the campaign seek to change fundamental behaviors, or chip away at more readily altered peripheral actions? Should the most resistant or most receptive segments be the focus of campaign efforts? What proportion of the resources should be devoted to direct influence on the focal segment versus indirect pathways via stimulating interpersonal influencers and leveraging or combating environmental determinants? Which influencers should be targeted? What is the optimum combination of awareness messages, instructional messages, and persuasive messages? How many messages should attack the competition (ranging from drinking initiation to drunkenness to impaired driving) versus promote healthy alternatives? Is it more effective to disseminate the messages via expensive TV channels or to utilize primarily minimedia? Should the campaign messages be scheduled in concentrated bursts or spread out over a lengthy period of time?

In media-based campaigns, development of the strategy entails sensitive application of mass communication theories and best practices principles. The strategic guidelines presented in this chapter draw on models, processes, generalizations, and recommendations in the voluminous research literature on media health campaigns, particularly theoretical perspectives and reviews by communication researchers such as Atkin (1981, 1994, 2001); Atkin and Wallack (1990); Backer and Rogers (1993); Backer, Rogers, and Sopory (1992); Bracht (2001); Cappella, Fishbein, Hornik, Ahern, and Sayeed (2001); DeJong and Winsten (1990, 1998); Donohew, Sypher, and Bukoski (1991); Dozier, Grunig, and Grunig (2001); Hale and Dillard (1995); Maibach and Parrott (1995); McGuire, (1989, 1994); Singhal and Rogers (1999); Slater (1999); Stephenson and Witte (2001); Wallack and DeJong (1995); and Wartella and Middlestadt (1991).

The applicability of the general principles depends on the specific context (especially types of audiences to be influenced and type of product being promoted), so effective campaign design usually requires extensive formative evaluation inputs and message pretests. Surveys, focus groups, and lab testing provide useful information to guide campaign development and to provide feedback on effective and ineffective components. Alcohol-related examples of formative evaluation are described by Atkin and Freimuth (2001).

In general, health campaigns that are targeted directly to the focal segment of the population tend to have a modest degree of impact, with limited effects on fundamental behavior patterns. But impact is highly variable, depending on the palatability of the advocated behavior and the receptivity of the target audience. Recent meta-analysis studies of comprehensive community-based campaigns show that the media contribute to a 5 to