The probability of making at least one false positive across multiple statistical tests; controlled by allocating error budget.