Exposure to Inconsistent message has been demonstrated as auseful method to alleviate belief in false rumor. However, thedata from previous research included unexplained variation inresponse to inconsistent message. Existing research alsoincluded the possibility that participants skipped out onreading and therefore they were not exposed to a message.We used an eye tracker to eliminate the possibility. Eyetracking data revealed that participants not only did not skipbut they paid more visual attention to inconsistent messagescomparing with consistent messages. Despite the overalleffectiveness of inconsistent message, some responsesshowed continued belief in rumors even after the exposure.Eye-tracking analyses demonstrated that when participantshad positive pre-belief for a rumor, more visual attention toinconsistent message predicted strengthened the belief. Wediscuss when exposure to inconsistent message does not workwell as a way for harnessing belief in false rumor.