PhenomenonReasoning & flexibility · Logical Fallacy · Entry 63
Berkson’s Paradox – when two things seem connected, but they aren’t
Does it seem like successful people are less kind because there aren’t many kind, successful people? That’s Berkson’s Paradox – a logical fallacy where an observed relationship between variables appears due to selection bias, not real causation.
Where it can show up
- Career & Personality – it seems like wealthy people are less kind, but kind, unsuccessful people just aren’t included in the sample.
- Hospitals – many COVID patients also have diabetes, but that doesn’t mean diabetes causes COVID – diabetics are just more likely to be hospitalized.
- College admissions – elite schools accept only top students, making it seem like there are no smart students with lower grades, when in fact, they just weren’t selected.
A practical countermeasure
- Watch for selection bias – not all groups are equally represented in the data.
- Check what actually causes the effect – sometimes, a connection appears simply due to how the data was collected.
- Don’t jump to conclusions – correlation does not imply causation.
