How Reliable Is Sports and Exercise Science?
PhD Cristian Mesquida Caldentey examined replication and reproducible issues in sports science. His work highlights selection bias, low statistical power, and the need for transparency and better methodological standards.
Cristian Mesquida Caldentey defended his PhD on January 8 at Eindhoven University of Technology, focusing on the reliability of sports and exercise science. He is part of the Human-Technology Interaction (HTI) research group within the Department of Industrial Engineering & Innovation Sciences.
Why It Matters
Sports and exercise play a major role in sports performance, health and well-being. Policymakers, coaches, and healthcare professionals often rely on scientific findings. But what if those findings are not always reliable? Mesquida Caldentey investigated how selection bias, low statistical power, and poor reporting practices affect the replicability and reproducibility of research.
What the Research Shows
His analysis of hundreds of studies shows that many published results in sports and exercise science support their hypotheses despite the low average statistical power, which suggests the presence of selection bias. Selection bias occurs when researchers and journals prefer studies that support the hypothesis, which can lead to inflated effects and misleading conclusions. His research also highlights that a priori power analyses are often poorly conducted and difficult to reproduce, and that meta-analyses frequently suffer from a lack of transparency and rigor.
Real-World Impact
Stakeholders shaping policies on sports and health need to deal with this uncertainty. For instance, injury prevention programs or training plans for older adults may rely on published research. If the underlying studies lack rigor, it can result in ineffective interventions and wasted resources.
The Challenge
Cristian Mesquida Caldentey demonstrates that improvements are possible. Greater transparency, adoption of Registered Reports, adherence to guidelines and better statistical training can raise research quality.
Cristian Mesquida Caldentey defended his thesis on January 8. Title of the thesis: . Supervisors: Daniel Lakens, Chris Snijders, and (TU Dublin).