Current methods would allow reasonable predictions of long-term effects of pesticide application if three changes were instituted. First, more population-based laboratory studies should be applied in predictive pesticide risk assessment. Second, ERA should include as much effort on collating and integrating ecological knowledge into the assessment in Tier 1 as is currently expended on gathering chemical and toxicological information on exposure and effects. Production of a formal conceptual ecological risk assessment model for each product or active substance for which authorization is sought would provide an appropriate framework for integrating and applying such knowledge. Third, in acknowledgment of the uncertainties in the predictive risk assessment process, more postauthorization monitoring should be done. The application of Occam's razor to pesticide risk assessment makes good sense, as it does in any other field of science. However, we must take care that simplicity in risk assessment process does not lead to oversimplification: Essentially all science is the study of either very small bits of reality or simplified surrogates for complex whole systems. How we simplify can be critical. Careless simplification leads to misleading simplistic conclusions. (Slobodkin 1994).