Understanding and Preventing Violence, Volume 2: Biobehavioral Influences (1994)
Chapter: RESEARCH NEEDS FOR ASSESSING THE EFFECTS OF SUGAR ON VIOLENT BEHAVIOR
have any effects and if the independent variable is operative (Hirsch, 1987). Moreover, even if sugar intake was reduced, the diets consumed during the two periods varied sufficiently in nutrient content to make it difficult to attribute any behavioral change to reduced sugar intake.
A second problem with these studies is that appropriate behavioral techniques were not used. None of these studies used standard double-blind procedures. Both subjects and institutional officials were aware of the dietary changes. The subjects' awareness of the changes, and also simply knowing that they were in a study, could have led them to alter their behavior. Furthermore, because the observers were aware of the nature of the studies, their expectations may have influenced their observations.
The nature of the dependent variable also poses a problem. Official records of disciplinary actions were used to assess changes in violent behavior. In many institutions, the staff have the discretion to, or not to, record an incident, and variation over time in the proportion of incidents reported may lead to erroneous results (Gray, 1986). Also, in some of these studies the dietary changes were made during the last portions of the subjects' institutional stay (Schoenthaler, 1983a). One might expect that the number of disciplinary actions would decrease as the juveniles either learned the rules or learned not to get caught.
In some of these studies, concern also must be expressed for the changing nature of the subject population. Some of the juveniles were included in both the control and the treatment condition, whereas others were in one condition but not the other. Finally, questions have been raised with respect to the statistical methods used in these studies (Gray, 1986; Pease and Love, 1986).
Taken together, the studies by Schoenthaler provide little convincing evidence for the claim that sugar intake contributes to antisocial behavior. These studies are flawed by faulty experimental design and inappropriate statistical analyses, and leave open the question of whether nonspecific factors were responsible for the changes attributed to diet.
RESEARCH NEEDS FOR ASSESSING THE EFFECTS OF SUGAR ON VIOLENT BEHAVIOR
Dietary replacement studies conducted in correctional institutions could contribute valuable information about the effects of dietary variables on antisocial behavior. However, these studies must be rigorously conducted and carefully controlled. The dependent
