In this media campaign, considerable effort has gone into survey sampling, questionnaire design, and survey administration to evaluate the program. However, because the campaign is being implemented nationwide, there are limits to what can be learned from the evaluation—as the evaluators note in their interim report (Westat, 1999:xiv). Interrupted time-series designs are considerably stronger when the media campaign is varied across media markets. For example, media markets can be randomly assigned to different messages, different initiation dates, or differences in the number and timing of messages (see Cook and Campbell, 1979). This approach was not taken in the National Youth Anti-Drug Campaign. As a result, the evaluation lacks a “counterfactual” condition—an indication of what would have happened to comparable youth in the absence of exposure to the campaign. In addition, a public information campaign of this sort may have fairly small effects on any given individual, even though the aggregate effects may more than justify the expense. Thus, it is likely that the eventual results of the evaluation will be ambiguous.
As reviews of drug prevention have suggested and the results in Table 7.1 document, among instructional programs, those that teach students about social influences to engage in substance use and provide specific skills for effectively resisting these pressures alone or in combination with broader-based life-skills training reduce substance use, particularly when these skills are taught using cognitive-behavioral methods. However, the evidence suggests that, on average, this type of prevention strategy is not likely to have a large or even moderate effect on substance use in the general population, and its effect on subgroups in the population that are at elevated risk for developing substance abuse problems is unknown. Some research suggests that other modalities, including small-group coaching delivered to younger, more vulnerable children (e.g., Lochman’s work, described above), or the application of group contingencies for desirable behavior (e.g., Kellam and Anthony, 1998), or broad improvement to school and discipline management (e.g., Gottfredson, 1996), none of which contain any focus on drugs, may be more effective at teaching social competency skills than the universal classroom-based instructional approaches.
Although social competency skill development is one element of ef-