Estimation versus falsification approaches in sport and exercise science

Area of Science:

• Sport and Exercise Science
• Statistical Inference

Background:

• Debates on statistical inference methods often overlook philosophical underpinnings and their alignment with scientific goals.
• Sport and exercise science frequently addresses
• influence-of-x-on-y
• questions, requiring robust inferential approaches.

Purpose of the Study:

• To compare estimation (Magnitude-Based Inference - MBI) and falsification (Null Hypothesis Significance Testing - NHST) statistical approaches.
• To evaluate the suitability of these methods for sport and exercise science research, particularly for quantifying effects and assessing measurement error.

Main Methods:

• Comparative analysis of MBI and NHST frameworks.
• Application of these methods to typical sport and exercise science research questions, including effect size estimation and method agreement.

Main Results:

• Falsification and NHST have limited value in addressing the nuances of sport and exercise science problems.
• Estimation approaches, specifically MBI, are more effective for quantifying the magnitude of effects and their uncertainty.

Conclusions:

• Adopting an estimation approach, such as MBI, is recommended for sport and exercise science.
• Expressing uncertainty in effect sizes and evaluating practical/clinical value against pre-defined magnitudes using MBI enhances scientific rigor.