Coming soon · Book Profile
The Art of Statistics
A leading statistician explains how to think clearly about data, drawing reliable conclusions from imperfect numbers while guarding against the many ways statistical reasoning goes wrong.
A profile of this book is on the way.
What it’s about
Built around real-world questions—from how Harold Shipman's murders could have been detected to whether bacon sandwiches cause cancer—The Art of Statistics reframes statistics not as a dry bag of mathematical tools but as a problem-solving discipline for learning about the world from data. David Spiegelhalter guides readers through the full investigative cycle (Problem, Plan, Data, Analysis, Conclusion), showing how to summarize and visualize numbers, infer from samples to populations, distinguish correlation from causation, build predictive algorithms, quantify uncertainty with probability, test hypotheses, and reason like a Bayesian. Crucially, the book is candid about the limits and abuses of statistics: framing tricks, questionable research practices, the reproducibility crisis, and misleading media coverage. With minimal mathematics and maximum conceptual clarity, it equips readers to produce honest analyses and to critically assess the statistical claims they meet every day, making data literacy an essential skill for the modern world.
The through-line
- Who it’s for
- A curious student, professional or citizen who wants to understand and trust the numbers they encounter at work and in everyday life.
- The problem
- Statistical claims are everywhere—headlines, studies, algorithms—and it is hard to tell which are reliable and what they actually mean. They feel intimidated by statistics, anxious about being misled, and uncertain whether they can ever judge data confidently.
- The plan
- Frame any inquiry as a problem-solving cycle: Problem, Plan, Data, Analysis, Conclusion.
- Learn to summarize and visualize data honestly using appropriate averages, spread and graphics.
- Understand how to infer from samples to populations and to quantify uncertainty.
- Separate correlation from causation and respect the role of randomized experiments.
- Evaluate predictive algorithms for accuracy, calibration, over-fitting and transparency.
- The payoff
- The reader confidently critiques media and research claims, spotting framing tricks and exaggerations. · They design and communicate their own analyses honestly, acknowledging uncertainty. · They make better decisions and avoid being misled by spurious correlations or significance.
See our guide
Related profiles we’ve built
- Applied Multivariate Stats Social Sciences Stevens →
- Bayesian Multilevel Models for Repeated Measures dаta A Conceptual and Practical Introduction in R →
- Beyond Multiple Linear Regression Applied Generalized Linear Models And Multilevel Models in R →
- Fundamentals of Social Research →
- Predictive HR Analytics →
- Handbook of Regression Modeling in People Analytics →
- Reliability and Validity Assessment →
Additional reading
- Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences · Cohen, J., & Cohen, P.
The book recommends this text for its excellent, accessible narrative discussions on regression, particularly for explanatory purposes in the social sciences.
- Classical and Modern Regression with Applications · Myers, R.
Cited as an excellent resource for its modern approach to regression analysis, especially its strong treatment of regression diagnostics for checking assumptions and identifying influential data points.
- Multivariate Statistical Methods in Behavioral Research · Bock, R. D.
This text is frequently cited by the author for more advanced or technical explanations of concepts in MANOVA, repeated measures, and step-down analysis.
- The Analysis of Covariance and Alternatives · Huitema, B.
Recommended as a very comprehensive and thorough text for readers wishing to gain a deeper understanding of Analysis of Covariance (ANCOVA).
- Structural Equations with Latent Variables · Bollen, K. A.
The guest-authored chapter on SEM heavily references this book as a key source for understanding fundamental concepts like model identification.
- Hierarchical Linear Models: Applications and Data Analysis Methods · Raudenbush, S., & Bryk, A.
The guest-authored chapter on Hierarchical Linear Modeling (HLM) cites this as the seminal text on the topic and the basis for the HLM software.
- Applied Discriminant Analysis · Huberty, C.
The chapter on Discriminant Analysis introduces this book as an excellent, current, and very thorough resource on the topic.
- Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan · John Kruschke
Recommended for reading 'During this book' to get additional information and a different perspective on Bayesian statistics and modeling.
- Regression and Other Stories · Andrew Gelman, Jennifer Hill, & Aki Vehtari
Recommended for reading 'During this book' as a supplementary text for a broader understanding of regression and Bayesian modeling.
- Statistical Rethinking: A Bayesian Course with Examples in R and Stan · Richard McElreath
Recommended for reading 'After this book' as a next step for readers who have mastered the concepts and wish to deepen their understanding of Bayesian modeling.