The Distinguishing-Variable Test

mental-model

To explain an exceptional case, eliminate factors it shares with ordinary cases. The useful candidate is a variable that actually distinguishes the target from the comparison set.

Why is India exceptionally populous? “Because people have sex” sounds causal—until you notice that the whole world enjoys sex. A fact can be necessary for an outcome and still explain none of the difference you are trying to understand.

E1

Explanation requires contrast

An explanation depends on a comparison: India versus other countries, a successful firm versus its peers, or one failure versus similar successes. If a proposed cause appears on both sides, it cannot by itself account for the gap between them. The test therefore has three moving parts: define the target, choose a defensible comparison set, and search for variables whose presence, intensity, or timing differs. This does not prove causation; it filters out explanations that have no distinguishing power.

E1

Where it shows up

The population non-explanation

Sex is relevant to population growth, but its near-universality makes it useless for explaining why India, specifically, became the most populous country. The question must shift from “What exists in India?” to “What differs between India and otherwise comparable countries?”

E1

Difference is not yet cause

A distinguishing variable is only a candidate explanation. It may be an effect, a coincidence, or a proxy for a Confounding Variable. The test also fails when the comparison set is badly chosen: change the peers, and the supposedly distinctive feature may disappear.

Build the comparison before the story

Tomorrow, when explaining an outlier, write down three genuinely comparable cases first. For every proposed cause, ask: Is this also present in the peers? Cross out anything shared, then investigate the remaining differences for timing, mechanism, and alternative causes.

Episodes that teach this