Youth and Caregivers Rarely Agree on Who Stays Highly Irritable Through Adolescence

Anger-related irritability is one of the most common reasons children and adolescents are referred to mental health services, yet the field has long depended on a single perspective to track it: the caregiver’s. A new longitudinal analysis of the Adolescent Brain Cognitive Development (ABCD) Study, published in Child Psychiatry & Human Development, asked what happens when young people are allowed to describe their own tempers across the transition to adolescence. The answer is striking. When researchers compared annual reports from more than 9,000 youths and their caregivers between roughly ages 10 and 13, the two sides rarely agreed on which children were persistently and intensely irritable.

The study, led by Kristin Romvig Overgaard of the University of Oslo together with colleagues including Esten Leonardsen, drew on the ABCD Study’s annual data release 5.1. The researchers started with 11,875 participants, whose average age was 9.9 years at the first time point and 47.7 percent of whom were female. Because the analysis required complete irritability data from both informants at three annual assessments, the final analytic sample included 9,411 participants with complete caregiver and youth data, and after excluding the 9.9 percent whose profiles changed across waves in ways the classification scheme could not accommodate, 8,579 participants remained. The three time points captured approximately ages 10, 11, and 12, spanning the age range of 9 to 13 years.

Irritability was measured with a deliberately narrow instrument. Youths completed the Brief Problem Monitor-Youth, which contains a single irritability-related item, “hot temper,” defined as “bad temper, gets mad easily.” Caregivers completed the Child Behavior Checklist, and the researchers used its matching “temper tantrums or hot temper” item to ensure conceptual equivalence between informants. Both items are rated on a three-point scale from 0 to 2, with higher scores indicating greater irritability. The authors chose this single item because it is strongly associated with, and serves as a central indicator of, the irritability factor within the Achenbach System of Empirically Based Assessment, making it a meaningful proxy for phasic, anger-related irritability—the rapid-onset, overt kind, such as temper outbursts—as opposed to tonic irritability, the longer-lasting grouchy mood between outbursts.

Rather than letting a statistical algorithm carve the sample into trajectory groups, the researchers pre-specified a priori categories based on raw scores across the three waves. The resulting classifications were low-stable, low-to-moderate, moderate, and high-persistent irritability. The high-persistent group, the study’s central focus, was defined as endorsing a score of 2—”very true”—on at least two of the three annual waves according to either informant. Participants whose profiles changed across time points were set aside because the analyses did not account for the timing or sequence of changes, such as whether irritability rose early or late in the window.

The headline numbers reveal a substantial informant gap. Youths placed 5.1 percent of themselves (95% CI 4.6–5.6) in the high-persistent category, while caregivers assigned only 1.6 percent (95% CI 1.4–1.9) of the same children to it—roughly a threefold difference. The low-stable class dominated both perspectives, covering 69.7 percent of youths’ self-reports and 83.4 percent of caregiver reports. Low-to-moderate irritability accounted for 13.5 percent of youth reports versus 8.1 percent of caregiver reports, and the moderate class appeared in 11.8 percent of self-reports against 6.9 percent of caregiver reports. In every category above low, youths rated themselves more irritable than their caregivers did.

Sex patterns diverged sharply depending on who was doing the reporting. Based on youth self-reports, the proportion in the high-persistent group was virtually identical for females (4.9 percent, 95% CI 4.3–5.6) and males (5.2 percent, 95% CI 4.6–5.9). Caregiver reports told a different story: only 1.2 percent of females (95% CI 0.9–1.6) versus 2.0 percent of males (95% CI 1.6–2.4) were classified as high-persistent, a statistically significant difference (p = 0.01). This male-biased caregiver pattern echoes earlier trajectory studies that relied exclusively on caregiver reports, which found males over-represented in high-persistent or declining trajectories. The authors had hypothesized that females would self-report more irritability than their caregivers, which held true (4.9 percent vs. 1.2 percent, p < 0.001), but they also expected males to self-report less than their caregivers—a prediction the data contradicted, since males reported high irritability more often than their caregivers did (5.2 percent vs. 2.0 percent).

Agreement between informants looked deceptively good until chance was taken into account. Raw percentage agreement across the four classifications was high, at 65.3 percent overall (95% CI 64.3–66.3), and barely differed by sex. The prevalence-adjusted bias-adjusted kappa, which corrects for the fact that most children fall in the low-stable class, reached 0.88 overall. But when the researchers applied weighted Cohen’s kappa with squared ordinal weights—a statistic that penalizes larger disagreements more heavily and accounts for chance agreement—the overall value dropped to a modest 0.22 (SE = 0.01). Sex-specific values showed slightly better concordance for males (κ = 0.25, SE = 0.02) than for females (κ = 0.18, SE = 0.02), indicating that informant disagreement is not uniform across sexes. The gap between the high raw agreement and the low weighted kappa reflects the dominance of the low-stable classification; agreement on the clinically relevant moderate and high categories was weak.

The most consequential finding concerns what happened when one informant flagged persistent high irritability. When a caregiver reported a high-persistent classification, the youth’s own reports concurred in only 21.4 percent of those cases (95% CI 15.4–28.9). When the youth alone reported being persistently hot-tempered, the caregiver agreed in a mere 6.9 percent of cases (95% CI 4.9–9.7). Both informants designated the same child as high-persistent in just 0.3 percent of the sample (95% CI 0.2–0.5). Sex shaped even this slender overlap: among youths who self-reported high-persistent irritability, caregivers of males agreed more often than caregivers of females—9.5 percent versus 3.9 percent (OR = 2.55, 95% CI 1.11–5.87, p = 0.02). Among caregivers who reported high-persistent irritability, agreement did not differ significantly by sex (24.7 percent for males vs. 15.7 percent for females, OR = 1.75, 95% CI 0.72–3.42, p = 0.28).

Why would children and the adults who raise them see the same temper so differently? The authors point to developmental and measurement explanations. The transition to adolescence brings increasing autonomy, gradual detachment from caregivers, and the growing importance of peer relationships, processes that may reduce how much emotional information adolescents share at home. The single “hot temper” item may also be interpreted differently: youths may focus on internal feelings of anger while caregivers notice observable outbursts, and phasic irritability often flares in contexts caregivers rarely witness, such as peer interactions or school. Reporter-specific biases likely contribute as well—youths are aware of internal states invisible to others, while caregivers are attuned to overt disruptive behavior. The authors note that such discrepancies are not unique to irritability and have been linked to how similar caregivers perceive their children to be to themselves. Notably, the youth-reported prevalence of 5.1 percent sits within the 2 to 13 percent range that earlier caregiver-based trajectory studies found for high-persistent irritability, and it approximates community prevalence estimates for chronic irritability conditions, which range from 3.3 percent under strict DSM-5 criteria to 5.7 percent with relaxed exclusion criteria.

The authors are careful about what these numbers do and do not mean. The study is population-based, drawn from a cohort that may not reflect the complexity or severity of clinical cases, and excluded participants were more often non-white and from lower-income households, introducing a slight demographic bias, although baseline irritability levels did not differ meaningfully between included and excluded participants. The single-item measure captures only one facet of a multifaceted construct and cannot speak to tonic irritability, which prior cross-sectional work suggests females endorse more often than males. Because the high-persistent group is small, base-rate differences constrain the overlap between reports and reduce the precision of agreement estimates. The a priori classification scheme, while robust across sensitivity analyses that redefined the intermediate categories, does not rule out that alternative cut-offs would produce different subgroups. The authors also stress that the informant divergence documented here is not a direct estimate of clinical recognition. Still, the practical implication stands: relying on caregiver reports alone may miss critical aspects of the longitudinal irritability picture—particularly for females—and incorporating youth self-reports deserves consideration in both research and clinical assessment.

Subject of Research: Longitudinal agreement between caregiver and youth reports of anger-related irritability during the transition to adolescence

Article Title: Longitudinal Patterns of Irritability During the Transition to Adolescence: Examining Caregiver and Youth perspectives and Sex Differences

Article References: Overgaard, K. R., Brænden, A., Dahl, A., Westlye, L. T., & Leonardsen, E. (2026). Longitudinal Patterns of Irritability During the Transition to Adolescence: Examining Caregiver and Youth perspectives and Sex Differences. Child Psychiatry & Human Development. https://doi.org/10.1007/s10578-026-02116-1

Image Credits: AI Generated

DOI: 10.1007/s10578-026-02116-1

Keywords: irritability, adolescence, ABCD Study, caregiver reports, youth self-reports, sex differences, longitudinal, hot temper, informant discrepancy, child mental health, phasic irritability, Cohen's kappa