Naked Statistics: Stripping The Dread From The Data by Charles Wheelan describes the proper ways to employ statistical analysis, as well as incorrect or misleading uses of the discipline.
The author begins by explaining how statistics is used to study interrelationships through the review of large data sets to arrive at inferences as to what the data means. Mass data is summarized to determine underlying phenomena. Sometimes, the data is analyzed to determine qualitative
aspects or defects as in the Poisson Distribution.
Wheelan explains classic problems in interpreting data. For instance, practitioners may not understand when events are independent. Occasionally, clusters impact the interpretation of the data. In other cases, practitioners wrongly assume that events are independent when they are not.
The author discusses the benefits and limitations of regression analysis. In addition, correlation does not necessarily imply causality. Wheelan discusses survivorship bias which applies classically to assumptions about students as they progress in seniority from freshman year through to the senior year.
Wheelan provides a very interesting limitation for regression. In essence, regression is not necessarily accurate for non-linear cases. The area of fractal geometry may be introduced to enhance the regression equation to deal with
non-linearity. The author does not discuss fractal geometry applications to probability/statistics.
The author explains the basic principles of how casinos operate and how people can lose money by betting in accordance with wrong assumptions. At bottom, casinos operate so that the outcome of one event has no discernible impact on another.
Naked Statistics is an important book on the proper application of statistics for solving problems of inference and even conditional probability. The presentation is easy to read and understand for a wide constituency of readers in academe, the professions and the general public.
This book should be read in connection with designing and interpreting statistical experiments.Powered by Sidelines