Discriminant analysis of biomechanical differences between novice, good and elite rowers

Publisher Website
Google Scholar
Abstract
Oar force and oar angle data resulting from a 6‐min maximal rowing ergometer test undertaken by novice (n = 9), good (n ‐ 23) and national (n = 9) level male rowers, were used to identify biomechanical performance variables which accurately discriminated between rowers of differing ability levels. The variables included two work capacity measures, mean propulsive power output per kilogram of body mass (W kg‐1) and propulsive work consistency (%), and two skill variables, stroke‐to‐stroke consistency (%) and stroke smoothness (%). Discriminant function analysis indicated the presence of two functions, both of which clearly indicated the importance of mean propulsive power output per kilogram of body mass as a discriminating variable. Function 2 gave greater weight to stroke‐to‐stroke consistency and stroke smoothness than function 1; however, function 1 was the most powerful discriminator. Classification procedures were used to predict the ability level to which a rower most likely belonged and involved defining the ‘distance’ between each rower and each ability level centroid, with the rower being classified into the ‘nearest’ ability level. These procedures indicated that 100% of the elite, 73.9% of the good, 88.9% of the novice and 82.9% of all rowers were correctly classified into their respective skill levels. Stepwise discriminant analysis included the variables in the following order: mean propulsive power output per kilogram of body mass, stroke‐to‐stroke consistency, stroke smoothness and propulsive work consistency (P < 0.001). The results of this study indicate that biomechanical performance variables related to rowing capacity and skill may be identified and used to discriminate accurately between rowers of differing skill levels, and that, of these variables, propulsive work consistency is the least effective discriminator.

This publication has 4 references indexed in Scilit: