ON REJECTION RATES OF PAIRED INTERVENTION ANALYSIS: COMMENT

Murtaugh (2000, 2002) claims the Before-After, Control-Impact (BACI) approach to assessment of long-term local effects of a planned environmental alteration (such as a development) ignores serial correlation and assumes the Control and Impact values would have "parallel trajectories" in the absence of an impact. In the BACI approach, observations are taken at the altered ("Impact") site at a set of times before the alteration and another set after it, and matching observations are taken at one or more unaltered "Con-