence of local affiliates for 1 or more of the broadcast networks (e.g., FOX, UPN, WB) on which “truth” campaign commercials aired and variation in cable television market presence (e.g., MTV).

Study Population

Our study used data from the 1997–2002 Monitoring the Future (MTF) annual spring surveys, designed to monitor alcohol, tobacco, and illicit drug use among youths in the United States.10 The survey, funded primarily by the National Institute on Drug Abuse and conducted by the University of Michigan’s Institute for Social Research, included approximately 18 000, 17 000, and 16000 8th-, 10th-, and 12th-grade students per year, respectively.

MTF surveys are conducted in approximately 420 public and private secondary schools per year and use a multistage random sampling design to provide nationally representative samples of students in each grade level. The sampling procedure selects geographic areas in stage 1, 1 or more schools within those geographic areas in stage 2, and classes within each school in the final stage. Schools are selected such that their probability of selection is proportionate to the size of the classes being sampled. Sample weights are applied to all respondents in the survey to adjust for school differences in probabilities of selection and school size. Up to 350 students are surveyed from each selected school. All surveys are self-administered in school classrooms during normal class periods. The average yearly student response rates for 8th-, 10th-, and 12th-grade students in the 1997–2002 MTF surveys were 89.0%, 86.2%, and 82.8%, respectively.

The primary study outcome was a dichotomous indicator for reporting any quantity of smoking in the past 30 days, based on the question “How frequently have you smoked cigarettes during the past 30 days?” Our indicator variable equaled zero for students who responded “none per day” and 1 for students who responded less than 1 cigarette, 1 to 5 cigarettes, about 1/2 pack, about 1 pack, about 1 1/2 packs, and 2 packs or more per day.


We used cumulative gross rating points (GRPs) for the “truth” campaign in each of the 210 television markets in the United States to measure each student’s exposure to the campaign (note that MTF surveys do not ask about awareness of the “truth” campaign). The GRPs measure the total volume of delivery of a media campaign to a target audience. It is equal to the percentage of the target audience that is reached by the campaign times the frequency of exposure.

To illustrate the variation in potential exposure, we grouped the 210 media markets into 1 of 5 levels of exposure, on the basis of the range in total GRPs (647 to 22 389) that accumulated in each market from campaign launch in February 2000 until the second quarter of 2002 (Figure 1). The lowest-exposure group received an average of 3867 GRPs over this 2-year period, whereas the highest-exposure group received an average of 20367 GRPs. Market-level variation in GRPs is primarily due to the availability of television stations on which “truth” campaign advertisements aired (e.g., FOX, UPN, WB). Markets with fewer stations received lower GRPs, whereas markets with more stations received higher GRPs.

A student’s exposure was defined as the cumulative number of “truth” campaign GRPs that were delivered in a school’s media market from the beginning of the campaign to the time of each spring survey in 2000, 2001, and 2002. Consistent with our pre/post evaluation design, we included students surveyed from 1997 to 1999 to serve as a historical unexposed (GRP = 0) comparison group. To account for the possibility that markets that received relatively high doses of the campaign might experience diminishing returns to additional GRPs, we included a quadratic term for cumulative GRPs in our models.

Potential Confounders

Following a socioecological model11 that recognizes multiple levels of influence on health behaviors (e.g., intrapersonal, interpersonal, community, media, policy, economic factors), we controlled for a wide array of potential confounding influences described in the following sections.

Individual Level.

Our multivariable models included individual-level data from the MTF such as grade, race/ethnicity, gender, parental education, and weekly income. We created indicator variables for grade, race/ethnicity (African American, Hispanic, Asian, other race, with White as the reference), gender, and parental education for mother and father separately (high school graduate, at least some college, with less than high school diploma as the reference category). To account for missing data on race and parental education, we included indicator variables for those with unspecified race and father’s and mother’s education in the MTF survey. We also included a measure of students’ weekly income, based on 2 MTF survey questions that assess how much money students earn during an average week from a job and from other sources such as allowances. To adjust for inflation, we used the 2002 consumer price index to express our measure of income in 2002 dollars.

Media Market Level.

In light of the source of variation in the media market dose of the “truth” campaign, there might be factors that determined both the dose of the “truth” campaign and the level of smoking at the media market level. For example, low-exposure markets tended to be more rural, White, and less educated and have lower incomes—all factors associated with smoking—than markets with high campaign exposure. Failing to control for these factors could lead to a spurious negative correlation between campaign exposure and smoking prevalence. We implemented 2 approaches to statistically model possible correlations between preexisting media market smoking rates and the subsequent campaign dose. We first treated each of the 210 media markets as fixed effects in a logistic regression model that included indicator variables for 209 of 210 media markets (with 1 market as a reference).12 The fixed-effects approach was equivalent to controlling for average market-level smoking rates, effectively making each market its own control group. Our second approach included direct media market–level measures of potential confounders—2002 data on the median household income, percentage of the population who were college graduates, and population size.

State Level.

To account for potential state-level influences, we collected data on inflation-adjusted cigarette prices13 and investments in tobacco control programs corresponding to the 1997–2002 MTF and the location of a student’s school. Previous research has shown that cigarette prices and state tobacco control programs influence youth smoking prevalence.4,14,15 Our measure of tobacco control investments is based on state per capita tobacco control program funding derived from Centers for Disease Control and Prevention State Highlight reports16–18 and supplemented by data from state programs.

Analytic Approach
Descriptive Statistics.

We began by examining the overall change in the prevalence of youth smoking from 1997 to 2002 overall and by grade. We also compared the annual rate of change for the period leading up to the campaign (1997–1999) and during the campaign (2000–2002) as a simple indication of whether declines in youth smoking prevalence appeared to accelerate after campaign launch.

Multivariable Logistic Regression.

To more precisely isolate the association between current youth smoking prevalence and “truth” campaign exposure, we used population average logistic regression models to estimate current youth smoking prevalence as a function of individual-, media market–, and state-level influences (noted earlier). We also included a linear time variable, taking values from 0 to 6 for the MTF years from 1997 to 2002, to control for the national downward trend in the prevalence of youth smoking that began in 1997 in order to isolate the effects of the campaign from the national trend. Students from the 1997–1999 surveys served as an unexposed comparison group.

We estimated all regressions by combining the cross-sectional 1997–2002 MTF surveys to relate the odds that an individual smoked to his or her media market dose of the campaign, measured at the time of the survey. Cumulative campaign GRPs were scaled such that the estimated odds ratios indicated the odds of smoking, given an increase of 10000 GRPs. All models were estimated separately for 8th, 10th, and 12th grades and all grades combined. All analyses were estimated with sampling weights that corrected for nonresponse and sample design. Standard errors were adjusted for clustering at the school level (schools were the primary sampling unit) using Stata’s (version 8.0; Stata Corp, College Station, Tex) SVYLOGIT command. We estimated 2 sets of logistic regressions: 1 with media market indicator variables and 1 with the specific market-level variables—median household income, percentage of the population who were college graduates, and average population size of telephone exchanges within each media market.

A final set of regressions examined the differential effects of the campaign in 2000, 2001, and 2002 to test the hypothesis that the campaign effects on smoking rates in spring 2001 and 2002 were substantially greater than campaign effects measured in the spring 2000 MTF survey—only a few months into the campaign.

We employed 2 methods to illustrate the relationship between youth smoking rates and campaign exposure from the logistic regressions. First, to illustrate to what extent there was evidence of diminishing campaign effects at increasing levels of exposure, we plotted the combined odds ratio for the linear (GRP) and quadratic (GRP2) exposure variables over a range of GRPs. Second, the inclusion of a precampaign period (1997–1999) allowed us to predict the trend in youth smoking prevalence in the absence of the campaign and hence to estimate the proportion of the decline in youth smoking prevalence attributable to the “truth” campaign after 1999. To do this, we estimated the probability of smoking for each year on the basis of the multivariable logistic regressions, setting campaign exposure to zero in each postcampaign year for all youths. The difference between the predicted and the actual smoking rates indicated how much lower smoking rates were as a result of the campaign.

The MTF data showed a large decline in current youth smoking prevalence overall and for each grade between 1997 and 2002 (Table 1). Among all grades combined, current smoking prevalence decreased by 36% from 1997 to 2002. Eighth-grade students exhibited the largest percentage decline during this period at 45%, whereas 12th-grade students showed the smallest decline at 27%. The descriptive MTF data also indicated that the decline in current smoking prevalence accelerated after the launch of the campaign between 2000 and 2002 (Table 1). The annual percentage decline for all grades was 3.2% before the campaign launch (1997–1999) compared with 6.8% after the campaign launch (2000–2002). T tests, based on the observed differences in ratios and a Taylor series approximation of the standard errors of these differences, showed that the post–“truth” campaign annual declines were significantly greater than the pre–“truth” campaign annual declines overall and by grade (P<.001). As shown later, the accelerated decline likely occurred in the latter 2 years of the campaign as a lagged effect.

Results of the logistic regression for all grades indicated that there was a statistically significant dose–response relationship between “truth” campaign exposure and current youth smoking prevalence (odds ratio [OR] = 0.78; 95% confidence interval [CI] = 0.63, 0.97; P<.05) (Table 2). The odds ratio for the quadratic GRPs provides evidence that the effect diminished at higher levels of exposure (OR = 1.11; 95% CI = 1.00, 1.25; P<.07). Figure 2 illustrates the overall relationship between youth smoking prevalence and “truth” campaign exposure between 2000 and 2002. As shown, the effect of the campaign continued to increase through 10000 GRPs and then began to attenuate as markets experienced higher average cumulative doses of the campaign but the odds ratio remained below 1. This suggested that the campaign could have a larger overall impact if it were feasible to redistribute GRPs from the highest-exposure markets to those with relatively low exposure.

Results calculated from the data presented in Table 2 indicate that between 1999 and 2002, the prevalence of smoking among students in all grades combined would have declined by only 5.7 percentage points to 19.6% (95% CI = 18.6%, 20.6%) instead of the actual decline of 7.3 percentage points to 18.0% had the campaign not existed. Therefore, roughly 22% (95% CI = 8.2%, 35.6%) of the total decline in youth smoking prevalence between 1999 and 2002 is attributable to the campaign.

In addition, as hypothesized, there was no statistically significant relationship between overall youth smoking prevalence and the campaign after only a few months of the campaign in 2000, but the effect was statistically significant in 2001 (OR = 0.66; 95% CI = 0.45, 0.98; P<.05) and 2002 (OR = 0.63; 95% CI = 0.45, 0.88; P<.01) (Table 2). To further illustrate the increasing effects of the campaign over time, we calculated the odds of smoking, based on the estimated odds ratios from Table 2, at varied levels of cumulative campaign exposure during each of the postlaunch cumulative periods (Figure 2). These results suggested that the relationship between overall youth smoking prevalence and the campaign strengthened over time and, as expected, the campaign showed little effect in 2000.

Separate regressions by grade show the largest effects for 8th-grade students (OR = 0.61; P<.05), followed by statistically non-significant effects for 12th- (OR = 0.79; P=.198) and 10th- (OR = 0.98; P=.884) grade students, respectively (Table 2). We also estimated a set of regressions excluding the quadratic GRP term (GRP2) (results available on request). In this set, the effect was marginally statistically significant for 12th-grade students (OR = 0.879; P<.07) but statistically nonsignificant overall and for 8th- and 10th-grade students.

The dose–response relationship for the “truth” campaign is robust across the alternative set of models that control for potential confounding of media market characteristics, such as median household income, percentage of the population who were college graduates, average population size of telephone exchanges within each media market, and state-specific indicator variables (results available on request).

The “truth” campaign was associated with significant declines in youth smoking prevalence; thus, its approach to appeal to youths with hard-hitting ads that show at-risk youths rejecting tobacco and that reveal deceptive tobacco industry marketing tactics appears to be effective. Previous research revealed that in its first year, the campaign reached three fourths of American youths and was associated with campaign-related attitudes toward tobacco and the tobacco industry among youths. The current results further validate these early markers of the campaign’s success. These findings are consistent with previous research on the effectiveness of antismoking campaigns in general1,2 and recent state tobacco industry manipulation campaigns.5,19

In addition to being consistent with previous findings, this study improves on previous research by reaching generalized conclusions about the effects of antismoking campaigns for youths across the United States and by implementing a pre/post quasi-experimental design that controlled for potential threats to validity, such as secular trends in smoking prevalence, the influence of cigarette prices, state tobacco control programs, and other factors.

Descriptive statistics show that smoking rates declined faster after the launch of the campaign. More significant, this result was confirmed in multivariable analyses that controlled for confounding influences and indicated a dose–response relationship between “truth” campaign exposure and current youth smoking prevalence. To address concerns over the environmental nature of the “truth” campaign exposure measure, we performed sensitivity analyses that showed internally consistent and intuitive results—no effect in the early months of the campaign, diminishing returns, and no statistically significant association between the campaign and drinking among youths (described later). We found that by 2002, smoking rates overall were 1.5 percentage points lower than they would have been in the absence of the campaign, which translates to roughly 300 000 fewer youth smokers based on 2002 US census population statistics. To put our findings in perspective, research indicated that youth smoking prevalence declined by about 1 percentage point per year between 1967 and 1970 during the period of the Fairness Doctrine.20 It is important to note that these results may also reflect residual impacts from the 1964 Smoking and Health: Report of the Advisory Committee to the Surgeon General of the Public Health Service21 and federal government policies requiring health warnings on all cigarette packages and in all cigarette advertising.

Our findings provide some evidence that the campaign may have the largest impact among 8th-grade students, which is consistent with evidence from Florida that indicates the Florida “truth” campaign led to declines in smoking rates and that smoking rates declined by 50% among middle school students (grades 6 through 8) and by 35% among high school students (grades 9 through 12) from 1998 to 2002.

Our analyses are not without their limitations. Our measures of youth smoking prevalence are self-reported and may be subject to social desirability bias so that youths are less likely to report smoking in areas with high exposure to the campaign than in areas with lower exposure. This would lead to an overstatement of the campaign’s effects. However, previous studies have found that underreporting of smoking by youths is minimal.22–24 Our results also rely on repeated cross-sectional surveys, not repeated measures on the same students, which weakens the strength of our causal inference. However, we included youths surveyed before 2000, so students in the 1997–1999 surveys served as an unexposed control group.

Finally, it is possible that the estimated campaign effects may have been due to other unmeasured youth-focused prevention activities (e.g., in-school substance abuse–prevention programs, the national antidrug campaign by the Office of National Drug Control Policy) that were correlated by chance with the “truth” campaign exposure. To assess this possibility, we used data from the 2000 and 2002 National Youth Tobacco Surveys to examine the correlation between the “truth” campaign exposure and exposure to tobacco use prevention education in schools. In addition, if there was a spurious correlation between “truth” campaign exposure and other prevention activities, we would also have expected to find a correlation between the “truth” campaign and other risk behaviors such as underage drinking. We addressed this potential problem by estimating a series of models identical to those presented in Table 2, using indicator variables for any drinking in the past 30 days and any binge drinking in the past 2 weeks as outcome variables.

Although exposure to multistrategy tobacco use prevention education programs in schools has been linked to lower smoking prevalence among middle school students,25,26 we found no statistically significant differences in reported exposure to tobacco use prevention education programs in schools and exposure to the “truth” campaign. We tested this relationship by comparing self-reported recall of tobacco use prevention education programs in schools from the National Youth Tobacco Surveys by grouping students into the same 5 levels of “truth” campaign exposure illustrated in Figure 1.

In addition, we did not find any evidence of a relationship between “truth” campaign exposure and any drinking in the past 30 days (OR = 0.981; P=.848) nor any relationship between “truth” campaign exposure and any binge drinking within the past 2 weeks (OR = 0.857; P=.189), suggesting that “truth” campaign exposure is not spuriously correlated with other prevention efforts. Our results further suggest that the measured “truth” campaign effects on smoking prevalence are not the result of other efforts such as in-school tobacco use prevention education programs and prevention activities aimed at other risk behaviors such as underage drinking.

Under the Master Settlement Agreement, the tobacco industry was obligated to fund Legacy’s Public Education Fund for 5 years (through 2003) and is obligated thereafter in any year in which the tobacco companies participating in the Master Settlement Agreement achieve a combined 99.05% market share of US tobacco sales. The continuation of Legacy’s efforts, including the “truth” campaign, is presently in question because of these terms. Our findings are consistent with those of other studies that demonstrate that effective antismoking campaigns are critical for public health and that their elimination will likely erase gains that have been made to date in reducing youth smoking prevalence.27–31

TABLE 1— Changes in Current Smoking Prevalence Among US Students Before and After the Launch of the “truth” Campaign in 2000: 1997–2002
TABLE 1— Changes in Current Smoking Prevalence Among US Students Before and After the Launch of the “truth” Campaign in 2000: 1997–2002
 Prevalence of Current Smoking, %Average Annual Percentage Change (95% Confidence Interval)
All28.018.0−35.7−3.2 (−3.8, −2.6)−6.8 (−7.5, −6.1)
8th19.410.7−44.8−3.4 (−4.6, −2.1)−9.0 (−10.4, −7.6)
10th29.817.7−40.6−4.6 (−5.6, −3.6)−8.7 (−9.8, −7.5)
12th36.526.7−26.8−1.8 (−2.7, −1.0)−5.1 (−6.1, −3.9)
TABLE 2— Impact of “truth” Campaign on Current Smoking Prevalence Among US Students: Monitoring the Future, 1997–2002
TABLE 2— Impact of “truth” Campaign on Current Smoking Prevalence Among US Students: Monitoring the Future, 1997–2002
 All Grades, Odds Ratios (95% Confidence Intervals)1997–2002 by Grade, Odds Ratios (95% Confidence Intervals)
 1997–20021997–1999 + 20001997–1999 + 20011997–1999 + 20028th Grade10th Grade12th Grade
“truth” exposure variables
    “truth” exposure (GRP)0.78** (0.63, 0.97)0.90 (0.50, 1.62)0.66** (0.45, 0.98)0.63*** (0.45, 0.88)0.61** (0.39, 0.94)0.98 (0.73, 1.31)0.79 (0.56, 1.13)
    “truth” exposure squared (GRP2)1.11* (1.00, 1.25)0.90 (0.34, 2.38)1.29 (0.94, 1.77)1.25** (1.04, 1.51)1.29** (1.01, 1.66)0.98 (0.84, 1.14)1.06 (0.87, 1.29)
Other explanatory variables
    Time0.89*** (0.86, 0.94)0.92*** (0.87, 0.98)0.93*** (0.87, 0.98)0.92*** (0.87, 0.97)0.88*** (0.81, 0.96)0.90*** (0.85, 0.95)0.94** (0.88, 0.99)
    Grade 101.50*** (1.40, 1.61)1.47*** (1.34, 1.60)1.48*** (1.36, 1.62)1.44*** (1.33, 1.57)
    Grade 121.97*** (1.84, 2.10)1.84*** (1.71, 1.99)1.89*** (1.76, 2.04)1.90*** (1.77, 2.04)
    African American0.28*** (0.26, 0.30)0.26*** (0.24, 0.29)0.28*** (0.26, 0.30)0.28*** (0.26, 0.30)0.37*** (0.33, 0.42)0.27*** (0.24, 0.31)0.24*** (0.21, 0.27)
    Hispanic0.69*** (0.64, 0.74)0.68*** (0.63, 0.74)0.71*** (0.65, 0.77)0.71*** (0.65, 0.76)0.88** (0.79, 0.98)0.59*** (0.54, 0.67)0.64*** (0.57, 0.71)
    Asian0.51*** (0.46, 0.56)0.48*** (0.43, 0.54)0.51*** (0.46, 0.57)0.51*** (0.46, 0.57)0.49*** (0.41, 0.59)0.55*** (0.47, 0.65)0.51*** (0.44, 0.58)
    Other race/ethnicity0.91*** (0.86, 0.97)0.89*** (0.83, 0.95)0.92** (0.87, 0.99)0.93** (0.87, 0.99)0.95 (0.87, 1.04)0.92* (0.84, 1.00)0.86*** (0.77, 0.95)
    Male0.91*** (0.88, 0.93)0.90*** (0.87, 0.94)0.89*** (0.86, 0.93)0.89*** (0.86, 0.93)0.88*** (0.83, 0.94)0.83*** (0.79, 0.87)0.99 (0.94, 1.04)
    Weekly income1.01*** (1.01, 1.01)1.01*** (1.01, 1.01)1.01*** (1.01, 1.01)1.01*** (1.01, 1.01)1.01*** (1.01, 1.01)1.01*** (1.01, 1.01)1.01*** (1.01, 1.01)
    Mother is high school graduate0.87*** (0.83, 0.91)0.86*** (0.82, 0.91)0.89*** (0.85, 0.94)0.88*** (0.83, 0.92)0.76*** (0.69, 0.82)0.88*** (0.82, 0.94)1.02 (0.95, 1.10)
    Mother had some college0.82*** (0.78, 0.86)0.82*** (0.78, 0.87)0.84*** (0.79, 0.88)0.82*** (0.78, 0.87)0.65*** (0.59, 0.70)0.78*** (0.73, 0.84)1.06 (0.99, 1.14)
    Father is high school graduate0.72*** (0.75, 0.82)0.80*** (0.76, 0.84)0.80*** (0.76, 0.84)0.79*** (0.75, 0.83)0.70*** (0.65, 0.76)0.77*** (0.61, 0.71)0.88*** (0.81, 0.95)
    Father had some college0.67*** (0.64, 0.70)0.68*** (0.64, 0.72)0.70*** (0.66, 0.74)0.69*** (0.65, 0.73)0.53*** (0.49, 0.58)0.66*** (0.61, 0.71)0.79*** (0.74, 0.86)

Note. GRP = gross rating point. All models include sampling weights, and standard errors were adjusted for school-level clustering. Control variables included media market indicator variables, inflation-adjusted state per capita tobacco control expenditures, state-level average cigarette prices, and indicator variables for missing or unspecified race and parental education data. Odds ratios for the “truth” campaign GRPs represent the odds of smoking for every increase in GRPs by 10 000.

*P<.10; **P<.05; ***P<.01.

This study was supported by the American Legacy Foundation.

We express our appreciation to Lloyd Johnson and Patrick O’Malley as the principal investigators of Monitoring the Future (MTF) for their cooperation in providing timely access to the MTF data. We also thank Timothy Perry for his contributions to analysis of the MTF data and Susan Murchie for editorial review.

Human Participant Protection The University of Michigan institutional review board approved the MTF study and the consent information provided to the respondents. No protocol approval was needed for the analysis of the MTF data.


1. Farrelly MC, Niederdeppe J, Yarsevich J. Youth tobacco prevention mass media campaigns: past, present, and future directions. Tob Control. 2003;12(suppl 1): i35–i47. Crossref, MedlineGoogle Scholar
2. Siegel M. Antismoking advertising: figuring out what works. J Health Commun. 2002;7(2):157–162. Crossref, MedlineGoogle Scholar
3. Wakefield M, Flay B, Nichter M, Giovino G. Effects of anti-smoking advertising on youth smoking: a review. J Health Commun. 2003;8(3):229–247. Crossref, MedlineGoogle Scholar
4. Bauer UE, Johnson TM, Hopkins RS, Brooks RG. Changes in youth cigarette use and intentions following implementation of a tobacco control program: findings from the Florida Youth Tobacco Survey, 1998–2000. JAMA. 2000;284:723–728. Crossref, MedlineGoogle Scholar
5. Sly DF, Hopkins RS, Trapido E, Ray S. Influence of a counteradvertising media campaign on initiation of smoking: the Florida “truth” campaign. Am J Public Health. 2001;91:233–238. LinkGoogle Scholar
6. Farrelly MC, Healton CG, Davis KC, Messeri P, Hersey JC, Haviland ML. Getting to the “truth”: evaluating national tobacco countermarketing campaigns. Am J Public Health. 2002;92:901–907. LinkGoogle Scholar
7. Niederdeppe J, Farrelly MC, Haviland ML. Confirming “truth”: more evidence of a successful tobacco countermarketing campaign in Florida. Am J Public Health. 2004:94:255–257. LinkGoogle Scholar
8. Evans WD, Price S, Blahut S, Ray S, Hersey JC, Niederdeppe J. Social imagery, tobacco independence, and the truth® campaign. J Health Commun. 2004;9: 425–441. Crossref, MedlineGoogle Scholar
9. Hersey JC, Niederdeppe J, Evans WD, et al. The effects of state counterindustry media campaigns on beliefs, attitudes, and smoking status among teens and young adults. Prev Med. 2003:37(6 pt 1):544–552. Crossref, MedlineGoogle Scholar
10. Johnston LD, Bachman JG, O’Malley PM, Schulenberg J. Monitoring the Future: A Continuing Study of American Youth (8th, 10th, and 12th Grade Surveys). Ann Arbor, Mich: Inter-University Consortium for Political and Social Research; 2003. Google Scholar
11. Glanz K, Lewis FM, Rimer BK. Health Behavior and Health Education. San Francisco, Calif: Jossey-Bass Publishers; 1997. Google Scholar
12. Heckman JJ, Hotz VJ. Choosing among alternative nonexperimental methods for estimating the impact of social programs: the case of manpower training. J Am Stat Assoc. 1989;84:862–880. CrossrefGoogle Scholar
13. Tax Burden on Tobacco. Historical Compilation. Vol 37. Arlington, Va: Orzechowski & Walker; 2002. Google Scholar
14. Reducing Tobacco Use: A Report of the Surgeon General. Atlanta, Ga: National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2000. Google Scholar
15. Farrelly MC, Pechacek TF, Chaloupka FJ. The impact of tobacco control program expenditures on aggregate cigarette sales: 1981–2000. J Health Econ. 2003; 22:843–859. Crossref, MedlineGoogle Scholar
16. State Tobacco Control Highlights 1999. Atlanta, Ga: National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2002. Google Scholar
17. Investment in Tobacco Control: State Highlights 2001. Atlanta, Ga: National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2002. Google Scholar
18. Tobacco Control State Highlights 2002: Impact and Opportunity. Atlanta, Ga: National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2002. Google Scholar
19. Sly DF, Trapido E, Ray S. Evidence of the dose effects of an antitobacco counteradvertising campaign. Prev Med. 2002;35:511–518. Crossref, MedlineGoogle Scholar
20. Lewit EM, Coate D, Grossman M. The effects of government regulation on teenage smoking. J Law Econ. 1981;24:545–569. CrossrefGoogle Scholar
21. Smoking and Health: Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington, DC: US Department of Health, Education and Welfare; 1964. Google Scholar
22. Bauman KE, Koch GG. Validity of self-reports and descriptive and analytical conclusions: the case of cigarette smoking by adolescents and their mothers. Am J Epidemiol. 1983;118(1):90–98. Crossref, MedlineGoogle Scholar
23. Bauman KE, Koch GG, Bryan ES. Validity of self-reports of adolescent cigarette smoking. Int J Addict. 1982;17:1131–1136. Crossref, MedlineGoogle Scholar
24. Messeri P, Haviland ML, Mowery P, Gable J, Farrelly MC. A biochemical validation study to assess effects of the “truth” campaign on truthful reporting of current smoking among high school students. Paper presented at the 130th Annual Meeting of the American Public Health Association; November 2002; Philadelphia, Pa. Google Scholar
25. Peterson AV Jr, Kealy KA, Mann SL, Marek PM, Srason IG. Hutchinson Smoking Prevention Project: long-term randomized trial in school-based tobacco use prevention—results on smoking. J Natl Cancer Inst. 2000;92:1979–1991. Crossref, MedlineGoogle Scholar
26. Wenter DL, Blackwell S, Davis KC, Farrelly MC. Legacy First Look Report 8: Using Multiple Strategies in Tobacco Use Prevention Education. Washington, DC: American Legacy Foundation; 2002. Google Scholar
27. Glantz SA. Changes in cigarette consumption, prices, and tobacco industry revenues associated with California’s Proposition 99. Tob Control. 1993;2: 311–314. CrossrefGoogle Scholar
28. Pierce JP, Gilpin EA, Emery SL, et al. Has the California tobacco control program reduced smoking? JAMA. 1998;280:893–899. Crossref, MedlineGoogle Scholar
29. Centers for Disease Control and Prevention. Effect of ending an antitobacco youth campaign on adolescent susceptibility to cigarette smoking—Minnesota, 2002–2003. MMWR Morb Mortal Wkly Rpt. 2004; 53(14):301–304. MedlineGoogle Scholar
30. Fichtenberg CM, Glantz SA. Association of the California tobacco control program with declines in cigarette consumption and mortality from heart disease. N Engl J Med. 2000;343:1772–1777. Crossref, MedlineGoogle Scholar
31. Goldman LK, Glantz SA. Evaluation of antismoking advertising campaigns. JAMA. 1998;279:772–777. Crossref, MedlineGoogle Scholar


No related items




Matthew C. Farrelly, PhD, Kevin C. Davis, MA, M. Lyndon Haviland, DrPH, Peter Messeri, PhD, and Cheryl G. Healton, DrPHMatthew C. Farrelly and Kevin C. Davis are with RTI International, Research Triangle Park, NC. At the time of the study, M. Lyndon Haviland was with, and Cheryl G. Healton is with, the American Legacy Foundation, Washington, DC. Cheryl G. Healton is also with the Mailman School of Public Health, Columbia University, New York, NY. Peter Messeri is with the Mailman School of Public Health, Columbia University. “Evidence of a Dose—Response Relationship Between “truth” Antismoking Ads and Youth Smoking Prevalence”, American Journal of Public Health 95, no. 3 (March 1, 2005): pp. 425-431.


PMID: 15727971