Salivary Cortisol, Socioemotional Functioning, and Academic Performance in Anxious and Non-Anxious Children of Elementary and Middle School Age

Abstract

Research Findings: Individual and contextual variables were examined in relation to children's ability to cope with socioemotional and academic challenges in a sample of typically developing (n = 51) and anxious (n = 72) children of elementary and middle school age. Anxious children had greater social difficulties than controls and showed different directions of cortisol response to a public speaking task, but there were no group differences in academic performance. Across the sample, greater salivary cortisol reactivity to the speech task was associated with both increased social difficulties and higher academic achievement. For a subset of younger (M = 6.9 years) but not older (M = 12.3 years) children, social functioning was also highly sensitive to contextual variables (e.g., family income level). Practice or Policy: Findings suggest that social functioning in elementary school is as sensitive to contextual variables as it is to physiological ones. Interventions promoting stress recovery, social interaction, and social competence may be beneficial during the elementary school years.

ACKNOWLEDGMENTS

This research was supported by a Lawson Foundation Post-doctoral Fellowship awarded to Karen J. Mathewson, a Vanier doctoral fellowship from the Natural Sciences and Engineering Research Council of Canada (NSERC) awarded to Vladimir Miskovic under the direction of Louis A. Schmidt, and operating grants from the Ontario Mental Health Foundation Grant awarded to Charles E. Cunningham and from the NSERC and the Social Sciences and Humanities Research Council of Canada awarded to Louis A. Schmidt. We thank the participants, their families, and the teachers for their cooperation and Lindsay Bennett, Diana Carbone, Sophia Fanourgiakis, Sue McKee, Renee Nossal, Matilda Nowakowski, and Jamie Sawyer and for their help with data collection and data entry.

Notes

Note. All values are uncorrected for covariates. MA = mixed anxiety; SM = selective mutism; S1 = Sample 1; S2 = Sample 2; S3 = Sample 3; BCFPI = Brief Child and Family Phone Interview; PPVT-III = Peabody Picture Vocabulary Test–III; PIAT = Peabody Individual Achievement Test.

Note. The interaction term tests whether the relations differed for anxious versus typical children.

†p < .09. *p < .05. **p < .01. ***p < .001.

*p < .05. **p < .01. ***p < .001.

‡p < .08. *p < .05.

¹As the age range within the older group (8.25 years) was wider than that in the younger group (4.83 years), we regressed academic achievement on cortisol reactivity for older children again, adding child age to the covariates. In this analysis, both family income level (β = .31, p < .01) and age (β = .80, p < .001) predicted academic achievement in older children, whereas cortisol reactivity was nonsignificant (β = .15, p > .20). Socioemotional outcomes were analyzed similarly, but there were no significant predictors of Functional Impact, Social Skills, or Sadness (all ps > .13) in the older group. Unlike in younger children, in older children, family income was more closely aligned with academic achievement than with socioemotional outcomes.

**p < .01.

²It is possible that the relations among sadness, cortisol reactivity, and academic performance were simply due to age-related development in all three variables. However, we note that age and cortisol reactivity were less closely related, pr (62) = .39, p < .01 (controlling for time of day, family income, and gender) than were age and age-equivalent academic achievement scores, pr(68) = .91, p < .001. The strength of these correlations differed greatly when they were tested using the Fisher r-to-z transformation (z = 6.06, p < .001, two-tailed). By contrast, the correlations between age and cortisol reactivity versus age and Sadness, pr (62) = .27, p < .04, were very similar (z = .73, p > .45, two-tailed). Although we present cross-sectional rather than longitudinal data, the findings suggest that age-related change in cognitive capacity appeared to occur at a much faster rate than development in emotional or HPA regulation, suggesting that the positive relations among sadness, cortisol reactivity, and academic performance were not simply due to a common association with increasing age.

³Although the speech task did not elicit a robust cortisol response in any group, behavioral evidence from a subset of younger children aged 5–8 suggests that the speech was perceived as stressful (Fanourgiakis et al., 2011). Compared to typically developing children, children with SM directed their speech less consistently to either the camera or their parent (ps < .02) and produced less spontaneous speech, thereby eliciting more prompting from the parent (p < .01). Children with MA scored between these two groups on these measures.