Abstract
Valid measures of how people view risks due to combinations of hazards are needed. 320 adult smokers responded to four vignettes in which hypothetical men were described as having high or low levels of (1) smoking and (2) either (a) blood cholesterol or (b) family history of CHD (coronary heart disease). Ratings of the risk of a heart attack were made using one of three different rating scales: (a) nine‐point, (b) 101‐point, and (c) unbounded. The nine‐point scale yielded a strong sub‐additive interaction, the 101‐point scale yielded a weak sub‐additive interaction, and the unbounded scale yielded a weak synergistic interaction. Although respondents preferred the nine‐point scale, evidence from this and other studies suggest that scales with nine points or fewer should not be used to assess perceptions of risks due to multiple hazards.
Acknowledgements
David French was a Wellcome Trust Training Fellow in Health Services Research (reference number 060634/Z/00/Z) at the Universities of Cambridge and London (King's College) while this work was conducted.