Loading...
In 1974, Amos Tversky and Daniel Kahneman ran one of psychology's most unsettling experiments at the University of Oregon. They set up a wheel of fortune rigged to land on either 10 or 65, then invited participants to spin it. After the wheel stopped, they asked each person two simple questions: Is the percentage of African countries in the United Nations higher or lower than the number on the wheel? And what is your best estimate of the actual percentage? The wheel had nothing to do with African geopolitics. Everyone knew it was random. Yet the results were staggering. Participants who saw 65 estimated an average of 45% of UN members were African nations. Those who saw 10 guessed just 25%. A completely arbitrary number—one the participants watched being generated by a spinning wheel—sh...
Popular framing: Naive people fell for an irrelevant number; sophisticated thinkers wouldn't.
Structural analysis: Under uncertainty, the brain treats whatever number is salient as a starting point and adjusts insufficiently from there — even when the source is visibly random. The anchor distorts the prior in a way that bayesian-thinking corrections never fully overcome; framing makes a pure-noise input into apparent signal, and the map-territory gap between the spinning wheel and African geopolitics is invisible to the heuristic that produces the estimate. Experts fall to the same effect because the mechanism runs below deliberation; the architecture of estimation under uncertainty, not the sophistication of the estimator, sets the bias.
The popular framing locates the problem inside the individual, implying individual solutions (awareness, training). The structural framing locates the problem in the information environment and the sequential nature of belief updating — implying that the real leverage points are institutional: who sets the anchor, when, and with what accountability. Closing this gap matters because interventions built on the wrong model (educating individuals vs. redesigning elicitation processes) will consistently underperform.
