and independent variables must be adequately specified, and valid and reliable methods for measuring these variables must be determined.

The starting point in this research should be accurate measurements of nutritional intake. Appropriate dietary changes can then be instituted to reduce sugar intake. It is important that double-blind procedures be used. To accomplish this goal, artificially sweetened foods and beverages could be substituted for sugar-containing items. To eliminate the possibility of order effects, a crossover procedure in which half of the subjects receive sugar-containing foods first and then artificially sweetened items—whereas the remaining subjects are tested in the reverse order—should be employed. All subjects should be examined in both treatment conditions. Food intake must be measured to determine if dietary manipulations actually reduce sugar intake.

The second major question that must be addressed in this type of research is what is the dependent variable. An objective criterion of antisocial or violent behavior must be established. Additionally, by whom and when the behaviors will be measured must be adequately specified.

It must also be realized that a number of potent variables (e.g. overcrowding, drug use) may override the behavioral effects of dietary modifications. Diet is only one of the many variables that can influence human behavior. Therefore, it is important that studies on diet and behavior be conducted in institutions offering the least number of extraneous variables. Only then can definitive conclusions about the role of diet in antisocial behavior be reached

Attention deficit hyperactivity disorder (ADHD) is characterized by developmentally inappropriate inattention, impulsive behavior, and significantly elevated levels of motor activity. At home, attentional difficulties may be manifested by failure to follow through on tasks and the inability to stick to activities, including play, for appropriate periods of time. In school, the child with ADHD is inattentive and impulsive, and has difficulty organizing and completing work. In approximately half of the cases, the onset of ADHD is before the age of 4. As many as 3 percent of preadolescent children may suffer from ADHD, with the disorder six to nine times more common in boys than in girls (American Psychiatric Association, 1987).

With respect to violent behaviors, children with ADHD frequently have difficulties interacting with their peers and may be seen as overly aggressive. Additionally, many children with ADHD develop oppositional defiant disorder or conduct disorder later in childhood. Among those with conduct disorder, a substantial number have antisocial personality disorder as adults. Follow-up studies of children with ADHD have revealed that approximately one-third display some symptoms of the disorder as adults (American Psychiatric Association, 1987).

One pervasive idea about sugar is that it can lead to hyperactivity in children. This idea has been accepted by both educational professionals and parents (McLoughlin and Nall, 1988). Evidence for this association came first from a correlation study by Prinz and colleagues (1980) who compared sugar intake and behavior in hyperactive and normal 4- to 7-year-old children. Seven-day dietary records were obtained for all children. Trained observers, blind to the nature of the experiment, then rated the children during play for a variety of behaviors including destructive-aggressive acts (attempts to damage, strike, kick, or throw objects in the room); restlessness (repetitive arm, leg, hand, or head movements); and overall movement. Hyperactive and normal children consumed equivalent amounts of sugar-containing foods. However, for children with ADHD the amount of sugar products consumed, the ratio of sugar to nutritional foods (foods containing neither sugar nor refined carbohydrates), and the carbohydrate/protein ratio were all positively correlated with destructive-aggressive and restless behaviors. In contrast, in the normal group, sugar intake was not associated with destructive-aggressive acts but was correlated with total body movements. Although Prinz and coworkers (1980) carefully interpreted their work as only suggestive evidence of a role for sugar in ADHD, the data were rapidly interpreted in the popular press as demonstrating a causal relation between sugar and hyperactivity. Although the majority of studies on sugar and hyperactive behavior in the last decade have not confirmed the work of Prinz et al. (1980), the original interpretation of this work continues to be part of the folklore about diet and behavior.

During the last 10 years, both correlational and dietary challenge studies have assessed the effects of sugar on hyperactive behavior. For example, in an attempt to replicate the findings of Prinz et al. (1980), Wolraich and coworkers (1985) examined the association between sugar intake and performance on 37 behavioral/cognitive variables in hyperactive and normal boys. In contrast

to the Prinz et al. (1980) study, sugar intake was not reliably related to destructive-aggressive behavior in children with ADHD.

One obvious problem with correlational studies is that it is impossible to establish cause and effect relations. To determine causality, experimental studies that systematically manipulate sugar intake and observe its effects on behavior are required. The most common procedure used to accomplish this goal has been the dietary challenge. In the majority of these studies, the behavior of hyperactive and normal children has been rated for several hours after they have consumed a sugar-containing food or beverage, or a placebo containing an artificial sweetener. For example, Behar and colleagues (1984) investigated the effects of sucrose on behavior in 6- to 14-year-old boys whose parents believed that they had adverse responses to sugar. After an overnight fast, the boys were given a beverage containing either glucose, sucrose, or saccharin. Motor activity, spontaneous behavior, and performance on psychological tests were then measured for 5 hours. The negative behavioral effects reported by parents to occur after sugar intake were not observed. In fact, the only significant finding was that the boys were less active 3 hours after consuming the sugar-containing beverages than after consuming the saccharin-containing beverage. Similar negative findings have been reported in a number of studies (e.g., Wolraich et al., 1985; Ferguson et al., 1986; Mahan et al., 1988).

One problem with the preceding studies is that they were conducted in a laboratory environment. This rather unnatural setting may mask sugar's effects on behavior. To circumvent this problem, Milich and Pelham (1986) incorporated a dietary challenge into a treatment program for hyperactive boys. The boys were fasted overnight and at 8 a.m. given Kool-aid containing either glucose or aspartame. A double-blind crossover design was used. Beginning one-half hour after drinking the Kool-aid, the boys were evaluated for positive (e.g. following rules, sharing) and negative (e.g. noncompliance, teasing, name calling) behaviors and for classroom performance. Sugar did not adversely affect behavior or classroom performance. Similar results were obtained by Kaplan et al. (1986) who examined the behavioral effects of breakfasts in which dietary protein, fat, and carbohydrate were controlled while sucrose and total calories were varied.

Because the original study by Prinz and colleagues (1980) suggested that younger children might be more susceptible to the adverse effects of sucrose, Kruesi and coworkers (1987) performed a dietary challenge study in preschool-aged boys with alleged sugar

reactivity. Again, no significant differences were observed as a function of sugar intake.

Taken together, the results of dietary challenge studies do not support the idea that sugar plays a major role in ADHD. In studies using hyperactive and normal children of varying ages and employing a range of experimental situations, sugar intake had no effects on behavior. However, although experimental evidence is weak, parents or teachers continue to supply anecdotal reports of the deleterious consequences of sugar. How do we reconcile these differences?

Several factors could contribute to the differing views of scientists and parents or teachers. One limitation of dietary challenge studies is that in most, only a single dose of sucrose was used. The amount of sucrose used in these studies may have been too small relative to the children's normal daily intake. Larger amounts might have produced negative reactions. Similarly, challenge studies can be criticized because they do not allow for the assessment of chronic sugar intake. Cumulative sugar intake may produce behavioral effects not detectable in single challenge tests. To help solve these problems, precise dietary histories and dose-response determinations of sucrose's effects on behavior are required.

Another difficulty with challenge studies is the choice of an appropriate placebo. In most studies, either aspartame or saccharin has been used as the placebo. Although this procedure successfully blinds the subjects to the item they are consuming, it does not control for the fact that the challenge food not only contains sugar, but also provides substantial calories. This presents the possibility that any changes in behavior could be attributed to calories rather than to sugar. Additionally, it has recently been proposed that aspartame may have negative effects on behavior.

Time parameters may also be important in determining sugar's effect on behavior. Most experiments have limited behavioral observations to one time after sugar intake and may have missed the critical period for its effects on behavior.

To overcome the objections raised about dietary challenges studies, and to further investigate the hypothesis that sugar intake has negative behavioral consequences, Wolraich and his colleagues (1994) evaluated hyperactive behavior and performance on cognitive tests of school-age and preschool children placed on diets high in sucrose, aspartame, or saccharin. To ensure that the subjects ate only the specified diets, all food was removed from the children's