Researchers followed 72 children and youth in Canada, from kindergarten to Grade 12, aged about 5 to 17. A parent reported on each child in Fall 2022 and again in Spring 2023. Families were recruited online, partly through ADHD organizations. The paper does not say how many children had an ADHD diagnosis.
Children missed about two days of school in the two weeks before each survey. In the fall, the most common reason was COVID, reported for 62.5% of children. By spring, reluctance or avoidance led the list, reported for 61.1%.
Missing more school in the fall was linked to more acting-out behavior in the spring, such as attention problems, defiance and rule-breaking. It was also linked to more anxiety and depression symptoms later. These links were small to moderate. For acting-out behavior, the pattern also ran the other way: children with more of it in the fall missed more school in the spring. No such return path was found for anxiety and depression.
The sample was small and not chosen at random. With only two timepoints, the study cannot show what caused what. Both models fit the data only borderline well, so the authors read the results with caution.
The authors suggest that adults who work closely with children like these track how often they miss school, whatever the reason, and watch children who act out so they can step in early.
The numbers Stats and what the terms mean
The statistics behind the plain-language claims above.
- Krause, 2025: average days of school missed in the past two weeks, 2.42 days (fall) and 1.95 days (spring).
- Krause, 2025: fall absences predicting spring acting-out behavior, beta = 1.38, p = .02.
- Krause, 2025: fall acting-out behavior predicting spring absences, beta = .04, p = .03.
- Krause, 2025: fall absences predicting spring anxiety and depression symptoms, beta = 1.32, p = .003.
- Krause, 2025: how well the acting-out model fit the data, CFI .88, RMSEA .08.
- Krause, 2025: how well the anxiety and depression model fit the data, CFI .88, RMSEA .098.
What the terms mean:
- beta How much one measure predicted the other in the model, after taking age and gender into account.
- p How likely a result this strong would be if there were no real link; below .05 is usually read as unlikely to be down to chance.
- CFI A score of how well a model matches the data, where values closer to 1 mean a better fit.
- RMSEA A score of how far a model strays from the data, where lower values mean a better fit.
Source: Amanda Krause, Maria Rogers, Yuanyuan Jiang, et al.. A longitudinal investigation of school absenteeism and mental health challenges among Canadian children and youth in the COVID-19 context. Frontiers in Child and Adolescent Psychiatry (2025). doi:10.3389/frcha.2025.1604431