Self-Report Assessment of Walking: Effects of Aided Recall Instructions and Item Order

Abstract

The effect of item format, item order, and aided recall instructions on walking scores were studied among 231 female university students under 8 experimental conditions: 2 aided recall by 4 item order conditions. Validity evidence was obtained from correlations between self-report measures of walking and a 1-week log of walking behavior. Four walking item response formats were studied: frequency by duration, closed versus open-ended, estimated total walking time, and a 4-part walking item. Mean walking, as assessed by log, was 453 min/week; self-report scores ranged from 58 to 246 min/week. Validity correlations with the walking log ranged from .31 to .61. Aided recall instructions yielded higher means and validities, but order of item presentation had no effect. The highest validity was found for the 4-part walking item that assessed time walking to and from school or work, during breaks, for errands, and for exercise. Differences in item format and recall instructions affect the validity of self-reports of walking.