INTRODUCTION: Performance of elite athletes is difficult to study with conventional experimental designs, because sample sizes are usually small, coaches are often unwilling to randomize athletes to treatments, and performance in tests may not reflect performance in competitions. We present here a more practical powerful competition-based research design for sports in which a squad of athletes competes frequently as individuals and the coach is prepared to implement an intervention for a competition. METHODS: We developed and evaluated the competition-based design by analyzing US competitive swimming performance times, assuming an intervention affecting performance in all strokes and distances had been applied to swimmers in one or more arbitrarily chosen squads for one or more competitions. The swim times were downloaded from USAswimming.org for the period September 2008 through August 2009 (post Beijing Olympics through Rome World Championships and US Open Championships). We analyzed data for swimmers who achieved >900 Hy-Tek points at the World Championship selection trials and who were in squads of >2 swimmers at the trials. Each swimmer`s best points score was then used to select their best event, and only competitions with >14 best-event swims were included. Times for the resulting 365 swims in 7 competitions by 146 athletes in 19 squads were then analyzed with a mixed linear model. Fixed and random effects were included in the model to estimate the intervention effect in a design equivalent to a parallelgroups controlled trial. RESULTS: Uncertainty in the estimate of the mean effect of an intervention in a given competition with a squad of ~13 swimmers was a minimum of ±0.7% (expressed as 90% confidence limits) when there were at least 50 other swimmers in that competition and all swimmers entered at least four other competitions. The same intervention applied to an extra squad (~25 swimmers in total) in a different competition reduced the minimum uncertainty to ±0.5%. DISCUSSION: The competition-based research design for interventions on athletic performance appears to provide outcomes that are more precise and trustworthy than those of conventional research designs, but clear outcomes with interventions producing trivial or smallest important effects for swimmers (~0.3%1,2) will require use of multiple squads in multiple competitions.
© Copyright 2010 Biomechanics and Medicine in Swimming XI - Abstracts. Published by Norwegian School of Sport Sciences. All rights reserved.