Simpson’s paradox

The Devil’s Advocate™ →Go uptown In which the curmudgeonly old sod puts the world to rights. Index — Click ᐅ to expand: The Devil’s Advocate Tell me more Sign up for our newsletter — or just get in touch: for ½ a weekly 🍺 you get to consult JC. Ask about it here.

Simpson’s paradox is a statistical phenomenon where a trend that is apparent appears in several groups of related data, when viewed in isolation, disappears or reverses when the groups are combined.

It plays particular havoc with social scientists when they try to draw causal inferences — or moral imperatives — from data they have gathered with the careful intention of illustrating their own precious hypothesis.

This can upset, for example, eugenicists, craniometrists, and those who tend to the idea that a measure of “intelligence quotient” has any necessary bearing on practical intelligence, expected success in life or personal value, let alone that it can be used to extrapolate into differences in intelligence between, say, races.

But it is equally offensive to those upset by Jordan Peterson and his fondness for pointing out how hard it is to draw firm conclusions — let alone moral ones — from any data set on which multiple, unrelated factors operate.

See also

• Correlation does not imply causation
• Multivariate factors
• Theory-dependence of observation