Self-report data are essential to behavioral research and clinical practice. Self-report is one of the primary means of obtaining information about a person, placing it at the heart of the research history that underlies much of cancer diagnosis and care. There are numerous benefits of retrospective self-reports, such as the expeditious development and modification of measures that are easy to administer and complete. Indeed, self-report is often the most cost-effective and valid means of collecting personal information (Baker & Brandon, 1990; Stone, Shiffman, & DeVries, 1999). Assessment of cancer-related behavior via self-report shares these strengths, although there are also limitations inherent to the use of self-report methods.
Behaviors (and their predictors) influencing cancer risk are often individual difference variables that require self-report (e.g., number of cigarettes smoked daily) or are constructs that cannot be readily observed (e.g., emotional experience, motivation towards preventive care). Although some behaviors have a gold standard that can be employed to test validity (e.g., tobacco use, some screening behaviors), for many other behaviors there is less clarity, ease, or a gold standard is not available (e.g., eating a nutritious diet, symptoms such as pain and fatigue). Improving the accuracy and utility of self-report measures would advance cancer research and clinical practice. We attempt to provide an overview of this issue by reviewing several key domains in the self-report of cancer-related behaviors.