Abstract
Contemporary information seekers can acquire health information from an unprecedented variety of sources. The reported study applied and extended channel complementarity theory to explain the use of multiple information sources in the health-information-seeking process. Channel complementarity was extended to consider four characteristics (i.e., access to medical expertise, tailorability, anonymity, and convenience) of health-information sources. The information-seeking behavior of 3,392 respondents from the 2007–2008 Health Information National Trends Survey was analyzed to test study hypotheses. Results indicate that, sources were used complementarily based on tailorability and anonymity during health-information sources. Additionally, the likelihood of using complementary sources based on all four characteristics changed during the search process.
Notes
1. The Internet is technically not mutually exclusive from several of the other sources; it is possible, for example, for one to read a magazine that is posted on the World Wide Web. In analyzing the sources used by respondents, we assumed that respondents distinguished the Internet from other sources, and as a result, reports about Internet use were not confounded with other sources. If, for example, participants read a magazine using the Web, we assumed that they would have reported that a magazine was their source. Granted the fact that 36% of the respondents completed the questionnaire via telephone interview and had the opportunity to seek clarification regarding the questionnaire items, we believe that this assumption is reasonable.
2. With additional information, these excluded sources could be coded based on the four complementarity characteristics. For example, knowing if a telephone information number presented automated information or had a person answering questions would make it possible to code this source based on each of the four characteristics.
3. Because Hypotheses 1 and 2 were tested using one-sample t-tests, the inclusion of control variables in these hypothesis tests was not possible.