The US Food and Drug Administration convened its Food Advisory Committee in March 2011 to review the evidence on food colors and hyperactivity in response to the Southampton study and a citizen petition requesting a ban. After reviewing the available studies, the committee voted 13-1 that more research was needed, but 8-6 against requiring warning labels.

The FDA concluded that the evidence did not establish a causal relationship between food color consumption and hyperactivity in the general population of children. The agency noted that while some children may be particularly sensitive to food colors, this did not warrant population-wide labeling or restrictions.

The FDA specifically criticized the effect sizes reported in the research as being too small to be clinically meaningful. An effect size of 0.18 means that food colors explain less than 1% of the variation in children''s behavior - meaning that more than 99% of hyperactive behavior is caused by other factors. The FDA maintained that the current approved uses of food color additives are safe.

The European Food Safety Authority published a highly critical assessment of the Southampton (McCann 2007) study in 2008. Despite the study being published in The Lancet, EFSA found significant methodological problems that undermined its conclusions.

EFSA noted considerable uncertainties in the study, including: inconsistent results between the two age groups tested, inability to identify which specific dyes or the preservative sodium benzoate was responsible for the observed effects, lack of dose-response data, and the modest magnitude of the behavioral changes observed.

EFSA concluded that there was no basis for changing the acceptable daily intake of any of the food colors tested or of sodium benzoate. This is notable because EFSA is generally considered more precautionary than the FDA, and even they did not find the evidence sufficient to warrant regulatory action beyond eventually requiring warning labels as a precautionary measure.

A 2012 analysis in Neurotherapeutics examined the clinical significance of the reported effect sizes for food dye effects on behavior. The meta-analytic effect size was 0.18 for parent ratings (dropping to 0.12 after publication bias adjustment) and 0.22 for high-quality studies of color additives alone.

To put these numbers in perspective: an effect size of 0.18 means that food colors account for roughly 0.8% of the variance in children''s hyperactive behavior. By comparison, the effect size for stimulant medication (like Ritalin) in treating ADHD is approximately 0.80-1.00, meaning medication explains about 16-25% of the variance. Food dyes explain about 20-30 times less behavioral variation than ADHD medication corrects.

For an individual child, an effect of this magnitude would be virtually imperceptible. The difference between a child''s behavior on a day with food dyes versus without would be too small for parents, teachers, or clinicians to notice in normal observation. Only sensitive statistical analysis of large groups can detect effects this small.

The meta-analyses consistently show a discrepancy between different raters. Parent ratings show a small but statistically significant effect of food dyes on behavior (effect size 0.18), but teacher and independent observer ratings show a much smaller, non-significant effect (effect size 0.07).

This discrepancy matters because parent ratings are the most susceptible to expectation bias. Even in double-blind studies, parents may unconsciously rate their child''s behavior differently based on their pre-existing beliefs about food dyes. Teachers and trained observers, who typically observe children in a more structured setting alongside other children, provide a more objective behavioral assessment.

If food dyes truly caused noticeable hyperactivity, the effect should be consistent regardless of who is doing the observing. The fact that the effect essentially disappears when independent observers measure behavior rather than parents raises serious questions about whether the reported effect is real or is partly an artifact of parental expectations.

A key limitation of the influential Southampton (McCann 2007) study is that it tested mixtures containing both artificial food colors and sodium benzoate (a preservative), not individual dyes in isolation. The two test mixtures each contained multiple dyes plus sodium benzoate.

This means the study cannot determine which ingredient caused the observed behavioral effects. Sodium benzoate alone could be responsible, one specific dye could be responsible, or the combination might produce effects that individual ingredients would not. Without testing individual ingredients, the blanket conclusion that "food dyes cause hyperactivity" is not supported by this study.

Subsequent research has not definitively resolved this question. Some studies have tested individual dyes, but with smaller sample sizes and less consistent results. The inability to identify the specific causative agent(s) has been a major barrier to regulatory action, since banning all food dyes based on a study that cannot identify which ones are problematic is a significant policy step.

Bateman and colleagues published the first Southampton study in Archives of Disease in Childhood in 2004, testing the effects of food colors and sodium benzoate on hyperactivity in 277 three-year-old children. The study found a significant effect of the active drink compared to placebo.

When the same research group expanded their study in 2007 (the McCann study), they tested both 3-year-olds and 8/9-year-olds. The results were inconsistent between age groups: one mixture affected one age group but not the other, and the pattern differed between mixtures. This inconsistency made interpretation difficult.

If food dyes have a genuine biological effect on brain chemistry and behavior, one would expect the effect to be consistent across age groups when the same dyes are tested. The inconsistent results between age groups and between the different dye mixtures suggest that the findings may not be robust and could be influenced by random variation in the data.

A fundamental limitation of the food dye and hyperactivity research is that most clinical trials have tested arbitrary mixtures of dyes at doses that may not reflect real-world consumption. The Southampton study used two specific mixtures, but these particular combinations were chosen for the study rather than representing what any child would typically consume.

Real-world exposure to food dyes involves dozens of different dyes consumed in varying amounts across many different foods throughout the day. The challenge study design - giving a single large dose of a specific mixture and measuring behavior hours later - does not replicate this pattern.

No individual food dye (Red 40, Yellow 5, Yellow 6, etc.) has been conclusively demonstrated to cause hyperactivity in a large, well-designed randomized trial testing that specific dye alone. Without identifying which specific dyes are problematic, regulatory action must be based on the precautionary principle rather than established causation for specific substances.

The 2012 Nigg meta-analysis included a publication bias analysis, which revealed that the reported effect size of 0.18 was likely inflated by the selective publication of positive results. After adjusting for publication bias, the effect size dropped to 0.12.

An effect size of 0.12 is classified as "trivially small" in behavioral science. In clinical terms, a child who consumes food dyes versus one who does not would show a behavioral difference equivalent to about 1-2 points on a 60-point hyperactivity scale - a difference that would be impossible for any parent, teacher, or clinician to detect without statistical instruments.

In the broader context of what causes hyperactive behavior in children, genetics accounts for approximately 70-80% of ADHD risk. Environmental factors like prenatal tobacco exposure, lead exposure, prematurity, and adverse childhood experiences each have much larger individual effect sizes than food dyes. Focusing on food dyes as a "cause" of hyperactivity is disproportionate to their actual contribution.

The hypothesis that food additives cause hyperactivity dates back to 1973 when Dr. Benjamin Feingold, a pediatric allergist, claimed that 30-50% of hyperactive children could be cured by removing artificial colors, flavors, and preservatives from their diet. His claims generated enormous public interest and led to multiple controlled studies.

The controlled trials conducted in the 1970s and 1980s to test the Feingold hypothesis produced mostly negative or inconclusive results. A 1982 National Institutes of Health consensus conference concluded that controlled studies had not confirmed that dietary changes could help more than a small percentage of hyperactive children, if any.

The modern resurgence of interest in food dyes and behavior is largely driven by the Southampton studies, but the longer historical context shows that this claim has been tested repeatedly over 50 years with consistently weak or negative results from controlled trials. Each generation of studies finds small, inconsistent effects that do not meet the threshold for clinical significance.

The 2012 Nigg meta-analysis analyzed both challenge studies (where dyes are added to the diet and behavior is measured) and restriction diet studies (where dyes are removed and behavior is measured). An interesting pattern emerged: restriction diets showed larger effect sizes than challenge studies.

This discrepancy is important because restriction diets typically change many things beyond just food dyes. Families that eliminate artificial colors tend to simultaneously improve their overall diet quality - eating more whole foods, fewer processed snacks, more fruits and vegetables, and preparing more home-cooked meals. The behavioral improvement may come from these broader dietary changes rather than from the specific removal of food dyes.

If the benefit were specifically due to removing food dyes, challenge and restriction studies should produce symmetrical effect sizes (adding dyes should worsen behavior by the same amount that removing them improves it). The asymmetry suggests that confounding dietary changes in restriction studies inflate the apparent benefit of dye removal.

A 2022 review published in Cureus examined the evidence specifically for blue food coloring (Brilliant Blue FCF / Blue 1) and ADHD symptoms in children. The review found that the evidence for this individual dye was extremely limited.

Most clinical trials have tested mixtures of multiple dyes, making it impossible to attribute effects to any single color. The few studies examining individual dyes had small sample sizes and produced inconsistent results. Blue 1, despite being one of the most commonly used food dyes, has been specifically tested in very few controlled behavioral studies.

This highlights a broader problem: the claim that "artificial food dyes cause hyperactivity" implies that all dyes share this property, but the evidence does not support such a broad conclusion. Different dyes have different chemical structures and may have different biological effects. Grouping all artificial colors together and attributing a single behavioral effect to the entire class is scientifically imprecise.