Cross-Validation of a Quarter-Mile Walk Test Using Two VO_{2 max}Regression Models

Abstract

This study was conducted to evaluate the accuracy of 2 regression models (Dolgener, Hensley, Marsh, & Fjelstul, 1994; Kline et al., 1987) in the prediction of VO2 max College-age men and women (N = 37) performed, in a counter-balanced order, a 1/4-mile walk test, a 1-mile walk test, followed by a maximal graded exercise test. For both walking tests, participants were instructed to self-select a fast (but less than maximal) steady exercise pace. For the 1-mile walk, the applicable data (e.g., age, gender, body weight, elapsed 1-mile exercise time, and ending exercise heart rate) were inserted into the Dolgener et al. (1994) equation and Kline et al. equation, respectively, to predict VO2 max A similar approach was taken to predict VO2 max for the 1/4-mile walk, except that elapsed exercise times were first multiplied by 4 to get 1-mile walk equivalents (because both equations are designed to predict VO2 max based on 1-mile data). The Kline et al. equation provided relatively accurate estimates of observed VO2 max values with mean residuals ranging from -0.36 to + 1.59 ml kg-1; min-1 and correlations ranging from .81 to .84. The percentage of predicted and observed VO2 max values within 4.5 ml kg-1; min-1; ranged from 67.6% to 75.7%. The Dolgener et al. (1994) equation, on the other hand, underpredicted observed VO2 max values with mean residuals ranging from -5.67 to -6.83 ml kg-11; min-1; and correlations ranging from .84 to .85. The percentage of predicted and observed VO2 max values within 4.5 ml kg-1; min-1; ranged from 18.9% to 43.2%. The results of this study provide evidence that the 1/4-mile walk predicts VO2 max with about the same accuracy as the 1-mile walk.