Book Summary
Naked Statistics is Charles Wheelan’s attempt to do for statistics what his Naked Economics did for economics: strip away the intimidation and leave the intuition. The book walks through descriptive statistics, correlation, probability, the central limit theorem, inference, polling, and regression analysis, each illustrated with real-world stories rather than equations. His promise is that you can understand what the numbers mean, and more importantly what they do not mean, without doing the math yourself.
The tour moves briskly. Descriptive statistics tells you what happened; Wheelan shows how means mislead, using examples from baseball salaries to standardized testing. Correlation gets its famous warning label, causation sold separately, with the reminder that clever researchers can find correlations between almost anything. Probability covers the Monty Hall problem, the prosecutor’s fallacy, and why our intuitions about randomness are reliably terrible. The central limit theorem chapter is the book’s quiet centerpiece: the reason samples work at all, and the foundation of everything from polling to quality control.
The back half is where the book earns its keep for professionals. Regression analysis, the workhorse of applied statistics, gets a full, example-driven treatment: what it can tell you, what it cannot, and the standard ways people lie with it, from omitted variables to confusing statistical significance with practical importance. The closing chapter distills everything into five questions to ask of any statistical claim, a checklist worth taping to your monitor.
Throughout, Wheelan’s voice is that of the engaging lecturer: funny, impatient with jargon, and allergic to the idea that statistics is only for specialists. The wager of Naked Statistics is that statistical literacy is a civic and professional necessity, and that anyone who reads charts for a living and cannot spot a bad regression is flying blind.
Buy Naked Statistics on Amazon
Who is Charles Wheelan?
Charles Wheelan is a former correspondent for The Economist who teaches public policy at Dartmouth College, with a PhD from the University of Chicago. He wrote the bestselling Naked Economics, a plain-English tour of economic ideas, and followed it with Naked Statistics and Naked Money. His project across all three books is translation: taking technical disciplines and rendering them in the language of an intelligent, busy nonspecialist.
Wheelan’s background matters because he is not a statistician showing off; he is a policy practitioner who has watched bad numbers drive bad decisions. The book’s examples come from education policy, criminal justice, sports, and business, the places where statistical claims actually move money and lives. That practitioner’s impatience with abstraction is the book’s tone throughout.
Lessons From Naked Statistics
The first lesson is regression to the mean, the most underappreciated force in performance data. Extreme outcomes tend to be followed by less extreme ones, not because of any mechanism but because of the math. Wheelan shows how this fools coaches, managers, and investors into seeing causation in noise: the sophomore slump, the fund manager who mean-reverts after a hot streak, the pilot punished after a bad landing who then improves. If you evaluate people or strategies on short track records, you are mostly measuring luck’s echo.
The second lesson is that correlation is a clue, not a conclusion. Naked Statistics is relentless on this point because the financial world is relentless in violating it: backtests that mistake correlated factors for causes, strategies built on patterns with no mechanism, data mining dressed up as discovery. Wheelan’s rule is simple and brutal: if you cannot explain why the relationship should exist, you do not have a finding, you have a coincidence with a p-value.
The third lesson is about base rates and false positives, illustrated with the classic mammogram example: even a good test produces mostly false alarms when the condition is rare. Investors live inside this problem constantly. Most stock tips, most signals, most hot sectors are low-base-rate environments, which means most apparent patterns are false positives. Thinking in base rates is the difference between updating on evidence and being jerked around by noise.
The fourth lesson is the book’s closing checklist, five questions for any statistical claim: What is the question being asked, and is the data actually answering it? How good is the data, and how was the sample chosen? Does the analysis confuse correlation with causation? Is the result statistically significant, and is it practically important? These five questions will defuse most chart crimes you encounter in earnings decks, fund marketing, and financial media. Laminate them.
Buy Naked Statistics on Amazon
Criticisms of the Book
The book’s lightness is both its virtue and its limit. Wheelan deliberately keeps the math shallow, which means readers finish with intuition but without the ability to actually run or audit an analysis. You will know what a regression is; you will not be able to build one or spot a subtle specification error. For a book aimed at professionals who consume statistics, that is a defensible trade, but it is a trade.
Some examples have also aged. The sports and policy anecdotes that felt fresh in 2013 now read as period pieces, and the book largely skips the Bayesian perspective that has reshaped applied statistics since, as well as the replication crisis that complicated several of its sunnier claims about what careful analysis can deliver. A revised edition could go harder on uncertainty instead of technique.
Finally, the regression chapters, the most valuable in the book, deserve more room. Omitted variable bias, the workhorse problem of applied work, gets a brisk treatment when it could carry a chapter of its own. Readers who finish hungry should follow up with something meatier; this book is the appetizer, not the meal.
Who is This Book For?
This is a book for anyone who reads charts, studies, or data-driven arguments for a living and never formally learned statistics: investors, managers, journalists, policymakers, founders. If you have ever nodded along to a regression you did not understand, Naked Statistics will repay an afternoon many times over in avoided mistakes.
It is not for working statisticians or quants, who will find it breezy to the point of thinness, and not for readers who want the mathematical machinery. Pair it with Thinking, Fast and Slow for the psychology of why we misread numbers, and with How Not to Be Wrong if you want the next level of mathematical depth.
Final Thoughts
Naked Statistics will not make you a statistician. It will make you a harder person to fool, which in markets is nearly as valuable. The investor who internalizes regression to the mean, base rates, and the correlation-causation firewall has acquired three of the most expensive lessons in finance for the price of a paperback.
Read it, keep the five questions handy, and the next time a fund deck shows you a beautiful backtest, you will know exactly which questions to ask before you reach for your wallet. That is the whole book in one sentence: do not let anyone, including yourself, spend money on a number you do not understand.
