Findings from indirect, non-experimental, and non-human studies are inconclusive because of mixed results from animal studies, the correlational nature of the data used, and the vagueness of observational studies. Studies concentrating on behaviors similar to aggression have indicated complicated alcohol- related changes and shown that many factors, such as type of task or gender, may interact with alcohol consumption. Results of direct and experimental studies, based on the original Buss aggression- machine paradigm, indicate that alcohol intoxication does not per se increase aggressive human responding, but may interact with different types of provocative circumstances to increase aggression. This article describes the Gustafson studies, conducted in Sweden, which were based on refinements of Buss' paradigm. This laboratory research showed that the relationship between alcohol intoxication and aggressive behavior is a complex one and part of the pharmacological nature of alcohol consumption. Aggression may occur when certain psychological and social factors, in addition to alcohol, are present. 101 references