The Texas Sharpshooter Effect
They never miss. A perfect toss every time. We could all be easily fooled into believing that the “Dude Perfect” guys have extraordinary coordination and timing, but what we’re not seeing are the thousands of failed attempts.
This is the TEXAS SHARPSHOOTER EFFECT. Inconvenient data has been omitted. Someone analyzes acres of data, but only reports the findings they find interesting. It’s sort of like firing hundreds of gunshots at the side of a barn, then painting a bullseye around the cluster with the most hits.
When we ransack large data sets, random chance guarantees we’ll find clusters. Our brain is fooled into believing we’ve stumbled upon causality, when in fact all we’ve found is random noise.
As British economist Ronald Chase said, “If you torture the data long enough, it will confess to anything.”
The following telltale signs signal that you might be looking at research that tells a misleading story:
Weirdly specific sample groups:
“Cleveland residents between 30 and 40, who commute daily.”
Unusual start or end dates:
“People buying groceries in the third week of March 2019.”
Small sample sizes:
“35 people suffering from thyroid abnormalities.”
Population growth:
There are more people alive today, and therefore, more money, more crime, and more of everything than there was in the past. Increases are usually the norm.
Other mitigating factors:
For example, Swedish women have a higher mortality rate, but not because of bad healthcare. It’s because there are more elderly women in Sweden.
Mark Twain said it best: “There are three kinds of lies – Lies, Damned Lies, and Statistics.”