“All models are wrong”

George E.B. Box, a British industrial statistician known for his work in quality control and time series analysis, allegedly said:

All models are wrong, but some are useful.

This soothes me. I thought I always had to be right! I’m realizing that useful data analysis is not so much a matter of rightness as of goodness: good judgment, usefulness, practicality.

Dick DeVeaux of Williams College suggests that if math is music, then statistics is literature:

Literature is about the world, not about rules. It deals with life’s experience and the wisdom we develop over time.

Numbers are about the world too — at least they can be. With statistics (and related tools like data visualization) we can talk more meaningfully about the world.


One thought on ““All models are wrong”

Comments are closed.