Two research approaches are described that were used to determine behavioral changes following a diet of Lake Ontario fish. Approach 1 involved the correlational method, in which human subjects voluntarily ate contaminated Lake Ontario fish. Demographic information, data on the amount of Lake Ontario fish consumed, and control variables were obtained during an interview. Respondents' behavior, as well as the behavior of their children, then was measured. Because subjects were not assigned randomly to eat or not eat Lake Ontario fish, other variables that might have influenced both consumption of fish and behavior had to be considered. Therefore, confounding variables were measured and their influence controlled for using statistical techniques. Approach 2 involved the experimental method using laboratory rats, where subjects were assigned randomly to receive a diet of environmentally contaminated Lake Ontario salmon, relatively uncontaminated Pacific Ocean salmon, or no salmon. Since the rats fed Lake Ontario salmon behaved differently than the other two groups on nine tasks, it was concluded that the contaminants in Lake Ontario salmon caused behavioral changes. Random assignment of subjects to groups eliminated competing explanations. If similar behavioral (e.g., emotional or cognitive) results were obtained using the two approaches, then the results utilizing rats probably could be generalized to humans, and the correlational results found in humans probably were due to a cause and effect relationship.