After controlling for the demographic and perceptual variables, whether or not respondents were exposed to the campaign did not reduce personal perceptions of whether they would drive drunk or whether they actually did drive drunk. It is possible that the exposure to the campaign had no impact because people may not agree with the view that "Buzzed Driving is Drunk Driving." If subjects failed to accept this message, then perhaps no amount of exposure would influence their perception about the probability or the actuality of the behavior. A second unpredicted finding was that age was only barely, positively, and significantly related to perceptions of not driving drunk among the general market sample and was not significantly related in any of the other models. As expected, males had greater odds of actually driving drunk compared to females among both the general sample and the targeted sample, yet had lower odds of believing that they would drive drunk. The disconnect between belief and behavior revealed in this study complemented similar disparities found in victim-focused media campaign effects. This suggests that altering beliefs alone is insufficient for preventing problem behaviors, whether as a victim or an offender. The campaign was initiated in December of 2005 in an effort to impact drunk driving during the holiday season, but the campaign continued through the year and beyond. Ads were disseminated to media outlets nationwide. Data for assessing the campaign's effectiveness were obtained through a national opt-in panel with access to just over 2 million U.S. households. 3 tables, 4 notes, and 30 references