Objective: To examine a possible problem of overreporting and to describe the degree of error with the International Physical Activity Questionnaire (IPAQ) short telephone protocol.
Design: Cross-sectional study, using two different physical activity (PA) self-report protocols.
Setting: Telephone interviews about PA in Belgium.
Subjects: Fifty adults who had previously been interviewed with IPAQ in a national survey.
Results: Seventy-five per cent reported less PA with the modified procedure than with the IPAQ. Twenty-three of the 50 individuals were found to have reported some amounts of PA with the IPAQ (either walking, or vigorous or moderate PA) when they should have reported none. In total, based on their revised reports of PA, 50% fewer persons met PA recommendations than was the case with IPAQ. The overreporting could not be related to types of error-prone individuals.
Conclusions: Overreporting of PA in population samples is a serious problem that could be reduced by implementing procedure changes without changing the IPAQ items themselves.