Do statistical methods replace reasoning in exercise science research? How to avoid statistics becoming merely a solution in search of a problem

The nature of research in the exercise sciences can be described as lying along a continuum, at one end of which lies basic research and at the other applied research. Basic research is concerned with the corroboration or discounting of the theories that underpin the mechanisms pertaining to a particular phenomenon. It is this type of research that is typically involved in modelling physiological or psychological mechanisms. The kinds of theory-driven research questions developed by exercise scientists are typically analysed using classical hypothetico-deductive (Nevill 2000) methodologies. Here a hypothesis is formulated, an experiment is designed, data are collected and statistical methods are applied to test the hypothesis with respect to these data. Generally, this type of research asks binary questions, such as: if all other factors are controlled, does X explain Y? Basic research principles should allow the exercise science researcher to apply fundamental statistical theory and be fairly certain that, if all other factors other than X were controlled, then any changes in Y that are not due to a chance occurrence must therefore be due to X (Atkinson and Nevill 2001).