Previous research in the area of public health communication has predominantly focused on a direct association between information in the media and health behavior change. The dissertation seeks to broaden this theoretical framework by examining complementary routes of media effects on behavior. Specifically, it is argued that by stimulating response from social institutions and facilitating change in the social acceptability of problem behaviors, the media may set in motion formal and informal mechanisms of social control that lead to behavior change. This proposition was tested in relation to the decline in drunk-driving (DO) behavior in between 1978 and 1995. In the first step, an elaborated content analysis procedure revealed that the grassroots movement against DO was instrumental in setting the media agenda for the DO problem but was no more influential than policy-makers. In the next step, the results demonstrated that policy-makers' actions followed the frames and solutions advocated in the media and that the impact of news stories on policy-makers' attention and behavior was primarily manifested in periods of intensive policy-making. The third step of the analysis tested hypotheses regarding the media's contribution to the emergence of an unequivocal social norm against DO. The findings suggested a small independent contribution of media coverage to social disapproval of DD but could not convincingly distinguish between a direct effect (through social learning) and a mediated effect (through social interaction). In the final step, an analysis of all three routes of media effects on DD behavior (I.e., media-behavior, media-policy-behavior, and media-norm-behavior) provided some compelling evidence of mediated media effects on DO behavior and some ambiguous findings regarding direct media effects. In addition, there was evidence that media effects on DO behavior varied by the level of resistance demonstrated by drivers of different age groups to different types of social control efforts. Overall, the results of this study support the argument that by neglecting to consider media effects on health behavior change that are mediated by other social structures, previous studies may have underestimated the contribution of mass communication channels to processes of health behavior change.