Differing Effects of Mass and Interpersonal Communication on Breast Cancer Risk Estimates: An Exploratory Study of College Students and Their Mothers

Abstract

Research has demonstrated that women tend to overestimate the percentage of all breast cancers that result from genetic predispositions, and this article examines the knowledge of college students, as well as their mothers, on this subject, applying uncertainty management (Brashers, 2001) as the theoretical framework. The authors build on the literature by studying (a) the types of media outlets college students and their mothers use for securing information, and (b) the types of articles and programs within those outlets that may affect risk perceptions. The authors also address associations between these mass communication measures and interpersonal sources of information in the context of risk estimation. Respondents exposed to media reports about the role of genetics in breast cancer, in addition to study participants who had discussed this role within the family, tended to overestimate measures of genetic risk. Conversely, those who had attended to media reports about screening practices tended to offer lower risk estimates, indicating that such reports may have positioned genetics as just one factor in the overall equation of breast cancer risk. The authors discuss the implications of these and other findings for communication scholars and health practitioners.

ACKNOWLEDGMENTS

The authors thank the anonymous reviewers of this article, as well as Julie L. Andsager, a discussant at the 2004 Annual Conference of the Midwest Association for Public Opinion Research, for their insightful comments and feedback.

Notes

¹For verification of these statistics, see www.cancer.org, a Web site operated by the American Cancer Society.

²Question 1 gave the participant the option to provide an estimate as a ratio or percentage. Questions 2, 3, and 4 asked the participant to provide a percentage estimate.

³Checklists for the sources of information and topics were generated from a review of the literature and focus groups. The following topics were provided: “breast cancer of a celebrity,” “breast cancer of a person other than a celebrity,” “breast cancer screening recommendations,” “issues regarding the effectiveness of breast cancer screening practices,” “stories about women who had a gene that predisposed them to breast cancer,” “statistics about how often genes cause breast cancer,” “stories about how genes play a role in breast cancer,” “environmental factors related to breast cancer,” and an “other” fill-in option. For the specific sources of print information, the nearest metropolitan daily was provided, as well as “other major city newspaper,” “national newspaper,” “other hometown/local newspaper,” the university's student newspaper, “organizational magazine,” “women's magazine,” “news magazine,” an “other” fill-in option, and “don't know/can't remember.” Television options included “local news,” “national news,” “local programming other than news,” “national programming other than news,” “public broadcasting program or news,” “cable/satellite channel program,” an “other” fill-in option, and “don't know/can't remember.” For frequency analyses, participants were instructed to check all that applied for these items.

⁴These statistics were obtained from the National Cancer Institute, as well as the American Cancer Society, at the time of the research. Since then, based on adjustments to life expectancy, the estimation is closer to one in seven.

⁵Correct responses for these items were derived from Wonderlick and Fine (1997).

⁶Due to space limitations, some tables have been omitted. Tables illustrating the regression models for each dependent variable may be obtained from the first author.

⁷Note that with more categories in the response variable, the possibility for zero-count cells increases. To assist with this problem, SPSS allows researchers to set a coefficient termed Delta to between zero and one. In fitting the best models to the data, we set Delta between .01 and .5, and that value was thus applied to zero-count cells.