message strategy variation also will be persuasive for message execution variation. At the simplest level, assessment of alternative executions across target subgroups will provide some picture of whether something can be gained. At the next level, systematic testing of alternative executions, on the model of the three-armed strategy test described earlier with belief change as the outcome, would provide a more credible and more expensive exploration of adapting to diversity in message executions.
These are the forms of evidence that would support decisions about how much and how to adapt communication campaigns to maximize effectiveness across diverse subgroups. For each of these types of evidence, what is the extent of the evidence base? Our review of the available documents suggests that the evidence base is quite limited. Most of the evidence that was found that bears directly on the diversity issue is evidence about differential effectiveness of existing campaigns across subgroups, responding to the question just asked. In addition, there is little evidence about differential reach of projects across important subgroups. We uncovered no evidence that systematically compared effectiveness across diversity strategies.
The ongoing Florida Tobacco Pilot Program (currently known as the National Truth Campaign) aims to reduce teen smoking through a wide variety of activities, including media promotion, in-school education, contests, and enforcement. Initial results are positive for both middle and high school youth. Rigorous tracking and ongoing outcome evaluations of the program have revealed that knowledge of tobacco possession laws increased for all grades and ethnic groups and both genders, while the number of middle and high school students who bought cigarettes decreased in the first 2 years of implementation. Importantly, current cigarette use declined significantly in both the first and second years of the cam-