Parents in the Trondheim Early Secure Study first answered questions about their children when the children were four. Every two years after that, from six until eighteen, the forms came back, filled in by parents, by teachers and eventually by the teenagers themselves. One of the recurring questions was about food: when this child gets upset, do they eat?

The paper is careful about that. The headline was not.

Francesca Parisi, a PhD research fellow at the Norwegian University of Science and Technology, and three colleagues ran the answers from 878 of those children through a model built to catch change inside each child. The clearest thing it caught sat in childhood. When a child between six and ten became moodier than was usual for that particular child, more comfort eating tended to follow two years later. Then the press coverage arrived on October 30, and the News-Medical headline put the risk somewhere else: “Children with intense mood swings face greater risk of emotional overeating in adolescence.” The abstract of the paper in Child Development places the effect, in its own words, “in childhood.”

Each child measured against their own average

The method explains why the childhood result carries more weight than most findings of its kind. A random-intercept cross-lagged panel model compares each child with that child’s own average over the years, not with other children, so under the model’s assumptions the stable differences between children are set aside. That matters, because in observational research on kids those stable differences usually do most of the explaining. The question the model answers is narrower: not whether moody children eat more than calm ones, but whether a child who becomes moodier than usual goes on to eat more than usual. Parents and teachers rated mood on the Emotion Regulation Checklist, which scores two things separately. Lability and negativity covers the volatile end, the quick anger and sudden swings. Regulation covers a child’s ability to calm down and recover.

Rising mood swings predicted rising comfort eating two years later, with an average standardized effect of 0.11 on parent ratings and 0.10 on teacher ratings. Put plainly, a child who swung about one standard deviation moodier than their own norm showed roughly a tenth of a standard deviation more comfort eating at the next wave. That sounds small until you set it against the benchmark the authors adopted for within-person effects, which are typically tiny: 0.07 counts as medium and 0.12 as large. The pattern also held whether the child was rated at home or at school, by observers who agreed with each other only modestly, with parent and teacher ratings correlating between 0.14 and 0.38.

CHILDHOOD EFFECT BY RATER
0.11standardized effect
Parent ratings
0.10standardized effect
Teacher ratings
Within-child link between rising mood swings and comfort eating two years later, from mood measured at ages six to ten. Source: Parisi et al., Child Development (2026)

For mood measured at ages 12, 14 and 16, the same path was not statistically significant on either set of ratings. That is an association the model failed to detect, which is different from proof that none exists, and the authors say they could not formally test whether childhood and adolescence differ. Up to age 12, parents reported their children’s comfort eating on one questionnaire; from 14, the teenagers reported their own on a different one. The paper is careful about that. The headline was not. Nothing in the analysis tracks a volatile seven-year-old into an overeating fifteen-year-old, and its own abstract confines the detected effect to childhood.


The study set out to test Emotional Eating Theory, which holds that children eat for comfort because they feel too much and cannot manage it, and the authors write that their results only partly support it. The first half held in childhood. The second half did not show up inside each child: a change in a child’s regulation ratings, from parents or teachers, did not predict later comfort eating at any age. Across children there was a familiar pattern, with parent-rated better regulators eating less for comfort on average in childhood, but within each child the model found nothing. In these data, how hard the storm hit predicted more than how well the child weathered it.

The finding the authors call surprising ran the other direction. Teenagers who reported more comfort eating at 14 and 16 were later rated by their teachers as better at regulating their emotions, an average effect of 0.10. The model cannot say why. One possibility, and it is only a possibility, is that the food does what the teenager wants it to: a snack settles the moment, and the composure that follows is what the teacher sees.

The funding lines name two public bodies, the Norwegian Research Council and a joint committee of the regional health authority Helse Midt-Norge and NTNU, and nothing from the food or drug industries. The limits the authors list are the ordinary ones for this design. Two years between assessments can miss anything that happens over weeks. Every measure was a questionnaire, with no observed eating. The model cannot account for factors that change over time, such as a parent’s own stress, and the sample is a European community cohort, weighted to correct for deliberate oversampling.

Comfort food starts at the kitchen counter

The researchers’ advice costs nothing. Oda Bjørklund, an associate professor who supervised the work, said that “paying extra attention to children who express strong and unpredictable emotions could be a wise tactic.” Silje Steinsbekk pointed to earlier work from the same cohort. In an 801-child analysis released in 2017, parents who used food to comfort their children at ages four and six had children who ate more emotionally at eight and ten, and children who were easily soothed by food were more likely to be fed for comfort later. Children who got upset or angry easily at four were at higher risk of both. “Food may work to calm a child,” Steinsbekk said then, “but the downside is teaching children to rely on food to deal with negative emotions.” Her advice now is pointed: “This habit may be particularly important to avoid for parents who have a highly emotional child.”

Neither study tested an intervention. Read together, they suggest where one might be aimed: the early school years, when a volatile child, a tired parent and a snack cupboard keep meeting. If that reading holds, the useful change happens at the kitchen counter, and nobody needs a prescription for it.

The adolescent finding stays unexplained, and it was recorded by the teachers. The ones in Trondheim who scored those sixteen- and eighteen-year-olds as calmer were rating what can be seen from the front of a classroom. The Emotion Regulation Checklist asks how a student handles frustration, and it has no line for what the student ate first.

Sources

  1. Parisi et al., Child Development (2026)
  2. News-Medical: “Children with intense mood swings face greater risk of emotional overeating in adolescence” (October 30, 2026)
  3. EurekAlert: “Parents’ use of emotional feeding increases emotional eating in school-age children” (April 25, 2017)