In 19 elite schoolboy rowers, the relationships between anthropometric characteristics, metabolic parameters, strength variables and 2000-m rowing ergometer performance time were analysed to test the hypothesis that a combination of these variables would predict performance better than either individual variables or one category of variables. Anthropometric characteristics, maximal oxygen uptake ( V ·O 2max), accumulated oxygen deficit, net efficiency, leg strength and 2000-m rowing ergometer time were measured. Body mass, V·O 2max and knee extension correlated with 2000-m performance time ( r = −0.41, −0.43 and −0.40, respectively; P ⩽ 0.05), while net efficiency and accumulated oxygen deficit did not. Multiple-regression analyses indicated that the prediction model using anthropometric variables alone best predicts performance ( R = 0.82), followed by the equation comprising body mass, VO 2max and skinfolds ( R = 0.80). Although the regression equations increased the predictive power from that obtained using single variables, the hypothesis that a prediction model consisting of variables from different physiological categories would predict performance better than variables from one physiological category was not supported.
© Copyright 1998 Journal of Sports Sciences. Taylor & Francis. All rights reserved.