Abstract
Influencing people's perceptions through information has been the focus of much theory and research since the 1950's. Researchers have concentrated solely on messages from single sources. However, the public often receive information through the media mainly in the style of contestations, that is, information that is disputed or about which there is disagreement. In order to design effective persuasive communications there is a need to understand the mechanisms involved in changing perceptions when information is contested. In light of these issues four studies are reported that investigate changes in perceptions induced by contestations in different contexts related to the innovation of genetically modified (GM) food. The first study explores the representations held by mothers of weaning babies and by undergraduate students, for whom the issues relating to GM food are important. Using the results as a baseline, the other studies investigate the changes in students' perceptions relating to the technology and the sources involved in conflict. Based on Hovland's research on the influence of source and message characteristics (Hovland, Lumsdaine, & Sheffield, 1949), Moscovici's dual process theory of influence (Moscovici & Lage, 1976) and trust in the risk perception literature (Slovic, 1993; Renn & Levine, 1991), the studies examine the effects of credibility of the contesting experts, the numbers involved in contestations, the effect of citing academic reasons, uncertainty related to the research and vested interest of the experts as reasons for rejecting the innovation. The results indicate that, when experts provide conflicting information, their influence and the way they are perceived varies in relation to the credibility of the experts, the number of experts involved in the conflict, their perceived role in society and the nature of any associated uncertainty. The function served by trust in risk perception is illustrated in these studies. The implications of this research for the design of effective communication by government agencies, industry, non-governmental organisations and pressure groups are discussed. The advantages of extending Hovland's theory into the realm of group interactions are noted.