Large-Scale Perturbations: Opportunities for Innovation

Several approaches are discussed for statistical analysis of large—scale (and possibly unreplicated) ecological experiments. These include intervention analyses and comparisons of alternative models using Bayes' formula. Such techniques are unfamiliar to many ecologists and are not typically included in graduate curricula in ecology. I argue for increased training in these areas and for collaborations between statisticians and ecologists to develop innovative approaches to the analysis of large—scale perturbations.