Purpose: The purpose of this paper is to examine the role of independent vs interdependent self-construal in non-smokers' responses to an anti-smoking message that focuses on either personal or relational consequences of smoking. Design/methodology/approach: Two web-based experimental studies were conducted among US college non-smokers. In the first study, participants' self-construal was measured. Then participants were randomly assigned to view an anti-smoking message emphasizing either relational or personal consequences of smoking. Message evaluation, smoking attitudes, and behavioral intentions were assessed after message exposure. The second study followed the same procedure except that participants' self-construal was manipulated by randomly assigning participants to an independent or interdependent self-construal priming task prior to message exposure. Findings: Both studies showed a noticeable pattern of interaction between message focus and self-construal: non-smokers with a salient interdependent self-construal responded more favorably to an anti-smoking message emphasizing personal (vs relational) consequences of smoking whereas non-smokers with a salient independent self-construal responded more favorably to an anti-smoking message emphasizing relational (vs personal) consequences of smoking. However, the interaction effect was small in the first study. Originality/value: Findings from this study are original in that they run counter to the general belief that messages matching people's self-perceptions will be more persuasive. On the other hand, matching health risk messages with people's dominant self-construal may reduce the messages effectiveness due to defensive processing. As a result, communication practitioners should take a great caution of tailoring threatening smoking prevention messages to target audiences' self-perceptions.