Steelman
Regression fallacy
Also called regressive fallacy, ignoring regression to the mean.
Crediting a cause for what was really just a natural return to average.
What it is
Attributing a change to some intervention when it’s better explained by regression to the mean: extreme results tend to be followed by less extreme ones, because part of any extreme result is luck.
The pattern
An extreme result is observed. Something is done. The next result is closer to average. Conclude that what was done caused the change.
An example
Flight instructors noticed that trainees praised for an excellent landing usually did worse next time, while those scolded for a bad one usually improved. They concluded that praise hurts and criticism helps.
Exceptional landings, good or bad, tend to be followed by more average ones either way. The psychologist Daniel Kahneman described this case from his work with flight instructors.
Why it fails
Any measurement that involves some chance tends to drift back toward average after an extreme. That drift happens whether or not anything was done, so it can’t be credited to whatever happened in between.
When it’s not a fallacy
Most lists skip this part. Calling a fair point a fallacy is a mistake too.
It can be fair when
- A comparison group experiences the same regression, so any extra improvement in the treated group can fairly be credited to the treatment.
Try one
A question from the app. The wrong answers are its usual look-alikes.
Group chat. Dev: “The football game always puts whoever just had the best season of his career on the cover.” Sam: “And almost every one of them has a worse year after. My favorite player just got the cover this year. That cover is bad luck. He should have turned it down.” Dev: “Too late now.” Sam: “Yeah. I’m still buying it day one.”
What’s the flaw?
Players land on the cover right after a career-best year, and a career-best year is usually followed by a more ordinary one anyway. The drop is the player returning to normal, not the cover’s doing.
A fair reply
“They pick guys coming off their best year ever, so most of them were going to have a worse one next anyway.”
Often confused with
- Post hoc
The regression fallacy is a particular post hoc error: the “after” is just the result returning to normal.
- Gambler’s fallacy
Regression predicts a return toward average, not a swing to the opposite extreme.
Where the name comes from
To regress is to go back. “Regression to the mean” is results drifting back toward the mean, the average, after an extreme one. The fallacy gives some cause the credit for that drift, which would have happened anyway. “Regressive fallacy” is another name for it, and “ignoring regression to the mean” spells it out.
Francis Galton described “regression towards mediocrity” in the 19th century: the children of very tall parents tend to be tall, but less extremely so. The same effect is behind the so-called Sports Illustrated cover jinx, since athletes make the cover after exceptional streaks.
Sources
- The Regression Fallacy, The Fallacy Files
- Regression toward the mean, Wikipedia
- Fallacies, Bradley Dowden, Internet Encyclopedia of Philosophy
In the app
Practice spotting it on arguments that sound real.
Steelman has more examples of regression fallacy, how to answer it in a real conversation, and practice rounds that mix it with its look-alikes, so you learn the difference, not just the name. Coming soon to iPhone and Android.
You’re on the list.
We’ll email you the day Steelman is out.
Thanks for offering to test. We’ll write before the beta with how to join.