--- title: "Survivorship Bias" description: "Survivorship bias is drawing conclusions from the cases that made it through some selection process while ignoring those that did not. Because failures are usually invisible, the surviving sample looks representative when it is not, and the traits shared by survivors get mistaken for the causes o..." last_updated: "2026-08-02" --- # Survivorship Bias Survivorship bias is drawing conclusions from the cases that made it through some selection process while ignoring those that did not. Because failures are usually invisible, the surviving sample looks representative when it is not, and the traits shared by survivors get mistaken for the causes of survival. Drawing conclusions only from cases that made it past a selection filter while overlooking the failures that did not, distorting the picture. *Example: Studying only the founders who got rich, you copy their habits — and never see the many who did the same and failed.* ## What it looks like **In a business book** > All five of these founders dropped out of college, so dropping out frees you to build something great. *Why it fails:* The sample is drawn from successful founders. The far larger group who dropped out and failed is absent, so nothing can be inferred about dropping out as a strategy. **In a discussion about old buildings** > Buildings used to be made so much better — look how many centuries-old ones are still standing. *Why it fails:* Poorly built structures from the same era collapsed or were demolished long ago. What survives is the best of its period, compared against everything built today. **In an investment pitch** > This fund strategy has beaten the market for eight straight years. *Why it fails:* Funds that performed badly are closed or merged away, so the surviving list is filtered for success. Across enough funds, some streak that long occurs by chance alone. ## Practice ### Question 1 In WWII, Abraham Wald was asked where to add armour to bombers, based on damage patterns in returning aircraft. Damage clustered on wings and fuselage. What did he conclude? A. Armour the wings and fuselage, since that is where hits occur B. Armour the engines and cockpit — the areas showing no damage on returning planes, because aircraft hit there did not return to be surveyed C. The data was too sparse to support any conclusion D. Reduce armour everywhere to improve fuel efficiency **Answer: B.** Wald recognized that the sample consisted only of survivors. Undamaged regions on returning aircraft marked the hits that were fatal — planes struck there never came back to be counted. The insight inverts the naive reading of the data entirely, which is why it remains the standard illustration of the bias. ### Question 2 Which question best exposes survivorship bias in a claim about what successful people do? A. How successful are these people exactly? B. How many people did the same thing and did not succeed? C. Are these people telling the truth about their methods? D. Were these people successful for a long time? **Answer: B.** Asking about the missing denominator goes straight to the flaw. If 10,000 people took the same approach and five succeeded, the shared trait explains nothing. Option C raises a different concern about honesty, which matters but does not address the sampling problem — the accounts could be entirely truthful and the inference still invalid. ### Question 3 Why is survivorship bias particularly hard to notice? A. Because it requires advanced statistics to detect B. Because the missing cases are absent from the data by definition, so nothing in the visible evidence signals that anything is missing C. Because it only affects historical data D. Because survivors usually lie about their experience **Answer: B.** The defining difficulty is that absence leaves no trace. A dataset of successful companies does not contain a field indicating how many failed — you have to know to ask. This is why the countermeasure is a habit of questioning rather than a technique: always ask what selection produced the sample you are looking at. ### Question 4 A study finds that patients who completed a twelve-month treatment programme had excellent outcomes. What is the survivorship concern? A. The treatment may have been too long B. Patients who dropped out are excluded, and they may have left because the treatment was not working — so the completers are selected for response C. Twelve months is too short to measure outcomes D. The study should have used a larger sample **Answer: B.** Completion is itself an outcome, so conditioning on it selects a favourable subgroup. This is why intention-to-treat analysis, which counts everyone originally assigned regardless of whether they finished, is the standard in clinical trials — it is a direct structural defence against exactly this bias. ### Question 5 How does survivorship bias relate to hasty generalization? A. They are unrelated B. Survivorship bias explains why a sample is systematically unrepresentative; hasty generalization is the leap from that sample to a general conclusion C. Survivorship bias applies only to people, hasty generalization only to objects D. Hasty generalization is a type of survivorship bias **Answer: B.** Survivorship bias identifies a particular reason a sample fails to represent the population — selection by outcome. Hasty generalization is the inferential error of drawing a broad conclusion from an inadequate sample, whatever made it inadequate. Survivorship is one common cause; small sample size is another. ## Frequently asked ### How do I check for survivorship bias? Ask what process determined which cases you can see, and whether that process is related to the outcome being studied. If entry into the dataset depends on succeeding, surviving, or completing, the sample is selected on the outcome and cannot support conclusions about causes. ### Is survivorship bias the same as selection bias? It is a specific type of selection bias — the one where selection happens by surviving or succeeding. Selection bias is the broader family covering any systematic difference between who is in the sample and who is in the population, including self-selection and non-response. ### Why is survivorship bias common in advice about success? Because successful people are the ones who get interviewed, write books, and give talks. The failures who followed identical strategies are neither sought out nor inclined to publish, so almost all available advice is drawn from a sample selected on the outcome it claims to explain. ## Related - [Hasty Generalization](https://www.criticalthinkingexercise.org/pages/glossary/hasty-generalization.md) - [Base Rate](https://www.criticalthinkingexercise.org/pages/glossary/base-rate.md) - [Confirmation Bias](https://www.criticalthinkingexercise.org/pages/glossary/confirmation-bias.md) - [Statistical Fallacies](https://www.criticalthinkingexercise.org/pages/exercises/statistical-fallacies.md) - [Evaluating Research Studies](https://www.criticalthinkingexercise.org/pages/exercises/evaluating-research-studies.md) - [Lesson: traps-and-pitfalls/cognitive-biases-memory-self](https://www.criticalthinkingexercise.org/pages/learn/traps-and-pitfalls/cognitive-biases-memory-self.md) ## Sitemap See the full [sitemap](https://www.criticalthinkingexercise.org/sitemap.md) for all pages.