I for one vote for E. The argument states that the additives contribute to behavior problems in hyperactive children. E states that opposite: the additives in fact inhibit behavior problems. After the experiment, there can be, for example, 75% of problem children.
But there is no evidence for E. 60% had behavior problems before the diet, but only 30% had behavior problems after the diet. While some children may indeed have increased behavior problems after the diet, there is no evidence of that, and overall, behavior problems decreased.
B is wrong, for it says that only children who changed to a low-additive diet were studied. The argument does not support it. The group was on one diet; the same group was on the other. The same group was studied: originally 60%, after 30%.
B is correct in that there was only one group of children studied. It's true that this one group was evaluated before and after the diet change... but a a second group should be studied during the same time period that does not undergo a diet change. What if the school implemented a new discpline system during the study? If you only evaluated one group, you might attribute the decrease in problems to the diet change. But if you looked at two groups (one with a diet change and one without), and there were similar decreases in behavior problems, you wouldn't be as quick to say the decrease was due to a diet change. (In fact it would seem the decrease was due to the new discipline system.)