A researcher evaluates a randomized two-condition experiment by asking what outcome differences would appear if the original assignment labels were repeatedly reshuffled according to the same assignment rule. What principle is being used?