Naked Statistics: Stripping the Dread from the Data
by Charles Wheelan
Quick Summary
An accessible, non-technical introduction to statistics that explains why statistical concepts matter in the real world. Wheelan covers descriptive statistics, deceptive data presentation, correlation, probability, the Central Limit Theorem, statistical inference, polling, regression analysis, common regression mistakes, and program evaluation, using real-world examples from baseball, Netflix, the financial crisis, and public policy to make abstract concepts intuitive and engaging.
Detailed Summary
Charles Wheelan's "Naked Statistics" applies the same readable, example-driven approach of his earlier "Naked Economics" to the field of statistics. The book opens with Wheelan's personal confession that he hated calculus because no one explained why it mattered, and argues that statistics, properly taught, is one of the most useful and intellectually exciting disciplines because it provides the tools to answer questions that actually matter.
The opening chapter asks "What's the Point?" and establishes that statistics serves two primary purposes: summarizing large amounts of information into manageable form, and drawing conclusions from limited data about larger populations. These correspond roughly to descriptive statistics and inferential statistics, the two main branches covered in the book.
The descriptive statistics chapter uses the question of who was the best baseball player of all time to introduce measures of central tendency (mean, median, mode), measures of dispersion (standard deviation, variance), and the importance of choosing the right metric for the question being asked. The "Deceptive Description" chapter warns about how statistics can be technically accurate but grossly misleading -- through selective reporting, misleading visual presentations, and the conflation of different metrics.
The correlation chapter uses Netflix's recommendation algorithm to explain how correlation measures the relationship between variables, while carefully distinguishing correlation from causation. The probability chapters cover basic probability theory, expected value, and the Monty Hall Problem (a famously counterintuitive probability puzzle), before examining how overconfident misapplication of probability models by "math geeks" contributed to the 2008 financial crisis (through mispriced credit default swaps and mortgage-backed securities).
The Central Limit Theorem -- described as "the Lebron James of statistics" -- receives a dedicated chapter explaining how the sampling distribution of means approaches normality regardless of the underlying population distribution, which is the mathematical foundation for all statistical inference. The inference chapter builds on this foundation to explain hypothesis testing, confidence intervals, and statistical significance, using the personal example of how Wheelan's professor suspected him of cheating based on an unusually high test score (which fell far outside the expected distribution).
The regression analysis chapter introduces the "miracle elixir" of statistical analysis -- the method for isolating the relationship between variables while controlling for other factors. Wheelan explains simple linear regression, multiple regression, the interpretation of coefficients, R-squared, and the critical assumptions that must be met for regression results to be valid.
The chapter on common regression mistakes serves as a warning label, covering selection bias, publication bias, omitted variable bias, reverse causation, and the ecological fallacy (drawing individual-level conclusions from aggregate data). The program evaluation chapter addresses the gold standard of research design -- randomized controlled trials -- and the challenges of evaluating programs when random assignment is impossible, covering natural experiments, difference-in-differences, regression discontinuity, and instrumental variables.
