Regression to the mean. A threat to exercise science?

Area of Science:

• Exercise science
• Sports medicine
• Biostatistics

Background:

• Regression to the mean (RTM) is a statistical phenomenon that can introduce bias in research.
• It occurs when individuals are selected based on extreme initial values, leading to a natural tendency for subsequent measurements to be closer to the average.
• This bias is particularly relevant in studies lacking appropriate control groups.

Purpose of the Study:

• To highlight the impact of RTM on scientific investigations, especially in exercise science and sports medicine.
• To explain the causes and estimation of RTM.
• To propose methods for avoiding or minimizing RTM bias in research design.

Main Methods:

• The study discusses the statistical underpinnings of RTM, including intra-individual variance and measurement error.
• It outlines methods for estimating the magnitude of RTM.
• The abstract reviews strategies for mitigating RTM bias, such as controlled experimental designs and duplicate measurements.

Main Results:

• RTM can significantly bias interpretations in studies examining treatment responses or training adaptations based on initial values.
• The phenomenon can create apparent relationships, such as the inverse correlation between aerobic training response and maximal oxygen intake (VO2(max)), which may be largely due to RTM.
• Exercise scientists and sports physicians have given insufficient consideration to RTM's impact.

Conclusions:

• RTM poses a significant threat to the validity of research in exercise science and sports medicine.
• Awareness and application of appropriate research designs are crucial to prevent biased conclusions.
• Proper methodological approaches can help isolate true treatment effects from the statistical artifact of RTM.