A Comparative Analysis of the Performance of Alternative Measures of Exposure

Abstract

This paper compares eight approaches to measurement of exposure to cancer information in the mass media: general media exposure, exposure to health media, attention to health topics, quantity of health information, general and specific cancer exposure assessed by closed- and by open-ended questions. After demographic controls, the strongest predictor of cancer knowledge was the open-ended general measure (9.8% additional variance), then open-ended and closed-ended measures about diet and exercise (each 6.7%), attention to health topics (6.2%), quantity of health information (4.1%), and health media exposure (3.7%). General media exposure was not associated. However, other validity criteria lead to different conclusions: attention measures may confound motivation with exposure, and there may be upward bias in using knowledge to validate open-ended exposure measures; considering face validity, survey costs and respondent burden as well as criterion validity, we propose that the closed-ended specific questions may be most useful.

Notes

¹All of the questions are in Table 1; the roman numerals in the discussion refer to groups of questions in that table.