(Mis)remembering health messages
MetadataShow full item record
This dissertation focuses on attitude accessibility’s impact on message perceptions and future recall as prior research demonstrates accessibility is the mechanism through which media campaigns can have their strongest influence (Berger & Mitchell, 1989; Roskos Ewoldsen, Klinger, & Roskos-Ewoldsen, 2007). Specifically, this dissertation used attitude accessibility and message type to predict which anti-smoking Public Service Announcements (PSAs) young adolescents would accurately recall 12 weeks after message exposure. Part of a larger grant project, a 2 (Sex) x 2 (Race: African American vs. White) x 2 (Condition: viewed personal testimony vs. second-half punch messages) x 3 (Message Order: nested within condition) mixed design was used in a study conducted with high school students in low-income counties in Georgia. Ninth graders (N = 244) responded to all items and viewed PSAs on laptop computers. Twelve weeks later, follow-up telephone surveys were conducted to assess unaided (i.e., no cues) recall. Multilinear or probit regression models were estimated in STATA 10.0. Results show attitude accessibility did not have a direct impact on delayed unaided recall, however, moderating effects were found. Adolescents with highly accessible negative attitudes toward smoking engaged in more message elaboration and recalled anti-smoking PSAs more frequently than other adolescents. On the other hand, those with highly accessible positive attitudes toward smoking perceived messages as more biased and accurately recalled PSAs less frequently. Therefore, PSAs were successful for adolescents with highly accessible, message consistent attitudes and likely ineffective for adolescents with highly accessible, message inconsistent attitudes. Few significant differences were found among adolescents with low smoking attitude accessibility, suggesting attitude change and/or reinforcement may be successful with this population. In addition, two different types of anti-smoking messages were used and compared. Both general components and health advocacies of personal testimony or narrative PSAs were accuracy recalled more frequently than PSAs with surprise/twist endings (i.e., second-half punch).