Local Evidence Beats Imported Intuition

principle

When a claim may change across populations or settings, replace imported assumptions and personal instinct with local measurement. Context can alter the size of an effect without removing the underlying constraint.

Indians accustomed to extreme heat might seem less vulnerable to it than workers represented in Western research. Yet Indian factory data found productivity falling by 2–4% for every 1°C increase in temperature.

Adaptation changes exposure, not necessarily the constraint

Evidence travels badly when populations, workplaces, incentives, or environments differ. A foreign estimate may therefore miss the local magnitude of an effect. But rejecting foreign evidence does not make intuition a reliable substitute: familiarity with heat can feel like adaptation even while output declines. Local data resolves both errors by testing the proposed mechanism where the decision will actually operate. This is the practical bridge between WEIRD Sampling Bias and Validity of Intuition: question whether a finding generalizes, then measure rather than guess.

Where it shows up

Heat on Indian factory floors

The Indian studies did not merely import a Western productivity estimate. They tested the relationship locally and still found a measurable penalty, showing that population differences can revise an estimate without making the underlying effect disappear.

Local is not automatically causal

A local correlation can still mislead if heat rises alongside other productivity-reducing conditions. Locality answers whether the evidence comes from the relevant setting; it does not, by itself, settle Confounding Variables or prove that every workplace will experience the same 2–4% decline.

Measure the claim where it will be used

Before adopting or dismissing a context-sensitive claim, choose one local outcome and record it across the relevant condition—for example, output at different factory temperatures. Compare the observed relationship with the imported estimate, then investigate plausible confounders before acting.

Episodes that teach this

High Temperature Kills Productivity | Future IQ 4,607 views
The transcript challenges Western data for Indians, then cites Indian factory studies showing productivity fell 2-4% for every 1 C increase.