Conditions
under which an intervention might be expected to affect many aspects
of a probability distribution (cont.)
Example 3:
Effects on rates
·The
intervention affects rates
·The outcome
measures a cumulative state
·Arbitrarily
complex mean-variance relationships can result