Research studies of a controversial link between hyperactivity in children and food chemicals challenge claims by a California doctor that a special additive-free diet is widely effective in controlling the condition, a University of Utah pediatrician says.

Dr. Esther H. Wender, who spoke at a National Institutes of Health conference yesterday, said that the overall conclusions from several well-designed scientific studies were "very clearly negative."

But Dr. Ben Feingold, a San Francisco allergist who popularized the idea that food colorings, flavorings and other chemicals stimulate hyperactivity in sensitive children, maintained that "careful implementation" of his dietary approach was effective in as many as 50 percent of the cases.

And parents testified to the benefits in their own children.

Wender said, however, that the diet has failed to pass scientific tests seeking a consistent effect when it is compared with other diets under controlled clinical conditions in which participants are unaware of the diet they were receiving. In an interview, she cited studies in Wisconsin, Pennsylvania, New York and Utah.

Dr. Judith Rapoport, a National Institute of Mental Health scientist who also spoke at the NIH meeting, agreed that the additive-free approach advocated by Feingold "doesn't turn out to make a difference in how kids behave."

One recent Canadian study, which has been criticized by others in the field, found that artificial food dyes in high amounts impaired hyperactive children's performances on a laboratory learning test, although not their behavior ratings. This raises a "larger issue" requiring additional research on the effects these chemicals could have on the population at large, Wender admitted.

Dr. Bernard Weiss, a University of Rochester toxicologist who found that artificial food colors affected two out 22 children in his study, argued that there may be a sub-group of very young children who are particularly sensitive and that it might be worthwhile to put them on a diet without dyes.

The value of defined diets in controlling childhood hyperactivity was the subject of a three-day NIH "consensus development conference." An outside panel will make recommendations today for professionals and the public.

As the meeting made clear, hyperactivity is not a "very clearly defined entity," said Dr. Gabrielle Weiss, a psychiatrist from Montreal's McGill University.

She said that it is characterized by excessive motor activity for a child's age, difficulties in sustaining attention, and impulsive behavior, with a duration of at least a year.

Conservative estimates suggest that about 1 percent of children may be hyperactive.

Rapoport said the proportion of children with some hyperactivity could go as high as 20 percent, a number also cited by Feingold. About four times as many boys as girls appear to be affected.

Feingold, 81, presented his theory and first case studies in the early 1970s and published two popular books on the subject.

He dismisses all of the negative scientific evidence, saying that his diet is too difficult to study and maintains that "there's no harm in trying it" because it might help. The diet, he said, means eliminating all foods with synthetic flavorings, colorings or preservatives and certain foods containing natural chemicals called salicylates, including fruits such as apples, grapes, and apricots.

Wender attributed the belief of many parents that the Feingold diet works to the "placebo effect," in which there is an expectation fulfilled.