Various methods are available for investigating the effects of environmental conditions or stress on fish ranging from assessments at the individual level to the population and community level. Each method, however, has limitations and advantages, and the type of method used influences the interpretation of stressor effects on fish health. We compared and evaluated three methods for investigating the effects of contaminants on fish health for populations in a stream receiving a variety of industrial effluents including PCBs, heavy metals (primarily mercury), and chlorine. These methods are" (1) individual bioindicator variables including measurements of biochemical, physiological, histopathological, and organismal-level responses, (2) statistically based integrative indices based on multivariate responses, and (3) integrative ecologically relevant measurements such as population and community-level responses. Many of the individual-level measurements reflect biochemical and physiological conditions and are therefore short-term response indicators of stress. Use of several individual variables in multivariate discriminant analysis allows a more integrative approach for evaluating response of fish to stressors. Community-and population-level measurements integrate the responses to a variety of environmental conditions and therefore may be less reflective of contaminant-induced stress. The three methods used together provide the best strategy for evaluating fish health in order that environmentally sound decisions can be made regarding the implications of contaminant-related stress on aquatic ecosystems.