No. 233 · Evidence · Statistics and Probability
Error Bars
Every number comes with a margin of doubt, whether or not anyone draws it.
In the Almanac
“Millikan’s oil drops”, 1909–1913 · “Babylon Health”, 2018
Both stories in full, three ways to use it and where it misleads.
Get the AlmanacListen · Watch
Questions
What is Error Bars?
Error bars mark the range around a measurement where the true value plausibly lies. Every number carries some uncertainty from small samples, imperfect tools or natural variation, and a precise-looking figure can hide a lot of wobble. Overlapping ranges mean a difference may not be real.
How do you use Error Bars?
To use it, ask how sure a number is before acting on it. Give and request estimates as ranges, watch trends instead of single readings, and be wary of small samples. Remember that error bars cannot fix a biased method.
When does Error Bars matter?
Error Bars is one of the 68 mental models for evidence, the situations where you ask: Is this really true? Within that, it belongs to “Reading the numbers”, alongside Base Rate Fallacy, Bayes’ Theorem and Central Limit Theorem.
What is an example of Error Bars?
Two stories show it at work: Millikan’s oil drops (1909–1913) and Babylon Health (2018). Both are told in full in The Mental Models Daily Almanac, which has two stories for each of its 365 models.
Which mental models are related to Error Bars?
Overgeneralising from Small Samples, Anchoring and Statistical Significance. Error Bars itself comes from statistics and probability.