Background: Walking is the most prevalent form of leisure time physical activity (LTPA). Advances in measurement of walking depend on understanding sources of error in self report. We examined the effect of prompting “walking for exercise, recreation, and sport” (WERS) upon surveillance estimates of LTPA and assessed what types of walking were recalled when reporting LTPA generally and when WERS was prompted specifically. Methods: Data were collected by telephone survey from a random sample of 3,415 Australian adults (≥15yrs). Respondents were asked first to recall any type of LTPA they participated in (unprompted) and if walking was not mentioned, WERS was prompted. All walkers were asked to describe the type of walking they did. Open-ended responses were categorized according to physical activity measurement dimensions. Results: Forty three percent did not report WERS unless prompted to do so. The prevalence of meeting recommendations by all LTPA was reduced by 10% for both genders and across all age groups if not prompted to recall WERS. The interpretation of WERS was broad and included travel related walking and dog walking whether unprompted or prompted. Conclusions: Current challenges in walking surveillance include ensuring that both researchers and respondents understand WERS in a standardized manner.