Noise (Daniel Kahneman)

Kahneman's distinction between bias (systematic error in one direction) and noise (random, unwanted variability in judgments). Two doctors given the same patient should give the same diagnosis—but they often don't, and the variability isn't systematic bias but random noise. The same judge gives different sentences on Monday versus Friday. The same insurance adjuster values the same claim differently depending on when they see it. Noise is invisible because individual decisions seem reasonable; it only becomes apparent when you compare multiple decisions and see unexplained variability.

When to use it

When important decisions are being made by different people with different outcomes for similar cases; when consistency matters (hiring, pricing, diagnosis, sentencing); when you suspect judgment variability but haven't measured it; when designing decision processes that need to produce reliable, repeatable outcomes.

How it can help

Audit your organization's important judgments for noise: have multiple people independently evaluate the same case and measure variability. Typical findings shock organizations—the variability in professional judgment is far larger than anyone expects. Noise reduction strategies: structured decision protocols, independent evaluation before discussion, clear criteria applied systematically, and algorithmic decision aids for routine judgments. Unlike bias (which requires identifying its direction), noise is reduced by any intervention that increases consistency.