Seven theories have been offered to explain these improvements in behavior: placebo effects, maturation effects, reactive hypoglycemia, the reduction of marginal malnutrition, neurotransmitter (serotonin) uptake, food intolerances, and absolute low blood sugar. Juvenile facilities in Alabama, Florida, and Virginia empirically tested these seven theories. A 36-month time-series design involving 125 Alabama juveniles showed a 61 percent reduction in antisocial behavior which was only consistent with the malnutrition theory. The Virginia site used an a-b-a-b-a design for 24 months. Support was present for the marginal malnutrition hypothesis on 856 children. The testing of the reactive hypoglycemia-causing-crime theory with 35 Florida delinquents showed that hypoglycemia does exist in delinquents, but the frequency is too low to explain the 47-percent rate of improvements in behavior. However, a significant but weak correlation was found between low blood sugar and behavior which is also consistent with marginal malnutrition theory. The data was not consistent with placebo effects, maturation effects, neurotransmitter changes, food intolerances, nor reactive hypoglycemia as the primary causes of the 47 percent improvements in behavior. (Author abstract)