Cross-validation of three jump power equations. Med. Sci. Sports Exerc., Vol. 31, No. 4, pp. 572-577, 1999. The vertical jump-and-reach score is used as a component in the estimation of peak mechanical power in two equations put forth by Lewis and Harman et al.
Purpose: The purpose of the present study was to: 1) cross-validate the two equations using the vertical jump-and-reach test, 2) develop a more accurate equation from a large heterogeneous population, 3) analyze gender differences and jump protocols, and 4) assess Predicted Residual Sum of Squares (PRESS) as a cross-validation procedure.
Methods: One hundred eight college-age male and female athletes and nonathletes were tested on a force platform. They performed three maximal effort vertical jumps each of the squat jump (SJ) and countermovement jump (CMJ) while simultaneously performing the vertical jump-and-reach test. Regression analysis was used to predict peak power from body mass and vertical jump height.
Results: SJ data yielded a better power prediction equation than did CMJ data because of the greater variability in CMJ technique. The following equation was derived from SJ data: Peak Power (W) = 60.7 × (jump height [cm]) + 45.3 × (body mass [kg]) − 2055. This equation revealed greater accuracy than either the Lewis or previous Harman et al. equations and underestimated peak power by less than 1%, with a SEE of 355.0 W using SJ protocol. The use of one equation for both males and females resulted in only a slight (5% of power output) difference between genders. Using CMJ data in the SJ-derived equation resulted in only a 2.7% overestimation of peak power. Cross-validation of regression equations using PRESS reveals accurate and reliable R2 and SEE values.
Conclusions: The SJ equation is a slightly more accurate equation than that derived from CMJ data. This equation should be used in the determination of peak power in place of the formulas developed by both Harman et al. and Lewis. Separate equations for males and females are unnecessary.
Department of Physical Education, Health Fitness Studies, Central Connecticut State University, New Britain, CT 06050; Military Performance Division, U. S. Army Research Institute of Environmental Medicine, Natick, MA 01760-5007; and Computer Science Department, University of Massachusetts, Amherst, MA 01003-4610
Submitted for publication July 1997.
Accepted for publication March 1998.
Address for correspondence: Stephen P. Sayers, Department of Exercise Science, room 110, Totman Building, University of Massachusetts, Amherst, MA 01003. E-mail: email@example.com.