General information, not medical advice. If you have concerns about a child’s attention or behaviour, speak to a paediatrician.
Ask a room of parents whether sugar makes children hyperactive and you will get near-unanimous agreement, usually accompanied by a specific memory: a birthday party, a Halloween evening, a grandparent who overdid it. The evidence is right there. Everyone has seen it.
The controlled trials have looked, repeatedly, and found nothing. And one small experiment from 1994 explains the gap better than any amount of debunking.
What the trials found
In 1995, Mark Wolraich and colleagues published a meta-analysis in JAMA pooling 23 studies in which children were given either sugar or a placebo under double-blind conditions — neither the child, the parent, nor the person doing the rating knew which.
The pooled effect on behaviour was not significantly different from zero. The authors also looked specifically at the subgroups where an effect was most expected: children described by their parents as sugar-sensitive, and children diagnosed with attention-deficit disorder. No significant effect there either. The confidence intervals were narrow enough to rule out anything but a very small effect.
Wolraich’s own earlier trial, published in the New England Journal of Medicine in 1994, had gone further than most. Families were put on controlled diets for three weeks at a time — high sucrose, high aspartame, or high saccharin as placebo — with all food supplied by the study. Across 25 normal preschool children and 23 school-age children whose parents described them as sugar-sensitive, none of 39 behavioural and cognitive measures showed a meaningful difference between diets.
The experiment that tested the observers
The more revealing study was published by John Hoover and Richard Milich in the Journal of Abnormal Child Psychology in 1994, and its design is the whole story.
Thirty-five boys aged five to seven came to a laboratory with their mothers. Each mother was told her son would receive a large dose of sugar, or that he would receive a placebo. The boys were then given a drink and the pairs were observed interacting.
Every single boy received the placebo. Not one child in the study consumed sugar. The only variable that differed between conditions was what the mother had been told.
Mothers who believed their sons had received sugar rated the boys’ behaviour as significantly more hyperactive than mothers in the control condition — watching children who were, chemically speaking, in exactly the same state. The observed differences in the mothers’ own behaviour were just as striking: those in the sugar condition stayed physically closer to their sons, criticised more, and maintained more active supervision.
The belief did not merely bias the rating. It changed how the mother behaved toward the child, which is a mechanism by which the belief could partly create the behaviour it predicted.
Why the birthday party is not evidence
The everyday observation isn’t wrong about what people see. Children at parties are wound up. The inference about the cause is what fails.
Sugar arrives in a bundle. Birthday cake comes with fifteen other children, unfamiliar rooms, presents, games, staying up late, no nap, and a parent watching more closely than usual because there are witnesses. Halloween sweets come with costumes, darkness and running between houses. The one variable people isolate is the only one they can name, and it happens to be the one that correlates with every other variable in the room.
There is also a selection effect in memory. Nobody records the quiet afternoons when a child has a biscuit and then reads a book. The confirming instances are memorable; the disconfirming ones are not events at all.
What this does not mean
This is a narrow claim and it should stay narrow. The research is about acute behavioural effects of sugar over hours. It says nothing about dental caries, where sugar’s role is well established, and nothing about the long-run metabolic case for limiting added sugar, which rests on entirely different evidence.
There are also genuine dietary conditions affecting behaviour — hypoglycaemia, certain metabolic disorders, and for a small number of children, documented food sensitivities. Those are specific, diagnosable and rare. They are not what the sugar-hyperactivity belief describes.
And individual variation is real: it is possible that a small subset of children respond to a sugar load in ways the aggregate data washes out. Wolraich’s meta-analysis looked for exactly that subgroup and did not find it, but absence of detection in a pooled analysis is weaker evidence than a positive finding would have been.
The part worth keeping
The interesting result here is not really about sugar. It is that Hoover and Milich found a way to measure something usually invisible: the observer’s contribution to what is observed. Thirty-five children behaved however five-to-seven-year-olds behave in a strange room, and what their mothers saw depended on a sentence spoken to them beforehand.
That finding generalises well beyond snacks. Most of what parents believe about what sets their children off is built the same way the sugar belief was — from a small number of vivid, confounded, remembered occasions, with no control condition available.
Sources
- Wolraich ML, Wilson DB, White JW. The effect of sugar on behavior or cognition in children: a meta-analysis. JAMA 1995;274(20):1617-1621.
- Wolraich ML et al. Effects of diets high in sucrose or aspartame on the behavior and cognitive performance of children. New England Journal of Medicine 1994;330:301-307.
- Hoover DW, Milich R. Effects of sugar ingestion expectancies on mother-child interactions. Journal of Abnormal Child Psychology 1994;22(4):501-515.
