How to Be a Statistical Detective:
Simple ways to sleuth out statistical errors
February 5, 2020 | 10:00-11:00 am PST
Statistical errors are all too common in medical literature, and contribute to the reproducibility crisis currently plaguing science. Fortunately, you don’t need a degree in statistics to catch these errors. While some errors are impossible to spot without access to the underlying dataset, many are detectable just by reviewing the information available in the paper. In many cases nothing more than common sense and simple arithmetic is required. In addition, there is an ever-increasing number of free, easy-to-use online statistical tools that facilitate error detection.
In this session, Dr. Sainani will demonstrate how to apply these statistical sleuthing tools using real examples from medical literature. She will also review principles of effective statistics that can help prevent statistical errors from occurring in the first place.
You will learn: