A study exploiting an externally generated exposure contrast to strengthen causal inference. Causal interpretation still requires assumptions about the comparison, concurrent changes and selection; an event outside the investigator's control is not automatically as-if random.
