Base rate fallacy

There's a version of this that analytical, careful people believe doesn't apply to them: I look at the numbers. I don't get swayed by surface impressions.

And then a decision arrives - a hiring judgment, a medical assessment, an investment call - where the specific details of the case are vivid and compelling. And the base rate - the population-level statistic about how common a thing actually is - quietly retreats.

This is the myth the base rate fallacy undermines. The effect isn't a sign of intellectual weakness. It operates in precisely the domains where intelligent, trained professionals are making considered judgments.

What It Is

The base rate fallacy (also called base rate neglect) is the tendency to ignore general statistical information and focus instead on specific, vivid case details when making probability judgments - even when the general information is more relevant.

The term captures a precise error: the probability of any event depends on two things - the background frequency of that event in the relevant population (the base rate), and the specific evidence about this particular case. Base rate neglect occurs when the second gets all the weight and the first is treated as irrelevant or uninteresting.

How the Mechanism Works

The fallacy emerges from a fundamental cognitive shortcut: the representativeness heuristic.

When assessing the likelihood of something, the brain asks: how much does this case resemble the typical member of that category? A detailed, vivid description that closely matches the stereotype of a category feels like strong evidence that the case belongs there - regardless of how rare that category actually is.

The problem is that resemblance and probability are different things. A description can closely match a rare category while the actual probability of belonging to that category is low - because probability depends on the base rate, not just the match. If a category is rare enough, even a strong resemblance doesn't overcome the rarity.

Simultaneously, abstract statistical summaries are less compelling to the brain than concrete descriptions. A statistic about prevalence - "only 1 in 1,000 people in this population has this condition" - doesn't carry the visceral weight of a specific description of symptoms. The specific details feel more real, more diagnostic, more like actual evidence. The base rate feels like background noise.

A Decision in Context

During a recruitment drive, a hiring manager reviews two resumes. One applicant graduated from a prestigious university and has a polished LinkedIn profile; the other attended a local college and lists modest experience. The manager feels the Ivy League graduate is far more likely to excel in the role and offers the position without further assessment. In reality, the majority of high performers come from a variety of educational backgrounds, and the prestige of the school alone does not predict job success. By focusing on the vivid detail of the elite degree, the manager overlooks the broader base rate of performance across all graduates.

The manager's error isn't emotional or irrational in a simple sense. The Ivy League credential is a real signal - it means something. The error is in treating it as sufficient to override what the distribution of outcomes actually shows. High performers are distributed across educational backgrounds. The base rate for the local college candidate is not materially lower than for the prestigious university candidate, because performance outcomes are not tightly linked to institutional prestige at the population level.

The Research Evidence

Kahneman and Tversky demonstrated in 1973 that participants ignored base rates when judging the likelihood of occupational categories from personality sketches. Participants were given descriptions of individuals - specific, vivid character details - and asked to judge which profession each person belonged to. Even when explicitly told the statistical breakdown of professions in the group, participants weighted the personality description heavily and the stated base rates barely at all.

In medicine, the same pattern has been documented with experienced clinicians. Eddy (1982) found that physicians often overestimated the probability of disease when given positive test results, neglecting the disease's low prevalence. A positive test result for a condition that affects a small fraction of the population can still, after a correct positive result, be less likely than not to indicate actual disease - if the test's false positive rate is accounted for against the base rate. Physicians in these studies regularly gave probability estimates far above what the statistics would justify, because the positive result felt like diagnostic certainty.

Both findings share a structure: specific, salient information overwhelms the background probability.

Where It Shows Up

In medical assessment, base rate neglect produces both over- and under-diagnosis. A symptom that closely matches a serious condition generates high suspicion - regardless of how rare the condition is. Proper diagnostic reasoning explicitly integrates the prevalence of the condition in the relevant population before interpreting test results or symptom patterns.

In hiring and evaluation, the effect mirrors the approved example: specific, vivid signals - prestigious credentials, confident presentation, articulate description - carry disproportionate weight against the background distribution of performance outcomes.

In legal judgment, jurors and evaluators weight forensic matches and specific evidence compellingly, sometimes without adequately accounting for the base rate of the trait in the relevant population.

In investment decisions, investors focus on the specific story of an individual company while neglecting the base rate of outcomes for companies in that sector, at that stage, or in that market environment.

The Common Misunderstanding

The most persistent misunderstanding is that base rate neglect only occurs when people don't know the base rate. In fact, it regularly occurs when people are explicitly provided the base rate. The Kahneman and Tversky experiments gave participants the statistical information directly - and participants still largely ignored it. The specific information is more salient, more memorable, and cognitively more compelling than the abstract statistic, even when both are present.

A second misunderstanding: experts who work with statistics professionally are immune. Domain expertise reduces but does not eliminate the effect. Physicians, judges, and financial analysts all show base rate neglect under realistic working conditions.