Sugar does not make children hyperactive. The trials settled this in the 1990s.

Sugar does not cause hyperactivity in children. A meta-analysis published in JAMA in 1995 pooled 23 within-subject studies from 16 reports, all of which fed children a known quantity of sugar, used an artificial sweetener as placebo, and blinded the children, their parents and the research staff to which was which [s1]. Across 14 separate measurement constructs, the weighted mean effect sizes ranged from −0.14 for direct observation to +0.30 for academic tests, and the 95% confidence interval for every one of the 14 included zero [s1]. The authors' conclusion was that sugar does not affect the behaviour or cognitive performance of children, and that the strong belief held by parents may be due to expectancy and common association [s1].

The trial that used doses no child would normally reach

The year before, a double-blind controlled trial had tested the question at a scale designed to find an effect if one existed. Two groups took part: 25 normal preschool children aged three to five, and 23 school-age children aged six to ten whose parents described them as sensitive to sugar [s2]. Each family followed three different diets in three consecutive three-week periods — one high in sucrose with no artificial sweetener, one low in sucrose sweetened with aspartame, and one low in sucrose sweetened with saccharin as placebo — with all three essentially free of additives, artificial colouring and preservatives [s2]. Behaviour and cognitive performance were assessed weekly [s2].

The doses were substantial. Preschool children on the sucrose diet took in a mean 5,600 ± 2,100 mg of sucrose per kilogram of body weight per day; the school-age group described as sugar-sensitive took in 4,500 ± 1,200 mg/kg [s2]. Aspartame and saccharin intakes on the respective diets were 38 ± 13 and 12 ± 4.5 mg/kg in preschoolers, and 32 ± 8.9 and 9.9 ± 3.9 mg/kg in the older children [s2].

Among the children whose parents considered them sugar-sensitive, there were no significant differences between the three diets on any of 39 behavioural and cognitive variables [s2]. Among the preschoolers, 4 of 31 measures differed significantly across diets, with no consistent pattern in the differences observed [s2] — the kind of scatter expected when 31 comparisons are run. The trial's stated conclusion: even when intake exceeds typical dietary levels, neither dietary sucrose nor aspartame affects children's behaviour or cognitive function [s2].

The adult version of the same belief

The related idea — that sugar delivers a burst of energy and alertness — has been tested in adults and fails in the same direction. A 2019 meta-analysis of 176 effect sizes from 31 studies in 1,259 participants found no positive effect of carbohydrate consumption on any aspect of mood at any time point after ingestion [s3]. What it did find ran the other way: carbohydrate administration was associated with higher fatigue and lower alertness than placebo within the first hour after eating [s3]. The authors describe the "sugar rush" as a myth [s3].

What the studies did not rule out

The 1995 meta-analysis is careful about the limits of a null result, and its caveat is the one most often dropped when it is cited. A small effect of sugar, or effects on subsets of children, cannot be ruled out [s1]. The studies pooled were within-subject designs of modest size, and the analysis was conducted on the literature as it stood in the mid-1990s [s1].

Two further limits matter for a parent reading this. None of these studies asked whether children who habitually consume large amounts of sugar differ over years from those who do not — they tested acute administration against placebo [s1] [s2]. And none of them bears on the well-established reasons to limit added sugar in childhood, which concern dental caries, energy density and cardiometabolic risk rather than behaviour. A finding that sugar does not cause hyperactivity is not a finding that sugar intake is unimportant.

What the 1995 review counted, and how

The meta-analytic method matters here because the studies being pooled were small. Studies were eligible only if they intervened by having subjects consume a known quantity of sugar, used an artificial sweetener as the placebo condition, blinded subjects, parents and research staff, and reported statistics from which effect sizes could be computed [s1]. Sixteen reports met those criteria, yielding 23 within-subject studies [s1]. Variables extracted included publication year, study setting, subject type and number, sex, age, sugar and placebo type and dose, prior dietary condition, measurement construct, and the direction of effect [s1].

Fourteen separate measurement constructs were analysed, and every one of the 14 mean effect sizes had a 95% confidence interval spanning zero [s1]. A single null on a single measure would be weak evidence; fourteen nulls across the constructs researchers had thought worth measuring is a stronger statement, and it is the reason this literature has not been substantially revisited.

Why the belief persists

The 1995 authors offered a mechanism for its durability rather than merely dismissing it: expectancy, and the common association between the settings where children eat sugar and the settings where they are excited [s1]. Birthday parties, holidays and outings supply both at once. When the placebo is indistinguishable and the observer is blinded, as in these designs, the effect disappears [s1] [s2] — which is the definition of an expectancy effect rather than a pharmacological one.

This article is informational and is not medical advice. Concerns about a child's attention or activity level are a question for a clinician, not a dietary experiment.

Sources

• The effect of sugar on behavior or cognition in children. A meta-analysis — JAMA, 1995
• Effects of diets high in sucrose or aspartame on the behavior and cognitive performance of children — The New England Journal of Medicine, 1994
• Sugar rush or sugar crash? A meta-analysis of carbohydrate effects on mood — Neuroscience and Biobehavioral Reviews, 2019