🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
regression_fallacy
The regression fallacy attributes a natural statistical regression to the mean to a specific cause. After an extreme event (unusually good or bad performance), outcomes tend to return toward the average simply due to random variation. People mistakenly credit or blame whatever intervention happened between the extreme event and the regression, confusing a statistical inevitability with a causal effect.
"My back pain was terrible yesterday, so I tried a crystal healing session. Today it's much better -- the crystals clearly worked!" (The pain was likely to improve regardless due to natural fluctuation.)
A football coach benches his star player after an unusually poor game and plays a backup instead. The team wins the next match, and the coach concludes: 'Benching him was the right call — it completely turned our season around.' (The star player was statistically likely to perform closer to his average regardless.)
A student scores unusually low on a practice exam, then starts wearing a specific 'lucky' bracelet. On the next test, her score returns to normal, and she tells her classmates: 'This bracelet genuinely works — my grades shot back up the moment I started wearing it.'
Binary (yes/no) questions an LLM must answer to identify this aspect:
Did an extreme or unusual measurement precede a more typical one?
Type: binaryIs a causal explanation being given for what could be regression to the mean?
Type: binaryWould the return to normal be expected statistically without any intervention?
Type: binaryThe regression fallacy attributes a natural statistical regression to the mean to a specific cause. After an extreme event (unusually good or bad performance), outcomes tend to return toward the average simply due to random variation. People mistakenly credit or blame whatever intervention happened between the extreme event and the regression, confusing a statistical inevitability with a causal effect.
People seek causal explanations for changes they observe and do not intuitively understand regression to the mean. The temporal sequence of intervention followed by improvement creates a compelling but false causal narrative.
Explain regression to the mean: extreme values are statistically likely to be followed by less extreme ones, regardless of any intervention. Ask for controlled comparisons rather than before-after anecdotes.
Widespread in alternative medicine testimonials, sports commentary ('sophomore slump'), educational interventions applied after poor test scores, and business decisions based on one bad quarter.
Use these tools to detect, analyze, or train this aspect.