The study’s central methodological feature lies in how emotional intelligence was measured. Rather than asking parents or teachers to rate a child’s emotional abilities, the researchers used the Emotional Intelligence Test of the Botín Foundation for Children, a performance-based instrument grounded in the ability model of emotional intelligence. Performance-based measures present children with actual emotional problems to solve, and responses can be scored as correct or incorrect, making them comparatively more objective than self-report instruments, which are vulnerable to social desirability bias. The test comprises three subtests: Catch the Emotion, which assesses the ability to identify basic emotions from photographs of children’s faces; In Search of the Lost Emotion, which asks children to identify the emotions experienced by a protagonist across a series of stories; and The Emotional Detective, which measures the ability to select the most appropriate emotion-regulation strategy for a given situation. Its hands-on, visual format, in which emotions and situations are represented through illustrations and vignettes, makes the instrument accessible to young children whose language skills are still developing.

Each child was assessed individually in a quiet, distraction-free room at their school. Alongside the emotional intelligence test, the researchers administered the Battelle Developmental Inventory, a norm-referenced instrument that evaluates developmental status from birth to seven years and eleven months. The analysis focused on three of its domains, cognitive, communication, and motor development, following recent evidence that these three abilities form a general developmental factor underlying several early childhood inventories. Items on the inventory are scored on a three-point scale, from zero, indicating the child does not demonstrate the skill, to two, indicating the child fully demonstrates it. Internal consistency in the present sample was adequate across all measures, with coefficients ranging from 0.70 to 0.89, although the reliability estimate for emotion understanding was the lowest of the emotional abilities.

The gender findings were limited in scope. Girls scored significantly higher than boys in emotion perception, with a small effect size of Cohen’s d equal to −0.24, while boys scored significantly higher in cognitive ability, with d equal to 0.27. No significant gender differences emerged for communication, motor skills, emotion understanding, or emotion management. The researchers note that this pattern only partially supports their first hypothesis, which predicted that girls would obtain higher emotional intelligence scores overall. Previous studies have reported that girls and women tend to show higher ability-based emotional intelligence during adolescence and adulthood, and better emotion-regulation behaviour among girls has been documented in earlier research. The absence of a difference in emotion management here may reflect a distinction between general emotion-regulation behaviour and performance-based assessments of knowledge about emotion-management strategies. Given the small effect observed, the authors caution that the gender difference in emotion perception should be interpreted cautiously.

Age told a more nuanced story. Chronological age was positively associated with every general developmental domain and every emotional intelligence ability, but the strength of those associations differed markedly. Age correlated with cognitive ability at r = 0.56, with communication at r = 0.51, and with motor skills at r = 0.50, effect sizes consistent with the large developmental gains typical of this age range. By comparison, the correlations between age and the emotional abilities were smaller: r = 0.22 for perceiving emotions, r = 0.14 for understanding emotions, and r = 0.16 for managing emotions. The researchers fitted both linear and quadratic regression models to determine whether the age-related pattern was better characterised by a straight line or by a curve. Adding the quadratic term increased explained variance by no more than 1.2 percentage points, supporting a predominantly linear pattern across the age range examined.

The multivariable analyses then revealed ability-specific patterns that the bivariate correlations had obscured. In hierarchical stepwise regression models that entered age and gender first, followed by the developmental domains, cognitive ability independently predicted emotion perception, with β = 0.177 and p = 0.004, and predicted emotion management most strongly, with β = 0.322 and p < 0.001. Communication ability, by contrast, independently predicted emotion understanding, with β = 0.181 and p = 0.002. Age retained a small independent association only with emotion perception, while motor ability made no unique contribution to any of the final models. The authors interpret this differentiation through the lens of the ability model: interpreting emotional information may draw partly on language and communication resources, whereas identifying emotions and selecting appropriate emotion-management strategies may place greater demands on broader cognitive processes.

The final models explained 8.1 percent of the variance in emotion perception, 4.4 percent in emotion understanding, and 9.6 percent in emotion management. Those figures are modest, and the researchers are candid about why. Emotional intelligence in early childhood is complex and multifactorial, and unmeasured family, social, educational, and temperamental factors inevitably play key roles. Future research, they suggest, should integrate contextual variables such as parenting practices, parent-child emotional communication, parental emotional competencies, and broader social and educational environments. The relatively small correlations among the three emotional abilities themselves, ranging from r = 0.12 to r = 0.15, further suggest that perceiving, understanding, and managing emotions are not interchangeable during early childhood, reinforcing the conceptualisation of emotional intelligence as a set of related but distinct abilities.

The study addresses a gap in the literature. Research conducted since the 1980s has established that emotional understanding and regulation play a key role in developing social competence during the preschool years, and that children who apply greater emotional knowledge in emotionally charged situations respond more prosocially, are perceived by their teachers as more socially competent, and are better liked by their peers. Higher emotional intelligence has also been consistently associated with better mental health, subjective well-being, academic performance, and empathy across the lifespan. Yet integrative research examining emotional development alongside other developmental domains during early childhood has remained scarce, largely because adequate performance-based instruments for assessing emotional intelligence as an ability at these ages were lacking until recently. Only one previous study, a cross-sectional investigation of adolescents aged 10 to 13, had examined this relationship using an objective, performance-based measure.

The researchers are equally clear about the limitations. The cross-sectional design precludes conclusions about intraindividual developmental change or the direction of the observed associations, and the restricted age range of 48 to 83 months may have limited the variability needed to detect more complex age-related patterns. The lower reliability estimate for emotion understanding may also weaken observed associations involving that dimension. Hierarchical stepwise regression is a data-driven approach, so the identified models should be replicated in independent samples using theoretically specified predictors. The authors frame their practical suggestions as promising avenues for future intervention research rather than established protocols, given the correlational nature of the current study.

Even so, the findings carry implications for early education. The stronger associations between cognitive functioning and the abilities to perceive and manage emotions suggest that emotion recognition activities could require children to direct their attention and discriminate between relevant facial or contextual cues, while tasks could present everyday emotional situations where children consider different emotion-management strategies, anticipate their potential consequences, and select an appropriate response. The association between communication ability and emotion understanding points toward language-rich, interactive activities as a promising educational approach for fostering that ability, a finding the authors say warrants examination in future intervention studies. Taken together, the results indicate that performance-based emotional intelligence abilities form part of broader developmental processes while showing distinct associations with cognitive and communication abilities, and they highlight the potential value of integrating emotional learning with developmentally appropriate cognitive and communication activities in early educational settings to support children’s socio-emotional development during these formative early years.

Subject of Research: Associations between performance-based emotional intelligence and general developmental abilities in children aged four to six

Article Title: The development of emotional intelligence in childhood: associations with general developmental abilities

Article References: Cabello, R., Pérez-López, M., Llamas-Díaz, D., Gutiérrez-Cobo, M. J., Ruiz-Baeza, V. M., & Fernández-Berrocal, P. (2026). The development of emotional intelligence in childhood: associations with general developmental abilities. Frontiers in Psychology, 17, Article 1999245. https://doi.org/10.3389/fpsyg.2026.1999245

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1999245

Keywords: emotional intelligence, childhood development, preschoolers, cognitive ability, communication skills, motor skills, emotion perception, emotion understanding, emotion management, Battelle Developmental Inventory, cross-sectional study, ability model