Causal Inference in Environmental Impact Studies

摘要: Justification of a cause-and-effect relationship in environmental impact studies is com- plicated by inherent qualities data. Lack randomization and replication invalidate the use inferential statistics for inferring causal link, place special demands on descriptive arguments causation. Assembly rules have been developed epidemiology provide rigorous structure analysis. Explicit assembly making allows investigators to efficiently organize, study, present available evidence. Within this framework, statistical tests can be used determine if populations at study sites were different after impact, but establishment link between observed difference based an argument. Causal inference means argument consistent with scientific method strong increases likelihood correct conclusions.