The authors surveyed 1,998 Missourians to evaluate (a) awareness and (b) understanding of messages about the impact of tobacco use in Missouri, (c) belief in the accuracy of the messages, and (d) intention to vote for a tobacco tax increase on the basis of the messages. Using structural equation modeling, the relationships among these four constructs were evaluated. A comparison of smokers and nonsmokers indicated that these two groups were influenced by the messages in different ways, [chi][superscript 2](8) = 20.89, p less than 0.05, and should be modeled separately. The nonsmokers' model demonstrated significant (p less than 0.05) relationships between understanding the messages and belief in message accuracy (b = 0.41; R[superscript 2] = 0.17) and between belief in message accuracy and intention to vote in favor of the tobacco tax (b = 0.54; R[superscript 2] = 0.29). In the smokers' model, understanding the messages was not a statistically significant predictor of belief in message accuracy. However, belief in message accuracy had a significant and positive relationship with intention to vote in favor of the tax (b = 0.45; R[superscript 2] = 0.21). These findings indicate that media campaigns about tobacco use should approach smokers and nonsmokers differently. (Contains 2 tables and 2 figures.)