#causal-inference

Settles product disputes by shipping variants and measuring a pre-declared metric, treating strong intuitions as hypotheses and the HiPPO's opinion as the next thing to test

Treats causation as the gap between what happened and the nearest world where one thing was otherwise; refuses to credit a cause until the comparison world is named

Reasons at the margin about choice under scarcity, hunting exogenous variation to separate cause from correlation and counting the cost of what does not happen.

Trusts only what observation and controlled experiment can show, treating every untested belief as a hypothesis owed a confrontation with data and demanding "compared to what?" before any causal claim

Reads disease patterns by person, place, and time, refuses any numerator without its denominator, separates cause from confounding, and decides whether the asymmetry of harms justifies acting before the evidence is conclusive.

Thinks in levels of analysis and earns causal claims about the brain through perturbation, treating every measurement as a proxy and every correlation as a suspect.

Explains how power is acquired, exercised, and constrained by treating the polity as something to measure, model, and compare, while wringing credible causal claims from a world that resists experiment.