Assessing perceptions of cancer risk: Does mode of assessment or numeracy matter?

Background: Many existing models of health behavior advance perceived risk for disease as a key motivator of risk-reduction behavior. Thus evaluating contextual factors that may influence assessment of perceived risk is important. We examined (1) how mode of assessment (mail, telephone, web-based) and numeracy affect reported estimates of perceived risk of colon cancer, and (2) how the amount of missing perceived risk data differs as a function of mode of assessment and numeracy. Method: Women (N = 457; mean age = 61.3 years) with and without Internet access participated. Women without Internet access (n = 233) were randomized into telephone or mail modes of assessment, and women with Internet access (n = 224) were randomized into telephone, mail, or web-based modes of assessment. Numeracy and four different estimates of perceived lifetime risk for colon cancer (personal percentage, population percentage, comparative, binary) were assessed. Results: No significant differences were found in obtained risk estimates for any of the four risk perception items across the different modes of assessment. Greater numeracy was associated with lower percentage estimates of perceived risk. In general, the telephone mode of assessment yielded less missing data than the mail mode of assessment. Conclusion: Mode of assessment largely does not matter when it comes to assessing perceived colon cancer risk. However, numeracy does matter and specifically impacts percentage estimates of perceived risk. While web-based, mail, and telephone modes may be used interchangeably when assessing perceived cancer risk; less missing data may result with telephone data collection.