Chapter 28 Further Reading: Probabilistic Thinking and Uncertainty

Core Academic Research

1. Kahneman, D. (2011). Thinking, Fast and Slow. New York: Farrar, Straus and Giroux.

The magnum opus of behavioral economics and cognitive psychology, synthesizing Kahneman and Tversky's decades of research on judgment and decision-making under uncertainty. The book introduces the System 1/System 2 framework (intuitive vs. deliberate processing), covers the heuristics and biases program comprehensively, and discusses the Linda problem, prospect theory, and numerous other findings directly relevant to probabilistic reasoning and its failures. Particularly valuable: the chapters on the law of small numbers, availability heuristic, base rate neglect, and overconfidence. Despite occasional critics who argue the replication record of some individual studies is mixed, the core findings — including those directly relevant to this chapter — have robust empirical support.

2. Tetlock, P. E., & Gardner, D. (2015). Superforecasting: The Art and Science of Prediction. New York: Crown.

The accessible account of the Good Judgment Project and its findings, written for a general audience. Tetlock and journalist Dan Gardner explain the research methodology, introduce readers to actual superforecasters and their reasoning processes, and draw out the implications for expertise, punditry, and institutional decision-making. The book is simultaneously an argument about the nature of expertise, a practical manual for improving forecasting accuracy, and a critique of the media incentive structures that reward confident wrong predictions. Essential reading for anyone interested in calibration and probabilistic thinking.

3. Tetlock, P. E. (2005). Expert Political Judgment: How Good Is It? How Can We Know? Princeton: Princeton University Press.

The more academic predecessor to Superforecasting, presenting the full 20-year study of expert political prediction. More rigorous than the popular book, with detailed statistical analysis of forecast accuracy across different experts and domains. The core finding — that expert predictions are barely better than chance and that foxes outperform hedgehogs — is presented here with full methodological transparency. Particularly valuable for graduate students and researchers who want to evaluate the evidence base rather than just the conclusions.

Bayesian Reasoning and Statistics

4. Gigerenzen, G. (2002). Calculated Risks: How to Know When Numbers Deceive You. New York: Simon and Schuster.

Gigerenzen's accessible account of how probability is misunderstood in medicine, law, and everyday life. Central argument: most probabilistic fallacies (including the false positive problem and base rate neglect) can be dramatically reduced by reformulating probability as natural frequencies ("10 out of 10,000") rather than conditional probabilities ("0.1%"). The book presents extensive research showing that physicians, lawyers, and policymakers make significantly fewer errors when information is presented in frequency format. Essential reading for anyone who communicates probability to non-specialist audiences.

5. Spiegelhalter, D. (2019). The Art of Statistics: How to Learn from Data. London: Pelican.

Cambridge statistician David Spiegelhalter's comprehensive yet accessible treatment of statistical reasoning for non-specialists. Covers data visualization, hypothesis testing, confidence intervals, Bayesian inference, and prediction — all with careful attention to how statistical reasoning is communicated and misunderstood. The chapter on communicating risk is particularly valuable, including analysis of relative vs. absolute risk, the importance of denominators, and the pitfalls of single-study reporting. Spiegelhalter also runs the Winton Centre for Risk and Evidence Communication at Cambridge, whose research on risk communication is a valuable additional resource.

6. Silver, N. (2012). The Signal and the Noise: Why So Many Predictions Fail — But Some Don't. New York: Penguin Press.

FiveThirtyEight founder Nate Silver's account of probabilistic prediction across many domains — elections, sports, weather, economics, earthquakes, climate. Silver's central argument is Bayesian: good forecasters update from strong priors using high-quality evidence, while bad forecasters are overconfident in noisy signals. The book covers the importance of base rates, the challenge of distinguishing signal from noise in complex data, and the systematic failures of expert prediction in domains from financial markets to pandemic forecasting. Particularly strong on the distinction between probabilistic forecasting (which provides honest uncertainty estimates) and deterministic punditry.

Misinformation, Manufactured Uncertainty, and Decision-Making

7. Oreskes, N., & Conway, E. M. (2010). Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming. New York: Bloomsbury.

The definitive historical account of manufactured uncertainty as a deliberate strategy, documenting how a small group of scientists — often the same individuals across multiple campaigns — worked with industry to create apparent scientific controversy about tobacco, ozone, acid rain, climate change, and other issues where consensus was strong but regulations were opposed. Essential reading for understanding how genuine uncertainty and manufactured uncertainty differ and how the latter is created. The book shows how the "Merchants of Doubt" strategy exploits the asymmetry between establishing claims and introducing doubt, and how it leverages media norms of balance.

8. Gigerenzen, G., Gaissmaier, W., Kurz-Milcke, E., Schwartz, L. M., & Woloshin, S. (2007). "Helping Doctors and Patients Make Sense of Health Statistics." Psychological Science in the Public Interest, 8(2), 53-96.

A comprehensive academic review of how medical statistics are misunderstood by physicians, patients, and journalists, and how natural frequency formats improve comprehension. This paper provides the research basis for the clinical application of Bayesian reasoning discussed in Section 28.3 and Case Study 28-1. It includes analysis of how risk is communicated in patient leaflets, medical journals, and health news — and documents the systematic ways these communications inflate apparent benefit and understate absolute risk.

Philosophy and Epistemology

9. Bovens, L., & Hartmann, S. (2003). Bayesian Epistemology. Oxford: Clarendon Press.

An academic treatment of Bayesian epistemology as a theory of rational belief and belief change. Covers the formalization of coherence, confirmation theory, testimony reliability, and the aggregation of testimonial evidence. More technical than the other entries on this list, but provides the philosophical foundations for the practical Bayesian reasoning covered in this chapter. Chapter 4, on "Coherence and the Reliability of Information Sources," is directly relevant to the question of how to aggregate evidence from multiple news sources with different credibility levels.

10. Jaynes, E. T. (2003). Probability Theory: The Logic of Science. Cambridge: Cambridge University Press.

A comprehensive and highly influential treatment of probability as an extension of logic — the view that probability theory provides the correct normative model for reasoning under uncertainty, whether in science, law, or everyday inference. Jaynes argues that Bayesian probability is not one statistical methodology among many, but the uniquely correct formalization of plausible reasoning. This is a challenging text, primarily for readers with mathematical background, but its first four chapters are accessible and give the intellectual foundation for treating probability as epistemology rather than just statistics.

Calibration and Forecasting Practice

11. Mellers, B., Ungar, L., Baron, J., Ramos, J., Gurcay, B., Fincher, K., ... & Tetlock, P. E. (2014). "Psychological Strategies for Winning a Geopolitical Forecasting Tournament." Psychological Science, 25(5), 1106-1115.

The primary academic paper reporting the Good Judgment Project's findings on forecasting performance. Documents the characteristics that distinguish superforecasters from ordinary forecasters, the value of team aggregation, and the specific training interventions (active open-mindedness training, calibration training) that improved performance. Essential primary literature for researchers and students who want to evaluate the superforecasting evidence base rather than relying on the popular book.

12. Mandel, D. R., & Barnes, A. (2014). "Accuracy of forecasts in strategic intelligence." Proceedings of the National Academy of Sciences, 111(30), 10984-10989.

A complementary paper evaluating the accuracy of intelligence community forecasts, finding similar limitations to those documented by Tetlock in academic experts. Provides the evidence base for the claim that GJP superforecasters outperformed CIA analysts. Important for the policy implications of superforecasting research in national security contexts.

Communicating Uncertainty

13. Fischhoff, B. (1994). "What Forecasts (Seem to) Mean." International Journal of Forecasting, 10(3), 387-403.

A classic paper on the mismatch between intended and interpreted probabilistic communication, demonstrating that probability words and numbers are systematically misinterpreted by non-specialist audiences. Fischhoff's research predates and complements the IPCC's systematic approach to uncertainty communication, and provides the empirical foundation for preferring numerical over verbal probability expressions in high-stakes communication.

14. van der Bles, A. M., van der Linden, S., Freeman, A. L. J., Mitchell, J., Galvao, A. B., Zaval, L., & Spiegelhalter, D. J. (2019). "Communicating uncertainty about facts, numbers and science." Royal Society Open Science, 6(5), 181870.

A comprehensive review of research on uncertainty communication, covering how different formats (ranges, confidence intervals, probability distributions) affect understanding and decision-making among non-expert audiences. Includes analysis of when expressing uncertainty undermines trust versus when it enhances credibility. Directly relevant to the debates about science communication discussed in Section 28.8.

15. Frewer, L. J., Miles, S., Brennan, M., Kuznesof, S., Ness, M., & Ritson, C. (2002). "Public preferences for informed choice under conditions of risk uncertainty." Public Understanding of Science, 11(4), 363-372.

Research on how the public prefers to receive uncertain health risk information, with implications for communicators and policymakers. Finds that people generally prefer honest acknowledgment of uncertainty over false certainty, even when uncertainty is uncomfortable — contra the assumption that simplification is always necessary for public communication.

Note on Access

Tetlock's Superforecasting and Kahneman's Thinking, Fast and Slow are widely available in both print and digital formats. Gigerenzen's Calculated Risks and Silver's The Signal and the Noise are similarly accessible. The academic papers cited are available through university library databases (PsycINFO, Web of Science, JSTOR). Many authors post preprints on ResearchGate or Academia.edu. The Winton Centre for Risk and Evidence Communication at Cambridge (wintoncentre.maths.cam.ac.uk) provides freely available resources on risk communication that complement several of the academic papers cited above.