Seven common errors in statistical analysis by environmental scientists all stem from an outdated understanding of statistics. I'll define the seven 'perilous errors' and how each can be avoided. They revolve around old ideas about hypothesis tests, p-values, using logarithms of data, evaluating what is a good regression equation, evaluating outliers and dealing with nondetects. Understanding why each error is perilous can save the scientist from publishing incorrect statements, using inefficient analysis methods, and wasting scarce financial resources. These errors have persisted through the years -- break the cycle and step into the 21st Century.
Dennis Helsel (Ph.D. Environmental Science and Engineering):
"I'm a translator of statistical methods for scientists. My firm's name, Practical Stats, says it all. I've written/co-authored two textbooks. Statistics for Censored Environmental Data using Minitab and R (2012) pioneered statistical methods for data below detection limits. Statistical Methods in Water Resources (2nd edition coming this year) is cited by scientists throughout the world. I've taught webinars for the National Water Quality Monitoring Council and others; workshops for the American Statistical Association and others; courses such as AES for scientists in North America, Europe and Singapore since 1990.
I worked for 30 years at the US Geological Survey before starting Practical Stats. In 2003 I received the Distinguished Achievement Award from the American Statistical Association’s Section on Statistics and the Environment for my training courses in applied statistics."