Decisions in the Wild

Studies have shown that people understand statistical information when it is communicated in formats that appear in the natural environment and thus can be more easily processed cognitively (absolute numbers with reference classes).  Unfortunately, risks are often presented in other formats that are difficult to understand (relative percentages).

Imagine a woman who is deciding whether or not to participate in breast cancer screening with mammography. She will need to know both the benefits and potential harms (e.g. overtreatment) of mammography. Our research has shown, however, that European women dramatically overestimate the benefits of mammography. Why?

Many health brochures state that thanks to mammography, 20% fewer women die of breast cancer. But to what does the 20% figure refer? Studies have shown that without mammography screening, 5 out of 1,000 women die of breast cancer; with mammography screening, 4 out of 1,000 die. In relative numbers, this reduced breast cancer mortality of 5 to 4 out of 1,000 women translates into 20% (1 out of 5 women). In other words, the absolute risk reduction through mammography of 1 in 1,000 appears exaggeratedly large when communicated as a relative percentage.