Health information on the Internet: Influence of online sources on credibility and behavioral intentions

Drawn from Sundar and Nass's source typology, this dissertation examines the influences of online health information sources on users' perceived credibility of information and behavioral intentions towards information. Specifically, it explores (a) the effect of the type of selecting source (website vs. bulletin board vs. blog vs. homepage vs. Internet), (b) the effect of the type of original source (doctor vs. layperson), and (c) the interaction between the type of original source and the type of selecting source.;The study yielded a significant main effect for the type of selecting source on behavioral intentions. Respondents were more likely to take action based on the information sourced from a website than from a blog, a homepage and the Internet. The effect was mediated by perceived level of gatekeeping and perceived information completeness.;The research also produced a significant two-way interaction between the type of original source and the type of selecting source on perceived credibility, mediated by perceived appropriateness of source placement. When the information was posted on a website, it was rated much higher in credibility if it was attributed to a doctor than to a layperson. When the information was presented on a homepage, it was rated slightly higher in credibility if it was attributed to a layperson than to a doctor. However, the two-way interaction differed as a function of message.;The research contributes to literature on online sources by generating a new online source typology based on perceived level of gatekeeping. In addition, it indicates that perceived level of gatekeeping predicts behavioral intentions. It reinforces dual process models by suggesting that source and source placement are powerful cues in information processing. Practical implications for online health planners, physicians, users and policy makers are discussed, followed by limitations and suggestions for future research.;Pretest 1 (N = 213) prepared two messages used in the main study. Pretest 2 (N = 16) tested the experimental stimuli. A 2 (message) x 2 (original source type) x 5 (selecting source type) full factorial experiment was conducted online among 555 randomly assigned participants.