Do multiple choice questions make us biased?

Episode Summary

Executive Summary: The episode examines why humans are bad at truly random choices and probability, using quiz-answer patterns, the Monty Hall problem, roulette, and poker to show how biases like middle bias, gambler’s fallacy, hot-hand fallacy, and attribution bias shape decisions. Experts argue these biases can be useful in natural settings but mislead us in games of chance, exams, finance, and everyday judgment.

Main Topics: Quiz-answer pattern bias and “anti-A” tendencies (Priority: 5/5): The episode begins with a listener noticing that an Unexpected Elements quiz seems to favor B and C over A. Producers admit they unconsciously avoid A because it feels too obvious and less engaging for listeners. Middle bias and imperfect randomness (Priority: 5/5): Mathematician Kit Yates explains that people often choose middle options when unsure, which affects multiple-choice tests, games, and even everyday choices like selecting a toilet cubicle or placing battleships. Probability mistakes: birthday problem and Monty Hall (Priority: 5/5): The show uses famous probability puzzles to reveal how intuition fails. The birthday paradox shows how quickly shared birthdays become likely, and Monty Hall demonstrates that switching choices gives a 2/3 chance of winning. Gambler’s fallacy and hot-hand fallacy in gambling (Priority: 5/5): Economist Rachel Croson explains that people wrongly think past outcomes affect future independent events, as in roulette or coin flips, and also mistakenly infer personal streaks of success or failure. Behavioral finance and real-world decision errors (Priority: 4/5): The episode connects gambling biases to investing, specifically the disposition effect, where people sell winners too soon and hold losers because they expect reversals that are not statistically justified. Poker, process versus outcome, and attribution bias (Priority: 5/5): Poker player and psychologist Maria Konnikova argues poker is a laboratory for decision-making because it separates skill from luck and teaches players to judge process rather than outcomes, while resisting self-serving explanations for wins and losses. Cognitive biases as adaptive but costly pattern detectors (Priority: 4/5): The episode closes by framing biases as evolutionary byproducts of a brain tuned to detect patterns. That same pattern-recognition ability helps survival in nature but can misfire in modern probability tasks.

Key Arguments: Humans are not good at choosing randomly; we gravitate toward patterns such as middle options or familiar-looking answer positions. Multiple-choice quizzes can unintentionally reflect writer bias, but A can feel too obvious and reduce listener engagement. Middle bias can be adaptive when the true answer is likely near the center of a distribution, but it becomes a mistake when all options are equally likely. The birthday problem and Monty Hall both show that intuition about probability is often wrong, especially when options or combinations grow nonlinearly. In independent random systems like roulette or coin flips, past results do not change future probabilities, despite strong human intuition that they do. The gambler’s fallacy and hot-hand fallacy are distinct but related: one assumes reversal after a streak, the other assumes continued success after a streak. The disposition effect in investing mirrors gambler’s-fallacy thinking by encouraging people to sell winners and hold losers. Poker teaches that good decision-making should be evaluated by process, not outcome, because luck can obscure whether a choice was correct. Pattern recognition is evolutionarily useful, but it produces false positives in random environments, leading to cognitive bias.

Data Points: Monty Hall win rate when switching: 2/3 (about 66%) - Kit Yates explains that switching doors after one goat is revealed wins two times out of three. Monty Hall losing chance when switching: 1/3 - If the initial choice was already correct, switching loses. Birthday problem threshold: 23 people - The probability that at least two people share a birthday exceeds 50% with only 23 people in a room. Expected naive birthday guess: about 182 people - Most people incorrectly estimate the threshold by dividing 365 by 2. World Cup birthday example: 16 of 32 teams - In the 2014 men’s World Cup, exactly half the squads had at least one pair of players sharing a birthday. Roulette probability of hitting one number: about 1 in 38 - Rachel Croson describes the odds on an American roulette wheel. Roulette table observation: hundreds and hundreds of spins - A gambler tracked many spins and believed one number (18) was unusually frequent. Quiz game playthrough: 6 wins out of 10 - In the cup-based Monty Hall demo, switching won six of ten rounds, close to the expected rate.

Pivotal Quotes: "If there was a big red button that would just demolish the internet, I would smash that button with my forehead." — The Interface promo: Opening and closing promo unrelated to the main science segment. "A just doesn't feel very random, like the vibes are off." — Unexpected Elements producer: Explaining why quiz writers hesitate to make option A the correct answer. "Process and outcome are not the same thing." — Maria Konnikova: Her core poker lesson about separating decision quality from luck.

Implications: Listeners are encouraged to question instincts about randomness, streaks, and “obvious” answers. The episode suggests better decisions come from understanding probability, checking biases, and judging choices by process rather than outcome.