Augmenting a school-based intervention with a community intervention may enhance the impact on adolescent smoking behavior beyond that of the school-based program alone. One randomized, controlled trial among eight pairs of small rural communities in Oregon examined the effect of the addition of several community components to supplement a school-based tobacco-prevention program (Biglan, Ary, Smolkowski, Duncan, & Black, 2000). The study sample was nearly evenly split across genders, and was approximately 85% white, 8% Hispanic, 5% Native American, 1% African American, and 1% Asian. The program's community components included media advocacy to encourage adults in the community to support adolescent tobacco-prevention efforts, community youth anti-tobacco activities, activities to help parents to express disapproval of tobacco use to their children, and efforts to reduce tobacco access in the community. Supplementation of the school-based program with the community intervention dampened the upward trend in smoking prevalence-observed both in this study's control group and in the U.S. as a whole-over the course of five annual surveys of 7th and 9th graders in each community. In communities that received only the school-based intervention, the prevalence of weekly smoking among 7th and 9th graders increased by 6 percentage points, from 8% to 14%, from the first to the final annual survey. In contrast, communities that received the school and the community-based component saw only a 2 percentage-point increase, from 10% to 12%, over the same time period.

A second study (Worden, et al., 1996) evaluated a mass-media campaign designed to encourage a positive perception of non-smoking and a negative perception of smoking, to teach cigarette refusal skills, and to remedy inflated perceptions of the prevalence of smoking among peers. This intervention was targeted to communities with lower adult educational attainment and focused on adolescent girls. Two small metropolitan areas received both the mass-media campaign and the school-based program, and two other, similar communities received the school-based program only, in a quasi-experimental design. The media program designers invested heavily in pre-testing the media spots among samples of the targeted audience. About 2,500 subjects, nearly evenly split across genders, were followed annually starting from grades 4-6 to grades 8-10, and then again two years later in grades 10-12. (The ethnic distribution of the sample was not detailed in the published program evaluation.) Compared with adolescent girls in the school-based-only intervention, adolescent girls in the communities that received the media campaign in addition to the school program showed a 44% relative reduction in weekly smoking prevalence: 29.4% of school-based-only participants, compared with just 16.5% of school-plus-media program participants, reported smoking during the week prior to the 12th grade survey. There was a similar but not statistically significant difference for adolescent boys. Among both boys and girls, school-plus-media intervention communities also showed a smaller rise than did school-based-only communities in positive attitudes toward smoking, perceived prevalence of peer smoking, and intentions to smoke. In general, the factors that appear to distinguish effective media campaigns from unsuccessful ones are high levels of intensity and duration of the media exposure and targeting of the campaign to a relatively high risk group (Sowden & Arblaster, 2000). There is some evidence to suggest that these media demonstration project results may be applicable also to government-funded, statewide media campaigns (Siegel & Biener, 2000), although more research will be necessary to investigate this important question.