Sample Distribution and Research Design Are Methodological Dilemmas When Identifying Selection and Using Relative Age as an Explanation of Results

Authors: Torsten Buhre and Oscar Tschernij

Corresponding Author:
Torsten Buhre, PhD
Department of Sport Sciences
Malmö University
20506 Malmö
Sweden
+46-40-665845
torsten.buhre@mah.se

Torsten Buhre is the senior physiologist at the Department of Sport Sciences at Malmö University

Sample distribution and research design are methodological dilemmas when identifying selection and using relative age as an explanation of results

ABSTRACT
The use of a statistical test, such as the chi-squared test, to determine if selection has occurred within a sport has been used frequently in research. The assumed distribution of a sample could influence the occurrence of significant outcomes. The occurrence of significance is generally interpreted as RAE and explained as a result of selection within the sport. Most studies in this field have been done using a cross-sectional design. Therefore, the purpose of this study was to examine the influence of using different types of distribution when testing for significance, in swimming as an example, over a nine-year period of six cohorts in age by gender groups.

Results show that using either an assumed uniformed distribution or a proportional distribution of the national population distribution will lead to an increased number of significant results, in comparison to using either a distribution of the actual sample of the specific age by gender group or the distribution of the previous year within the age by gender group. In addition, when using a longitudinal design over a nine-year period, the occurrence of significance decreased over time. In order to interpret significant results as a consequence of selection within a sport the use of a sport specific and age by gender distribution and a longitudinal design is proposed.
Continue reading