Calling Bullshit

The practical companion to the theory books: how to spot a misleading statistic, a rigged graph, a selection effect or a bogus study in ordinary media, which is a daily-use skill in every country with a phone. Written by two working scientists from an actual university course.

The chapter guide

You do not need advanced mathematics to challenge data-shaped nonsense: inspect the source, scale, selection, causality, incentives, and presentation, then refute the claim clearly and proportionately.

1. Chapter 1: Bullshit Everywhere

   Quantitative language can give weak or deceptive claims an undeserved appearance of rigor, but ordinary reasoning can often expose them.

2. Chapter 2: Medium, Message, and Misinformation

   Digital media reward speed, novelty, emotion, and sharing in ways that help misinformation outrun correction.

3. Chapter 3: The Nature of Bullshit

   Bullshit is broader than lying and often protects itself by placing the evidence or method inside a black box that audiences feel unqualified to open.

4. Chapter 4: Causality

   Correlation alone cannot establish cause, so causal claims require attention to timing, confounding, comparison groups, and alternative explanations.

5. Chapter 5: Numbers and Nonsense

   Quantitative claims should be checked for units, denominators, definitions, plausible scale, false precision, and measures distorted by incentives.

6. Chapter 6: Selection Bias

   A pattern in observed data may be created or reversed by the process that determined who or what entered the sample.

7. Chapter 7: Data Visualization

   Axes, scales, areas, baselines, cumulative totals, and design choices can make accurate data tell a visually misleading story.

8. Chapter 8: Calling Bullshit on Big Data

   Machine-learning systems inherit the limits and biases of their training data, can overfit, and do not become trustworthy merely because their methods are complex or opaque.

9. Chapter 9: The Susceptibility of Science

   Scientific institutions can produce misleading results through incentives, p-hacking, publication bias, metric gaming, and replication failure even when individual researchers are sincere.

10. Chapter 10: Spotting Bullshit

    A practical detection routine checks who is speaking, how they know, what they want, whether the claim is plausible, and what corroboration or alternative explanations exist.

11. Chapter 11: Refuting Bullshit

    Effective refutation uses proportionate tools such as counterexamples, analogies, redrawn figures, and null models while aiming at the claim and adapting to the audience.

Scorecard

Universality

4 of 5

Evidence

5 of 5

Reread value

3 of 5