| Abstract: | Some combinations of hazards yield disease rates greater than would be expected from the risk attributable to each hazard in isolation. However, perceptions are often more consistent with the combined risk being seen as less than the sum of the individual factors, raising concerns over the validity of some measures of risk perception. Here, 249 adults estimated the risk of cardiac events for four hypothetical men, described as having high or low levels of smoking and family history of heart disease. Three distinct measures were used. A 9‐point scale produced a strong sub‐additive interaction, a 101‐point scale produced a weaker sub‐additive interaction, and an unbounded scale produced no interaction. In this study, as in all previous research, scales with relatively few points (here a 9‐point scale) yield sub‐additive interactions. Given the consistency of results yielded by such scales, irrespective of context, it is concluded that these scales are not valid for perceptions of multiple risk factors. Copyright © 2004 John Wiley & Sons, Ltd. |
