You only date people who are at least this attractive OR nice
50%
0.00
What's Happening?
Berkson's Paradox (1946) reveals how selection based on a "collider"
creates spurious correlations that don't exist in the full population. When you only observe
a subset selected by a criterion involving two variables, those variables appear negatively
correlated—even if they're completely independent or positively correlated in reality!
🎯 The Key Insight
In the dating pool: If you only date people who are at least attractive OR nice
(a minimum standard), then among your dates, these traits will appear negatively
correlated. Why? The nice-but-plain people in your pool "made the cut" via niceness,
while the pretty-but-rude ones made it via looks. The corner with low-both gets excluded!
⚠️ Real-World Dangers
Hospital studies: Patients are selected by illness. A spurious
negative correlation between two diseases appears because people with neither
aren't in the hospital.
COVID-19 research: Studies of hospitalized patients showed
misleading correlations because severe illness was the selection criterion.
Try it: Move the threshold slider and watch the correlation in the
selected sample become increasingly negative as you raise standards—even though the
true population correlation stays at zero (or even positive)!