The Cognitive Mediation Model: Factors Influencing Public Knowledge of the H1N1 Pandemic and Intention to Take Precautionary Behaviors

Abstract

This study uses the cognitive mediation model as the theoretical framework to examine the influence of motivations, communication, and news elaboration on public knowledge of the H1N1 pandemic and the intention to take precautionary behaviors in Singapore. Using a nationally representative random digit dialing telephone survey of 1,055 adult Singaporeans, the authors' results show that the cognitive mediation model can be applied to health contexts, in which motivations (surveillance gratification, guidance, and need for cognition) were positively associated with news attention, elaboration, and interpersonal communication. News attention, elaboration, and interpersonal communication in turn positively influence public knowledge about the H1N1 influenza. In addition, results show that the motivations have significant indirect effects on behavioral intentions, as partially mediated by communication (media attention and interpersonal communication), elaboration, and knowledge. The authors conclude that the cognitive mediation model can be extended to behavioral outcomes, above and beyond knowledge. Implications for theory and practice for health communication were discussed.

Acknowledgments

A previous version of this article won the Top Faculty Paper (First Place) Award from the Communicating Science, Health, Environment and Risk Division at the 2010 annual convention of the Association for Education in Journalism and Mass Communication, Denver, Colorado. This work was supported by Nanyang Technological University Academic Research Grant number M58061012.

Notes

¹Past studies have shown that the direct effects of media attention on health behaviors are small or almost negligible (e.g., Atkin & Wallack, 1990; Morley, Wakefield, Dunlop, & Hill, 2009; Rogers, 1973; Tambashe, Speizer, Amouzou, & Djangone, 2003). At the same time, a growing body of literature has demonstrated that the effects of media attention on behavior are indirect, as mediated by social and cultural factors (e.g., Beaudoin, 2009; Randolph & Viswanath, 2004; Rogers, 1973; Yanovitzky & Bennett, 1999). Therefore, given the inconclusive evidence, we opted not to examine the direct effects of news attention on individuals' intention to undertake H1N1 precautionary behavior in our study. By so doing, we could also ensure that our model is parsimonious.

²Our sample demographics are similar to the characteristics in the 2000 Census of Population (Singapore Department of Statistics, 2010) in terms of age, gender, education, and income. The median age groups were “35–39 years old” in the census and “40 years old” in our sample. Women comprised 50.1% of the population in the census, and women comprised 54.7% of the respondents in our sample. The median education level attained in the census was “secondary education” and the median education level of our survey respondents was “A level.” The median monthly household income was “$2000–$2999” in the census and “$3001–$4000” in our sample. In general, our sample was a relatively good reflection of the Singapore population.

³Formal education was categorized as 1 = “no formal education,” 2 = “primary 6 or below,” 3 = “some secondary education,” 4 = “N level/Institute of Technical Education,” 5 = “O level,” 6 = “A level,” 7 = “diploma,” 8 = “degree,” and 9 = “postgraduate.”.

Note. *p < .05; **p < .01; ***p < .001.

⁴Confirmatory factor analyses on the four motivations were conducted to assess the multidimensionality of the concepts. Confirmatory factor analysis result shows that surveillance gratification, anticipated interaction, guidance, and need for cognition were empirically distinct dimensions (χ²(df) = 165.68(38); χ²/df = 4.36; comparative fit index =.98; Tucker-Lewis index =.97; root mean square error of approximation =.04).

⁵We use the framework suggested by Baron and Kenny (1986) to test for mediation for the key variables in the original CMM. There are evidence for mediating effects when (a) the independent variable significantly influence the mediator, (b) the independent variables significantly influence the dependent variable in the absence of the mediator, (c) the mediator has a significant unique influence on the dependent variable, and (d) the influence of the independent variable on the dependent variable is reduced when the mediator is added to the model. The Sobel test could be computed to establish the statistical significance of the mediating relation (Mackinnon & Dwyer, 1993). The mediation tests were conducted separately for elaboration and interpersonal communication. The direct path from attention to elaboration was significant (β = .44, p < .001); the path from news attention to knowledge was significant (β = .18, p < .001); the path from elaboration to knowledge was significant (β = .19, p < .001); and the addition of elaboration resulted in a reduced influence of attention on knowledge (β = .12, p < .001). The Sobel test for the mediating path was significant, z = 2.33, p < .05. These results suggest that elaboration partially mediated the influence of attention on knowledge. The direct path from attention to interpersonal communication was significant (β = .53, p < .001); the path from news attention to knowledge was significant (β = .18, p < .001); the path from interpersonal communication to knowledge was significant (β = .18, p < .001); and the addition of interpersonal communication resulted in a reduced influence of attention on knowledge (β = .13, p < .001). The Sobel test for the mediating path was significant, z = 2.31, p < .05. These results suggest that interpersonal communication partially mediated the influence of attention on knowledge.