Sample Size Calculation

We calculated the sample size necessary for our research goals with a program developed by the Clinical Epidemiology and Biostatistics Unit of the Universidade da Coruña (A Coruña, Spain) (http://www.fisterra.com/mbe/investiga/9muestras/9muestras2.asp) [28]. We based our calculations on prior results obtained by McCulloch on the effect of pedal type on EMG activity patterns in muscle recruitment (16.9 ± 7.2% for a lateral translation pedal and 30.9 ± 8.5% for a standard pedal) [29]. With these results, a sample of 8 riders would give 80% analysis power (beta = 20%), to detect a significant effect (alpha of 0.05) in a two-tailed test for this cross-over experimental study.
