Biases in self-reports of self-care behaviours in type 2 diabetes

Abstract

We examined the effects of question wording on self-reports of self-care in patients with type 2 diabetes. Specifically, the effects of time span and category frame were explored using modified questions from the Summary of Diabetes Self-Care Activities questionnaire. We used a shorter (past week) and a longer (past month) time span to ask questions about self-care, within participants. We also examined the effects of framing the question either in a whole (past week) or a split (past 7 days) category format, between participants. Participants answered questions about their self-care daily for a week. On day 8 they were asked the same questions framed either in a ‘past week’ or ‘past 7 day’ format. Thirty days later, participants reported on their self-care over the past month. Weekly behavioural frequency estimates calculated from weekly, 7 day and monthly self-reports were compared against daily records. We found no evidence of a split category effect in any of the questions. Neither medication taking nor blood glucose testing behavioural frequency estimates were affected by time span. Patients overestimated the frequency with which they had eaten unhealthily and had failed to limit their caloric intake. The amount of fat and sweets they consumed over the recent past was also overestimated compared to daily records. The findings were discussed with reference to work on implicit theories of autobiographical recall.

Keywords:

• Self-reports
• diabetes self-care
• time span effect

Acknowledgements

The authors would like to thank Dr. N. J. Morrish and his team for allowing us access to their patients, and Professor Peter Hampson for his helpful suggestions in interpreting the data.