No. 149 of 500 No. 22 in AI for Beginners

Naked Statistics

Stripping the Dread from the Data

By Charles Wheelan

Published
January 13, 2014
Print length
304 pages
Reading level
Beginner
ISBN-13
978-0393347777
Publisher
W. W. Norton & Company
Written for
General readers and aspiring analysts who want statistical intuition before studying data science or AI.

Synopsis

Charles Wheelan, an economist who teaches at Dartmouth, sets out to explain the ideas behind statistical analysis without the formulas that discourage most students. He moves through descriptive statistics, correlation, basic probability, the central limit theorem, inference, polling and regression analysis, using examples drawn from sports, marketing, medicine and public policy. A beer company's live blind taste test illustrates probability, and a sausage festival illustrates the central limit theorem. Wheelan also spends substantial time on how statistics go wrong: biased samples, selection and publication bias, confusing correlation with causation, and the flawed risk models implicated in the 2008 financial crisis.

For readers interested in AI, the book supplies the statistical literacy that machine learning assumes. Regression, probability and inference underpin most predictive models, and Wheelan's account of what those tools can and cannot show transfers directly to judging model accuracy claims, poll-based forecasts and studies cited in technology debates. His chapter on program evaluation explains how researchers try to establish causation when controlled trials are impossible. The trade-off is depth: there are almost no exercises or formulas, Bayesian methods receive little attention, and readers who want to compute results will need a textbook. As a first book that makes the subject readable and memorable, it has few rivals.

Scorecard

Editorial score

How it scored

Editorial score: 71.8 / 1005 weighted criteria

  • Influence30%

    22.5/30
  • Authority25%

    17.5/25
  • Practical value20%

    13.0/20
  • Currency15%

    9.8/15
  • Readability10%

    9.0/10
  • Editorial score

    71.8/100

How the rank was computed

Rank No. 149 of 500 comes from a composite of 59.7: observable signals 27.3 of 55 points plus editorial review 32.4 of 45 points.

All 11 inputs
InputPoints
Observable signals27.3/55
Reach (Amazon ratings volume)13.1/16.5
Consensus across best-of lists2.8/13.8
Awards and bestseller evidence3.3/8.3
Reader rating, sample-adjusted6.2/8.3
Amazon sales rank1.9/5.5
Recency0.0/2.8
Editorial review32.4/45
Author authority7.9/11.3
Influence10.1/13.5
Practical value5.9/9.0
Readability4.1/4.5
Currency4.4/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A witty, intuition-first tour of core statistics that prepares general readers to question the data claims behind AI and the news.

Strengths

  • Humor that makes dry concepts memorable
  • Concrete examples from sports, health and policy
  • Strong treatment of statistical misuse and bias

Keep in mind

Contains few formulas or practice problems and largely skips Bayesian statistics, so readers who need to calculate results will require a textbook.

Key takeaways

  1. Descriptive statistics compress complex data into a few numbers, and the choice of which number to report can mislead as easily as it informs.
  2. The central limit theorem explains why reasonably sized random samples support reliable conclusions about much larger populations, the basis of polling and inference.
  3. Correlation does not establish causation; regression can control for observed variables, but omitted variables and selection bias can still produce false conclusions.
  4. Bad data defeats good methods, since biased samples, publication bias and survivorship bias distort results regardless of how sophisticated the analysis is.

Notable facts

  • Reached the New York Times hardcover nonfiction bestseller list shortly after its 2013 publication, according to the author's site. Source (opens in a new tab)
  • Science News reviewer Tom Siegfried called it an excellent book while faulting its omission of Bayesian statistics. Source (opens in a new tab)

About the author

Charles Wheelan

Charles Wheelan teaches at Dartmouth, where he is faculty director of the Tuck School's Center for Business, Government and Society after many years as a senior lecturer and policy fellow at the Rockefeller Center. He holds a University of Chicago PhD, was The Economist's Midwest correspondent from 1997 to 2002, wrote Naked Economics and Naked Money, and founded the political reform organization Unite America.

Questions

Is Naked Statistics a good introduction for someone interested in AI?
It is a strong starting point for the statistical side of AI. Machine learning rests on probability, inference and regression, and Wheelan explains those ideas in plain language with memorable examples. The book does not cover algorithms, neural networks or programming, so readers interested in AI should treat it as preparation for a machine learning introduction rather than a substitute.
How much math is in Naked Statistics?
Very little. Wheelan keeps formulas to a minimum and focuses on the intuition behind each method, so readers who have forgotten high school algebra can follow the arguments. Reviewers noted the trade-off: without equations or practice problems, few readers will come away able to calculate a standard error or run a regression on their own.
What is Naked Statistics about?
Naked Statistics explains the core ideas of statistical analysis for general readers: descriptive statistics, correlation, probability, the central limit theorem, inference, polling, regression and program evaluation. Charles Wheelan uses humor and real cases from sports, business, medicine and public policy to show how these tools answer practical questions, and he gives equal attention to the ways biased data and careless analysis produce misleading conclusions.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.