Obesity is among the CDC's most important public health concerns, but a growing body of evidence suggests research surrounding obesity freqently includes major statistical errors. Investigators at the University of Alabama at Birmingham recently identified several key statistical errors commonly seen in obesity research.
According to Brandon George, PhD, a statistician in the UAB Office of Energetics, the goal of the paper is to provide researchers and reviewers with a tutorial to improve the scientific rigor of future obesity studies.
"Investigators who conduct primary research may find the paper useful to read or share with statistical collaborators to obtain a deeper understanding of statistical issues, avoid making the discussed errors and increase the reproducibility and rigor of the field," said Dr. George. "Editors, reviewers and consumers will find valuable information allowing them to properly identify these common errors while critically reading the work of others."
The 10 most common types of statistical errors stem from errors in statistical design, analysis, interpretation and reporting. They include:
- Misinterpreting statistical significance
- Inappropriate testing against baseline values
- Excessive and undisclosed multiple testing and "p-value hacking"
- Mishandling clustering in cluster randomized trials
- Misconstruing nonparametric tests
- Mishandling missing data
- Miscalculating effect sizes
- Ignoring regression to the mean
- Ignoring confirmation bias
- Insufficient statistical reporting
David Allison, PhD, senior investigator of the authorship group and director of the UAB Nutrition Obesity Research Center, suggests giving obesity researchers rigorous statistical training and support, through coursework during graduate or postdoctoral training, workshops or short courses.
More articles on obesity:
Antibiotics taken before age 2 leads to greater risk of obesity by age 4
Obesity increases risk of CVT in women who use oral contraceptives, study finds
8 of the worst counties in the US for residents' health