The purpose of this study is to determine whether anti-smoking messages based on promising smoking-related beliefs increase anti-smoking intentions more than messages based on less-promising beliefs. Never smokers and former smokers will be randomly assigned to view different anti-smoking messages, and will answer questions measuring smoking-related beliefs, intentions, and message ratings online.

The goal of the main study is to test the hypothesis that messages promoting promising smoking-related belief themes identified by the Hornik and Woolf (1999) method will increase anti-smoking intentions to a greater extent than less-promising belief themes. This hypothesis will be tested by exposing research participants to messages based on a range of belief themes and comparing their anti-smoking intentions across conditions. Using the method described by Hornik and Woolf (1999), promising and less-promising smoking-related beliefs are identified using cross-sectional survey data that examines the association between smoking-related beliefs and intentions. The research team has previously used this method to advise campaign developers as to which smoking-related beliefs they should target in anti-smoking campaigns, although our cross-sectional evidence was only suggestive because it could not tease apart the causal order of beliefs and intentions (i.e., do people with anti-smoking beliefs have intentions not to smoke, or do people who don't intend to smoke develop more anti-smoking beliefs). Because our evidence has been cross-sectional, it is currently unknown whether anti-smoking messages based on promising smoking-related beliefs are indeed more likely to increase intentions not to smoke. Therefore, there are two parts to the study that we will conduct experimentally. First, in the pilot study, we will pretest a larger set of themes than will be used in the main study to ensure that the themes we do use (whether promising or unpromising) are equally convincing. Otherwise, there may be a difference in intentions based on treatment group simply because the manipulation did not work in one group and did in another, not because holding promising beliefs is actually more effective at increasing intentions. Our claim is that if it were possible to convince people of both types of beliefs (promising and less-promising), they would be less likely to smoke in the promising case than in the less-promising case because those beliefs are more powerful at changing intentions. For the main study, we will therefore use a smaller set of themes and experimentally manipulate which respondents are exposed to promising smoking-related messages and which are exposed to less-promising smoking-related messages. We expect that the randomly assigned groups will endorse promising and less-promising smoking-related beliefs to different extents. This will allow us to test our hypothesis that promising beliefs are more effective at increasing anti-smoking intentions by examining differences in smoking-related intentions based on treatment group (exposure to either promising or less-promising messages).

Inclusion Criteria:

• U.S. participants 18-25 years of age who are part of the Survey Sampling International (SSI) panel
• Never smoker (never puffed a cigarette) or former smoker (have at least puffed a cigarette but have not smoked in the past 30 days)

Exclusion Criteria:

• Less than 18 or greater than 25 years of age
• For the main study, they must not have participated in the pilot study in which we will test the how convincing these messages are
• For both studies, they must not have participated in a previous study in which we originally generated these theme sets
• Current smokers (people who have smoked in the past 30 days)