Sleep Troubles and Children’s Well-Being: Anxiety Carries the Link, Resilience Cuts Both Ways

Children in China who report more sleep problems also report noticeably lower well-being, and a new cross-sectional study suggests that anxiety may be the psychological conduit through which poor sleep and diminished happiness are connected. The research, published in Frontiers in Psychology, surveyed 606 children aged 9 to 12 years from two public primary schools in Guizhou Province and found a statistically significant negative correlation between sleep problems and well-being, with a Pearson coefficient of r = −0.39 (p < 0.001). Beyond this headline association, the study tested a more intricate question: whether anxiety explains part of the link, and whether resilience changes its strength. The answer to both, according to the authors, appears to be yes — but the way resilience behaves in the data is more surprising than a simple protective-buffer story would predict.

The research team, led by Junfeng Ding of Suzhou Polytechnic University together with Xinhua Hou and Guiping Dou, framed their investigation around the Double ABCX model, a long-standing framework from family stress theory. In this model, a stressor — here, sleep problems — is appraised cognitively and emotionally by the individual, producing a response such as anxiety, while personal resources like resilience shape how the stress-coping process unfolds. This theoretical scaffolding led to three hypotheses: that sleep problems would correlate negatively with well-being, that anxiety would mediate the association, and that resilience would moderate the indirect pathway. Data were collected between September and October 2024, with ethical approval from Zhoukou Normal University and informed consent obtained from the children’s legal guardians.

Participants were recruited through cluster sampling, initially comprising 667 children, of whom 61 (9.14 percent) were excluded during data screening for incomplete responses. The final analytical sample of 606 children had a mean age of 11.09 years (SD = 1.00) and was 53.3 percent girls. Each child completed four self-report instruments: the Athens Insomnia Scale, an eight-item measure of sleep difficulties originally developed for adults but validated for use in Chinese adolescents; the short eight-item Oxford Happiness Questionnaire, scored on a six-point scale; the seven-item anxiety subscale of the Depression Anxiety Stress Scale (DASS-21), assessing symptom frequency over the past week; and the 25-item Connor-Davidson Resilience Scale, which measured the capacity to maintain or restore psychological equilibrium under adversity. Internal consistencies were satisfactory for most measures, with Cronbach’s alpha values of 0.79 for the insomnia scale, 0.73 for the anxiety subscale, and 0.92 for the resilience scale, though the happiness questionnaire showed a relatively low alpha of 0.66.

Before testing hypotheses, the researchers took steps to guard against methodological artifacts. Harman’s single-factor test was used to assess common method variance, and an exploratory factor analysis found that the first unrotated principal component accounted for only 15.84 percent of total variance, well below the conventional 40 percent threshold. An independent samples t-test showed no significant gender difference in well-being, t(604) = 0.90, p = 0.370, so gender was excluded from further analysis. Age, which correlated significantly with well-being (r = 0.13, p = 0.001) and anxiety (r = −0.11, p = 0.007), was retained as a covariate. All continuous predictors were standardized via z-score transformation, so the reported coefficients represent the change in the outcome associated with a one-standard-deviation increase in the predictor.

The correlation matrix told a coherent story. Sleep problems correlated positively with anxiety (r = 0.43, p < 0.001) and negatively with both resilience (r = −0.35, p < 0.001) and well-being (r = −0.39, p < 0.001). Well-being was positively associated with resilience (r = 0.42, p < 0.001) and negatively with anxiety (r = −0.41, p < 0.001), while anxiety correlated negatively with resilience (r = −0.30, p < 0.001). These bivariate patterns set the stage for the central analysis: a moderated mediation model estimated with Hayes’ PROCESS macro, Model 58, using 5,000 bias-corrected bootstrap samples to generate 95 percent confidence intervals for direct and indirect effects.

The regression results supported the mediating role of anxiety. In the first equation, sleep problems strongly predicted anxiety (β = 0.46, 95 percent CI [0.37, 0.55], p < 0.001), and in the second, anxiety in turn predicted lower well-being (β = −0.20, 95 percent CI [−0.26, −0.14], p < 0.001), alongside a direct negative association between sleep problems and well-being (β = −0.16, 95 percent CI [−0.24, −0.09], p < 0.001). Resilience itself was a positive predictor of well-being (β = 0.26, 95 percent CI [0.20, 0.33], p < 0.001). Crucially, two interaction terms were significant: sleep problems by resilience predicted anxiety (β = 0.13, p = 0.001), and anxiety by resilience predicted well-being (β = −0.05, p = 0.029), indicating that the strength of the indirect pathway varied with children’s resilience levels.

Here the findings took an unexpected turn. When resilience was split into high (mean plus one standard deviation) and low (mean minus one standard deviation) groups, the positive association between sleep problems and anxiety was stronger among highly resilient children (β = 0.60, 95 percent CI [0.46, 0.73]) than among their less resilient peers (β = 0.34, 95 percent CI [0.24, 0.44]). Likewise, the negative association between anxiety and well-being was steeper in the high-resilience group (β = −0.25, 95 percent CI [−0.33, −0.17]) than in the low-resilience group (β = −0.15, 95 percent CI [−0.23, −0.08]). In other words, resilience did not buffer the pathway — it appeared to amplify it, a pattern the authors connect to the so-called dark side of resilience, in which higher resilience does not always confer protection and may sometimes intensify distress.

The overall indirect effect of sleep problems on well-being through anxiety was substantially larger for children with high resilience (indirect effect = −0.15, 95 percent CI [−0.23, −0.08], accounting for 48.4 percent of the total effect) than for those with low resilience (indirect effect = −0.05, 95 percent CI [−0.09, −0.03], accounting for 23.8 percent). The authors offer a tentative, explicitly speculative interpretation: highly resilient children may be more aware of or responsive to sleep-related difficulties, or may expend considerable effort trying to manage persistent problems, and this overinvestment could deplete limited coping resources, strengthening the anxiety-well-being association. They stress that the cross-sectional design makes such mechanistic explanations untestable in the present data.

Why does any of this matter? The authors point to epidemiological estimates suggesting that roughly 64 percent of Chinese children experience sleep problems, a figure that, combined with the established link between sleep quality and psychological functioning, underscores the scale of the issue. Prior research has tied childhood sleep difficulties to poor physical health outcomes such as obesity, impaired cognitive functioning, internalizing behavioral problems, and compromised social-emotional functioning. Within this landscape, the new study contributes something relatively rare: evidence from a Chinese pediatric sample about the potential psychological processes — anxiety as an appraisal of the sleep stressor, resilience as a modulating resource — that may underlie the sleep-well-being connection, rather than merely documenting the association itself.

The authors are careful about the study’s limits. The cross-sectional, non-experimental design precludes causal inference; all variables came from child self-report, with no objective or caregiver-reported sleep measures; potential confounders including socioeconomic status, health conditions, screen time, family functioning, and school characteristics were not assessed; and the low internal consistency of the happiness questionnaire may have attenuated some associations. The sample, drawn from two public primary schools in a single province, also limits generalizability. The practical suggestion that future interventions target both sleep quality and psychological factors such as anxiety and resilience remains, as the authors emphasize, a hypothesis for longitudinal and experimental research to test — not a recommendation that can yet be declared effective for improving children’s mental health.

Subject of Research: The association between sleep problems and well-being in Chinese children, mediated by anxiety and moderated by resilience

Article Title: The relationship between sleep problems and well-being among Chinese children: a moderated mediation model with anxiety as mediator and resilience as moderator

Article References: Ding, J., Hou, X., & Dou, G. (2026). The relationship between sleep problems and well-being among Chinese children: a moderated mediation model with anxiety as mediator and resilience as moderator. Frontiers in Psychology, 17, Article 1900442. https://doi.org/10.3389/fpsyg.2026.1900442

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1900442

Keywords: sleep problems, well-being, anxiety, resilience, Chinese children, moderated mediation, cross-sectional study, child mental health, Double ABCX model, Frontiers in Psychology, insomnia, bootstrap analysis