Base Rate Anchoring¶

Why It Matters¶

The "base rate fallacy" is among the most robust findings in judgment research. When given individuating evidence (this person tested positive, this startup has a great pitch), we ignore category frequencies (how rare is the disease, what percentage of startups succeed). This generates systematic overconfidence in diagnosis, prediction, and attribution. Base rate anchoring is the corrective discipline: start with the category frequency, then adjust for individual evidence.

The Key Move¶

Before considering any specific evidence about an individual case, ask: "What is the frequency of this outcome in the appropriate reference class?" State this number explicitly. This is your anchor. Then, as evidence arrives, ask: "How much does this evidence move me from the base rate?" Resist the temptation to ignore the base rate once you have specific information. If you find yourself at a conclusion far from the base rate, you need strong evidence to justify that distance.

Classic Application¶

Predicting startup success. A founder has impressive credentials, a compelling pitch, and strong early traction (individuating evidence suggesting success). But base rate: 90% of startups fail. Even very positive signals might only shift your posterior to 30-40% success probability - still pessimistic, but appropriately so. Ignoring the base rate leads to systematic over-investment in ventures that "look good" but face structural headwinds.

Surprising Application¶

Interpreting personality in social situations. You meet someone who seems anxious (individuating evidence). Base rate question: What fraction of people are anxious in this specific context (first day at new job, public speaking, meeting in-laws)? If the context makes most people anxious, this behavior is weak evidence about their general personality. The base rate anchors you to "this is contextually normal" before you infer dispositional traits. Prevents the fundamental attribution error's tendency to ignore situational base rates.

Failure Modes¶

Reference class selection - base rates depend on the class (startup success rate varies hugely by sector, funding stage, founder experience). Obsolete base rates - using historical frequencies when structural conditions have changed. Ignoring evidence - base rate anchoring isn't "always use the average"; strong evidence should move you. Outside view paralysis - sometimes individual details matter more than category frequencies, especially in novel situations.

Go Deeper¶

Kahneman, Thinking, Fast and Slow, Chapter 16 on base rate neglect; Meehl, Clinical Versus Statistical Prediction on base rates in clinical judgment.

Anchoring - Behavioral Economics: Reasoning About Systematic Deviations