Base Rate Fallacy
Edited by Paweł Raja, PhD · Published
Definition
Base Rate Fallacy is the base rate fallacy is the tendency to ignore or underweight how common something is in general (the base rate) and to rely instead on specific, case-level information when judging probability.
- Also known as:
- Base rate neglect, Base-rate neglect, Base rate bias
- Category:
- Judgment & Belief
- Impact:
- High
- Landmark study:
- Kahneman & Tversky (1973)
- Evidence here:
- 5 cited studies
Why it happens: the psychology of base rate fallacy
Bayes' rule says that a good probability judgment combines two things: the base rate (how common the outcome is to begin with) and the strength of the specific evidence. The base rate fallacy is giving the evidence almost all the weight. Bar-Hillel (1980) defined it as the tendency to ignore base rates in favor of individuating information rather than integrating the two.
Kahneman and Tversky (1973) showed this with the engineer–lawyer problem. People read personality sketches said to be drawn from a group of 70 engineers and 30 lawyers, or from a group of 30 engineers and 70 lawyers. Their probability judgments were essentially the same in both conditions — the base rate barely mattered once a description was available. The usual explanation is the representativeness heuristic: the sketch is judged by how well it matches a stereotype.
Base rates are not always ignored. Bar-Hillel (1980) argued that people rank information by perceived relevance and let the most relevant dominate. Base rates are used when they feel specific to the case, or when they can be read as a cause — for example, when one cab company is said to cause most accidents rather than simply owning most cabs.
The format of the numbers matters a great deal. Gigerenzer and Hoffrage (1995) found that stating the same problem in natural frequencies ('10 out of 1,000') rather than probabilities substantially increased the share of people who reached the correct Bayesian answer.
Base Rate Fallacy examples
In a 1978 study, staff and students at Harvard Medical School teaching hospitals were asked the chance that a person who tests positive for a disease with a prevalence of 1 in 1,000 actually has it, given a 5% false-positive rate. The correct answer is about 2%; the most common answer was 95%, and only 11 of 60 respondents answered correctly.
- Health: a person panics after a positive screening test for a rare condition, not realizing that most positives for rare conditions are false alarms.
- Money: an investor backs a new restaurant because the business plan is impressive, without considering how many new restaurants close within a few years.
- Work: a security team trusts an alert system that is '99% accurate' without noticing that, when real threats are very rare, most alerts will still be false.
- Everyday: a witness is fairly sure the car in a hit-and-run was blue, and people conclude it probably was — even if blue cars are rare in that town.
- Law: a jury hears that a forensic match is very unlikely by chance and treats it as near-proof of guilt, without asking how many people in the population could also match.
A worked example: why a positive test can still mean 'probably not'
Imagine a disease that affects 1 in 1,000 people and a test that always detects it but gives a false positive 5% of the time. Test 1,000 people: about 1 truly has the disease and tests positive, while about 50 of the 999 healthy people also test positive. Of roughly 51 positive results, only 1 is real — about a 2% chance.
The intuitive answer, 95%, comes from looking only at the test's accuracy and forgetting how rare the disease is. Writing the problem out as counts of people, as above, makes the base rate impossible to miss.
Base rate fallacy vs representativeness heuristic
The representativeness heuristic is a way of judging — by resemblance. The base rate fallacy is a specific error — ignoring prior probabilities. Representativeness is the most common route to base-rate neglect, but neglect can also come from how numbers are presented, from confusing 'the chance of a positive test if sick' with 'the chance of being sick if positive', or from simply not knowing the base rate.
How to avoid the base rate fallacy
- Ask 'out of how many?' first. Find the base rate — how common the outcome is in this population — before looking at the specific evidence.
- Translate percentages into counts. Picture 1,000 people and work out how many fall into each group; natural frequencies make Bayesian reasoning far easier.
- Separate two questions: how likely is this evidence if the hypothesis is true, and how likely is it if it is false? A test can be accurate and still produce mostly false positives for rare conditions.
- Use reference classes for forecasts. Check how similar projects, businesses or cases usually turned out, then adjust for what is genuinely different.
- Be wary of vivid single cases. A compelling description is evidence, but rarely enough to override a strong base rate.
How to design for it (nudge strategy)
Present risks as natural frequencies ('of 1,000 people tested, about 1 has the disease and about 50 test positive without it') rather than percentages and conditional probabilities, and show the base rate first.
Ethical use: design for choices people would endorse on reflection — a nudge, not sludge. Be transparent and keep opting out easy.
Key studies
In the engineer–lawyer problem, people gave essentially the same probability judgments whether the group was 70% or 30% engineers, largely ignoring the base rate.
Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Review, 80(4), 237–251. DOI →
Only 11 of 60 Harvard Medical School staff and students correctly estimated the roughly 2% chance that a positive test for a 1-in-1,000 disease was a true positive; 95% was the most common answer.
Casscells, W., Schoenberger, A., & Graboys, T. B. (1978). Interpretation by physicians of clinical laboratory results. New England Journal of Medicine, 299(18), 999–1001. DOI →
People let the information they see as most relevant dominate; base rates are integrated when they feel as specific or causally relevant as the case evidence.
Bar-Hillel, M. (1980). The base-rate fallacy in probability judgments. Acta Psychologica, 44(3), 211–233. DOI →
Presenting information as natural frequencies rather than probabilities markedly improved people's Bayesian reasoning without any training.
Gigerenzer, G., & Hoffrage, U. (1995). How to improve Bayesian reasoning without instruction: Frequency formats. Psychological Review, 102(4), 684–704. DOI →
Key research
- Errors in Probabilistic Reasoning and Judgment BiasesDaniel J. Benjamin · Handbook of Behavioral Economics: Applications and Foundations 1 (2019)
Related biases
Base Rate Fallacy: frequently asked questions
What is the base rate fallacy in simple terms?
It is forgetting how common something is. People focus on the specific evidence in front of them — a test result, a description, a witness — and ignore the background odds.
Is base rate neglect the same as the base rate fallacy?
Yes. Base rate neglect, base-rate neglect and base rate bias are all names for the same error: underweighting prior probabilities when judging a specific case.
What is an example of the base rate fallacy?
Assuming a positive result on a screening test for a rare disease means you almost certainly have it. When the disease is rare enough, most positive results can be false positives even from an accurate test.
Why do doctors and experts fall for the base rate fallacy?
Test accuracy is usually reported as conditional probabilities, which are easy to misread. Studies such as Casscells et al. (1978) show that medical training alone does not protect against the error, while frequency formats help.
How do you avoid the base rate fallacy?
Start from the base rate, convert percentages into natural frequencies, and only then adjust for the specific evidence according to how reliable it really is.
Cite this page
Raja, P. (Ed.). (2026, 8 Oct 2026). Base Rate Fallacy. Behavioral Economics Lab. https://www.behavioraleconomicslab.com/biases/base-rate-fallacy