Episode Summary
Executive Summary: Benjamin Felix interviews Alex Edmans about his book May Contain Lies, focusing on how misinformation can be true yet misleading. They discuss the ladder of misinference, confirmation bias, black-and-white thinking, cherry-picked facts, data mining, and the limits of evidence in books, media, and research. The episode emphasizes critical thinking, skepticism, and actively seeking disconfirming evidence.
Main Topics: The ladder of misinference (Priority: 5/5): Edmans introduces a framework that separates statement, fact, data, evidence, and proof, arguing that each step up requires stronger scrutiny and that many public claims fail because they are misquoted, cherry-picked, or misinterpreted. Confirmation bias and biased search/interpretation (Priority: 5/5): The conversation explains how people preferentially accept evidence that supports their prior beliefs and ignore contradictory information, including how this shows up in research, management, politics, and even AI-assisted search. Misleading uses of statistics in popular books (Priority: 5/5): The hosts examine examples such as the 10,000-hour rule and Why We Sleep, showing how famous claims can rest on weak, selective, or misrepresented evidence. Black-and-white thinking and oversimplification (Priority: 4/5): Edmans argues that people are drawn to simple binary claims, even when reality is nuanced and often depends on thresholds, context, and tradeoffs. Data mining, correlation, and causation (Priority: 4/5): The discussion highlights how researchers can produce impressive-looking results by trying many variable combinations, and why a correlation is not necessarily evidence of causation. Limits of generalization and proof (Priority: 4/5): The episode stresses that even causal evidence may only apply in a specific setting, using natural experiments and restricted samples to show why universal claims are rare. Practical habits for truth-seeking (Priority: 5/5): Edmans recommends checking sources, searching for counterevidence, encouraging dissent, and teaching critical thinking so listeners can better evaluate claims in work and daily life.
Key Arguments: A statement can be true but still misleading if it relies on an outdated draft, selective quotation, or omitted context. Confirmation bias is not just ignorance; it is an active preference for information that supports what we want to believe. People are often more skeptical of claims they dislike than of claims they agree with, which creates inconsistent standards of evidence. Highly educated people can be more vulnerable to motivated reasoning because they are better at constructing sophisticated justifications for what they want to believe. Anecdotes and single facts are not data; data must include counterexamples and the full distribution, not just the most persuasive cases. Correlation is not causation, and even large datasets can be distorted by data mining, arbitrary choices of measures, and selective reporting. Even when evidence is causal, it may only apply in one setting and should not be treated as universal proof. The right response to misinformation is not perfect certainty, but disciplined skepticism, source-checking, and an intentional search for opposing viewpoints. Leaders and experts should explicitly invite dissent so that others do not self-censor and echo chambers do not dominate decision-making.
Data Points: UK corporate governance inquiry year: 2016 - Edmans describes the select committee hearing where misleading evidence on CEO pay gaps was cited. Version gap in the pay-gap paper: 2010 draft vs 2013 final version - The witness cited a half-finished draft rather than the published paper, which found the opposite result. Practice rule: 10,000 hours - Discussed as the popular but weakly supported claim from Malcolm Gladwell’s Outliers. Sleep example graph cutoff: 5 hours of sleep omitted - In Why We Sleep, a bar for five hours was excluded to preserve a monotonic storyline. Diversity study measures tested: 24 measures - Edmans says he tested 24 gender-diversity measures for an investor and only 2 were positive. Positive results in diversity analysis: 2 of 24 - Most measures showed a negative link between gender diversity and performance. Negative results in diversity analysis: 22 of 24 - The majority of tested diversity measures did not support the desired thesis. Minimum wage natural experiment: Pennsylvania vs New Jersey - Used as an example of a quasi-experimental comparison across neighboring jurisdictions. West Point grit study sample: Highly fit cadets - Angela Duckworth’s grit evidence was based on a restricted range sample of West Point cadets. Company size policy threshold: Above a certain threshold - The UK law required companies above a threshold to publish pay ratios, following the inquiry.
Pivotal Quotes: "a statement is not a fact, it may not be accurate" — Alex Edmans: Explaining the first rung of the ladder of misinference. "we are actually addicted to seeing things that confirm what we'd like to be true" — Alex Edmans: Describing confirmation bias as psychologically rewarding and hard to resist. "perfect should not be the enemy of the good" — Alex Edmans: On why evidence can be useful even when it is not fully universal proof.
Implications: Listeners should treat viral claims, popular books, and even trusted institutions skeptically, verify sources, and seek counterarguments. For investing and decision-making, the best defense is disciplined critical thinking, not blind trust in headlines or single studies.
About The Rational Reminder Podcast
A weekly reality check on sensible investing and financial decision-making, from three Canadians. Hosted by Benjamin Felix, Cameron Passmore, and Dan Bortolotti, Portfolio Managers at PWL Capital.