Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jan 30.
Published in final edited form as: Sleep Med Rev. 2025 Nov 22;85:102206. doi: 10.1016/j.smrv.2025.102206

Unpacking Sleep and Mental Health Disparities Across Childhood and Adolescence: A Meta-Analytic and Systematic Review

Morgan J Thompson 1, Alexandra D Ehrhardt 1, Ekjyot K Saini 2, Emily F Brigham 1, Zeynep Su Altinoz 1, Tiffany Yip 3, Joseph A Buckhalt 1, Mona El-Sheikh 1
PMCID: PMC12853373  NIHMSID: NIHMS2126285  PMID: 41313963

Summary

Children and adolescents from racial and ethnic minority and socioeconomically disadvantaged backgrounds are at greater risk for health disparities. This meta-analysis examined the association between sleep and mental health, with attention to variations by: (a) race, ethnicity, and socioeconomic status, and (b) individual sleep parameters and mental health domains. The review included 104 studies (66 independent samples, 326,478 participants). Less optimal sleep was associated with worse mental health (r = −.20, 95% CI [−.23, −.16]). Subjective sleep quality was a stronger correlate of mental health relative to all other sleep parameters (bs ranged from .11 to .19, ps < .001). Sleep schedule was a stronger correlate than objective sleep duration and quality (bs = .08 and .07, respectively, ps < .01). Sleep consistency was also a stronger correlate than objective sleep duration (b = .05, p = .03). Socioeconomic status moderated two associations. Objective sleep quality was more strongly linked to mental health among middle- relative to mixed-SES samples (b = .11, p = .01). Sleep consistency was more strongly associated with mental health among low- compared to mixed-SES samples (b = .05, p = .002). Findings are discussed in the context of methodological challenges in studying health disparities.

Keywords: Sleep duration, Sleep quality, Sleep schedule, Chronobiology, Sleep consistency, Sleep health, Mental health, Health disparities, Children and adolescents, Meta-analysis

Sleep problems are pervasive among youth and are consistently linked to poor mental health [1–3]. Evidence suggests that socioeconomically disadvantaged and minoritized youth may be at greater risk for less optimal sleep [4,5] and mental health symptoms [6]. Prior work highlights the key role of sleep in contributing to racial, ethnic, and SES disparities among youth [7,8]. Although research on the role of sleep in mental health within this context is emerging, a synthesis is needed to guide conceptual and methodological decisions. Meta-analyses have examined associations between race and ethnicity or SES and sleep [4,5], as well as sleep and mental health (i.e., internalizing and externalizing symptoms) [3]. However, previous meta-analyses have not quantified the interplay among the sociocultural milieu (i.e., race, ethnicity, SES), sleep, and mental health. The primary aim of this meta-analysis was to quantify the association between youth sleep and mental health and examine the moderating effects of race and ethnicity and SES. Additionally, expanding on existing meta-analyses, we evaluated the specificity of sleep parameters and mental health domains to answer two questions: (a) Is a particular sleep parameter more strongly associated with mental health relative to others? and (b) Are sleep parameters more strongly associated with externalizing versus internalizing symptoms?

Key Constructs

Mental health problems among youth are commonly categorized as internalizing (e.g., anxiety) and externalizing (e.g., disruptive behavior) symptoms (“table s1” provides operational definitions) [9]. Over the past decade, rates of mental health symptoms and diagnoses among US youth have generally increased [10–12], with internalizing symptoms accounting for a growing proportion of cases [11]. This rising prevalence presents significant public health concerns, affecting not only youths’ quality of life but also placing considerable strain on the healthcare system [12]. Continued efforts to understand risk factors and identify youth most vulnerable are critical.

Sleep is a multifaceted construct and examining multiple parameters can provide a more complete understanding of developmental and clinical outcomes [13–17]. Well-established sleep parameters [18] including duration, quality, schedule/chronobiology (i.e., timing), and consistency (variability/regularity in sleep) have documented associations with mental health [15–18]. We define more optimal sleep as longer duration, better quality, earlier schedules and greater preferences for morningness, as well as more consistent sleep-wake patterns. Sleep can be measured subjectively (e.g., surveys, sleep diaries) and objectively (e.g., actigraphy, polysomnography), each method having its own strengths and limitations [18]. We included objective and subjective assessments in the current review to test for differential associations with mental health; such assessments are scarce in existing meta-analyses. Our review focused on six sleep parameters: objective duration, subjective duration, objective quality, subjective quality, schedule/chronobiology, and consistency (“table s1”).

SES is a multidimensional construct, including objective (e.g., income, education, occupation, housing and neighborhood characteristics) and subjective (e.g., perceived economic well-being) components. Most research on sleep and SES focuses on income and education [8], but subjective assessments offer insights into how economic pressure and stress impact families [8,19,20]. Quantifying SES across studies is challenging due to a lack of standardization, regional variations, economic fluctuations over time, and differences in data granularity [21,22]. SES-related experiences and reporting can depend on sample characteristics such as demographics, and individual fluctuations in SES are often unaccounted for in multi-wave data. To address these challenges, we used a meta-coding procedure [23] to categorize studies into low, middle, high, or mixed SES samples based on all available SES information, allowing for standardization across studies while retaining more effect sizes.

Race and ethnicity are unique constructs—whereas race is a social construct that refers to perceived phenotypic differences, ethnicity reflects shared sociocultural characteristics shaped by social and historical contexts [24]. A closer examination of the existing literature on racial and ethnic disparities and sleep reveals several gaps. First, prior research has primarily focused on AA and EA samples [8], even though other groups (e.g., Asian Americans, Hispanic or Latinx) together represent over 25% of the US population [25]. Second, previous research has largely relied on between-group study designs that assume equal vulnerability within groups and ignore within-group heterogeneity [26,27]. Third, classifying individuals into a single racial or ethnic group is a growing challenge as the number of US individuals identifying as multi-racial or multi-ethnic increases [28]. Lastly, sleep research is multidisciplinary, and there is no standardized guideline for measuring and reporting race and ethnicity. Consequently, these data are often presented inconsistently, if at all. We use the term “race-ethnicity” hereafter to reflect the diverse and sometimes conflated ways in which these constructs are derived. We emphasize, however, that the use of the term “race-ethnicity” should not be interpreted as implying that the constructs are interchangeable.

Given the early stage of this research, we retained as many studies as possible and reported all racial-ethnic information available across between- and within-group designs. Small cell sizes limited our ability to test within-group meta-analytic effects. We employed a between-group analytic approach that examined variations in weighted effects based on the proportion of one racial-ethnic group relative to all others.

Race-Ethnicity and SES Related Disparities in Links Between Sleep and Mental Health

Structural and systemic problems, including discrimination and racism, may increase risk for mental health symptoms and are known correlates for less optimal sleep [29,30]. Little research has tested race-ethnicity as a moderator of sleep and mental health [8]. This work shows that racial-ethnic differences in mental health are smaller among youth experiencing better-quality sleep relative to worse quality sleep [31]. Relatedly, some studies have found weaker links between discrimination and mental health among youth experiencing more versus less optimal sleep [8,32,33]. Together, these findings show the potentially protective role of sleep in mitigating racial-ethnic disparities in mental health. Risk for mental health symptoms and less optimal sleep also arise as a function of SES disadvantage [34,35]. Few studies have examined SES as a moderator of sleep and mental health and tended to report smaller SES differences in mental health among youth with poor sleep [36,37]. Studies demonstrate the protective role SES may play in reducing links between less optimal sleep and mental health.

As the primary aim of the current meta-analysis, we examined race-ethnicity and SES as moderators of the association between sleep and mental health. Although conceptually different from work examining sleep as a moderator of relations between race-ethnicity or SES and mental health, examining race-ethnicity or SES as moderators of the sleep–mental health link is statistically equivalent. Despite limited research on this interplay, meta-analyses offer a unique tool to fill this gap. By extracting bivariate correlations between sleep and mental health and the available racial-ethnic and SES data, we probed variations in effect sizes between sleep and mental health based on samples’ racial-ethnic composition and SES—a question not previously synthesized, marking a key contribution of this meta-analysis.

Additional Moderators

To determine the robustness of the sleep and mental health link, we examined additional sources of heterogeneity, including participant and methodological characteristics. Participant characteristics included youth gender [17,38–40] and age [13,17,41]. Methodological characteristics tested as moderators included mental health informant (youth, parent, teacher, multi-informant, objective or clinician rated) [42,43], subjective sleep informant (sleep duration and quality only), shared-informant variance (same versus different reporters for subjective sleep duration and quality and mental health), subjective versus objective assessments of sleep duration and quality [18], cross-sectional versus longitudinal associations [44], and time lag (i.e., amount of time elapsed between measurements) [45].

Present Study

This meta-analysis makes significant contributions to the literature. First, the primary goal was to examine how the sleep and mental health link varied as a function of race-ethnicity and SES. Second, specificity analyses probed whether any given sleep parameter (objective duration, subjective duration, objective quality, subjective quality, schedule/chronobiology, and consistency) was more strongly linked to mental health relative to another parameter by examining sleep parameter as a moderator of the association between overall sleep (i.e., across all sleep parameters) and mental health. Additionally, to understand whether a particular sleep parameter is more strongly associated with a particular mental health domain, we examined the moderating role of externalizing versus internalizing symptoms. Third, because specific sleep parameters can offer additional insights on racial-ethnic and SES disparities, we examined race-ethnicity and SES as moderators at the level of each sleep parameter to offer a more comprehensive picture of sleep and health disparities. Likewise, to account for heterogeneity at the individual sleep parameter level, we assessed the robustness of the association between each sleep parameter and mental health across participant and methodological characteristics. These analyses provide insight into who is most vulnerable to the negative effects of poor sleep and identify areas for refining and developing measurement tools.

Method

The report follows PRISMA guidelines [46].

Search Strategy, Study Selection, and Coding Procedures

Literature searches for relevant records were initially conducted in August 2022 using PsycINFO, PubMed, Web of Science, CINAHL, and ProQuest. The search included all accessible journals, chapters, and dissertations and theses. “File s1” describes the search strategy. Briefly, key search terms reflected sleep parameters, mental health, and race-ethnicity and SES. “Figure s1” shows the flow of study reports into the meta-analysis.

After duplicate removal, title and abstract screening was completed by two research assistants trained to reliability by the first author. The two coders overlapped on 31% of records (κ = .64) [47,48]. The first and fourth authors overlapped on 100% of full-text screening (κ = .66). “File s1” provides a detailed explanation of inclusion criteria. Studies were included if they: 1) were quantitative, 2), were in English, 3) were published in 1990 or after, 4) concerned youth 5–17 years old, 5) had an N of ≥ 28, 6) concerned community samples, 7) were conducted in the US, 8) measured sleep concurrently or before mental health, 9) reported on independent samples or nonoverlapping effects, 10) provided at least one assessment of sleep, excluding those from the Child Behavior Checklist, 11) provided at least one assessment of mental health, 12) reported on sample SES or racial-ethnic makeup, and 13) reported necessary statistical information.

The first and fourth authors independently coded all records using a detailed coding system (“table s2”). Reliability was moderate to strong for continuous (ICCs = .93–1.00; MICC = .99) and categorical (κs = .69–1.00; Mκ = .96) variables. Coders discussed discrepancies and consensus codes, which were utilized in the final analyses [48].

Two updated literature searches using the same procedures as the initial search were conducted in September 2024 and March 2025. These updates identified studies published during manuscript preparation and expanded the search terms to include “sleep health” and “circadian misalignment”. “Figure s2” shows the flow of study reports from the updated searches into the meta-analysis. The second and fifth authors reviewed all titles and abstracts (first update: κ = .76; second update: κ = .88). The first and second authors independently reviewed all full texts (first update: κ = .84; second update: κ = .82).

The first and second authors coded articles included in the two updated searches using the same procedures as the initial round. Both authors independently coded 30% of the articles to assess reliability; the remainder were divided between them. Reliability was acceptable to strong for continuous (ICCs = .99–1.00, MICC = .99) and categorical (κs = .50–1.00, Mκ = .82) variables. Discrepancies were resolved through consensus and were used in the final analyses [48].

Risk of Bias

Methodological procedures vary in rigor. Variation in the weighted effect across diverse methods can indicate potential bias. Risk of bias was assessed across assessment method (subjective versus objective sleep), mental health and sleep informants, shared-informant, study design (cross-sectional, longitudinal), and time lag.

Bias in meta-analytic effects can arise from publication bias (e.g., locating, selecting, and combining studies) [49]. To assess or mitigate potential biases, several measures were implemented. We included both published and gray literature (e.g., dissertations and theses) in the search strategy and analyzed how weighted effects varied between published or unpublished sources. We conducted a modified Egger’s regression test [49] for multilevel meta-analysis (MLMA) [50,51]. Lastly, we evaluated potential reporting biases using a commonsensical approach of contour-enhanced funnel plots [52].

Data Analysis

Analyses were conducted in accordance with best practices in R using the metafor package. [53] We utilized Pearson’s r to capture the association between sleep and mental health. Interpretation of effect size magnitude followed established guidelines [54]. Most records reported Pearson’s r; however, we also considered Cohen’s d, partial eta squared (η2), and chi-squared (χ2) effect sizes, which were transformed into r. Effects were examined for directionality (with lower scores indicating worse sleep and better mental health)—a negative r indicated that less optimal sleep (lower scores) was associated with worse mental health (higher scores).

Our analysis plan tested three models reflecting the three aims of the study. Model 1 examined how the association between sleep and mental health varied across sleep parameters and mental health domains. Model 2 probed variations in the association between each sleep parameter and mental health based on race-ethnicity and SES. Model 3 assessed how the association between individual sleep parameters and mental health depended on participant and methodological characteristics. See “file s1” for model details.

Continuous moderators were examined when at least three studies reported relevant information. We provide the average weighted effect (r) when the moderator is zero (i.e., the intercept) and the slope coefficient (b) to express the degree of change for every 1-unit increase (−b) or decrease (+b) in the moderator. Although this appears counterintuitive, our effect sizes were recorded to reflect a negative correlation between sleep and mental health—less optimal sleep (lower values) is associated with worse mental health (higher values). Thus, a −b indicates a strengthening effect whereas a +b suggests a weakening effect. Categorical moderators were examined when each cell contained at least three studies. We provide the average weighted effect (r) of the reference category (i.e., the intercept) and the slope coefficient (b) to quantify differences from comparison categories. A −b indicates a stronger effect whereas a +b suggests a weaker effect in the comparison category relative to the reference category.

We utilized a hierarchical meta-analysis to capitalize on all possible effects. Hierarchical models account for potential bias by adding an intermediary level representing a cluster effect that organizes effect sizes based on the study from which they were drawn [55]. We assessed three levels: Level 1—sampling variance between effect sizes within a record; Level 2—variance between records within a sample; Level 3—between sample variance. We measured the degree of heterogeneity that could be attributed to variance at Level 1 relative to Level 2 and Level 3 as well as ICCs to determine the degree of homogeneity in effect sizes across clusters. ICCs ranged from small to moderate (“table s3”; note that ICCs for Models 2 and 3 paralleled Model 1.3).

Transparency and Openness

Data and coding materials from the meta-analysis are available online on the Open Science Framework (https://osf.io/ejhc6).

Results

We included 104 studies, consisting of 1,208 effects, 66 independent samples, and 326,478 participants (52% girls; 54% European American/White, 17% African American/Black, 6% Asian American, 11% other race-ethnicity, 17% Latinx/Hispanic; Mage at sleep assessment = 8.00 years old; Mage at mental health assessment = 8.46 years old). Samples represented relatively diverse socioeconomic backgrounds: 30% low SES, 23% middle SES, 7% high SES, and 17% mixed SES. More studies reported cross-sectional effects (97%) than longitudinal (19%) effects. Short-term longitudinal studies (i.e., daily diary) were categorized as cross-sectional because data were aggregated across the week and examined as between-group effects. The average lag between assessments was 11 months. Effects were predominantly from published sources (80%). “Table s2” provides descriptive statistics for study variables. “Table s4” summarizes study characteristics.

Heterogeneity

There was significant heterogeneity across Models 1.1–1.3 (“table s3”). Except for Models 1.3A (objective sleep duration), 1.3C (objective sleep quality), and 1.3E (sleep schedule/chronobiology), differences between effect sizes within studies accounted for the largest proportion of variance.

Model 1: Specificity Analyses

“Table s5” provides results from specificity analyses.

Model 1.1: Overall Sleep and Mental Health

Sleep was significantly and modestly associated with mental health, r = −.20, 95% CI [−.23, −.16], p < .001.

Model 1.1A: Variations Across Sleep Parameters.

Each sleep parameter showed a significant effect, ranging from weak to moderate (rs ranged from −.08 to −.27, ps < .01). Effect sizes significantly varied across sleep parameters (“figure s3”). Subjective sleep quality showed a stronger effect relative to objective sleep duration (b = .19, p < .001), subjective sleep duration (b = .14, p < .001), objective sleep quality (b = .19, p < .001), sleep schedule (b = .11, p < .001), and sleep consistency (b = .14, p < .001). The effect for sleep schedule was stronger than objective sleep duration (b = .08, p = .003) and objective sleep quality (b = .07, p = .005). The effect for sleep consistency was stronger than objective sleep duration (b = .05, p = .03).

Model 1.1B: Variations Across Mental Health Domains.

Overall sleep showed a significant and modest effect on each domain of mental health (rs = −.18 [externalizing], −.20 [internalizing], ps < .001). Effect sizes varied across mental health domains (“figure s3”), with internalizing showing a stronger effect relative to externalizing (b = .03, p = .02).

Model 1.2: Associations Between Overall Sleep and Each Domain of Mental Health

Moving toward greater specificity, we examined whether one sleep parameter was more strongly associated with externalizing (Model 1.2A) and internalizing (Model 1.2B) symptoms.

Model 1.2A: Externalizing Symptoms.

Overall sleep showed a significant, yet small effect on externalizing symptoms, r = −.12, 95% CI [−.15, −.09], p < .001. With the exception of objective sleep duration, each sleep parameter had a significant yet small to modest effect, (rs ranged from −.09 to −.17, ps < .001). The omnibus moderation analysis was significant (“figure s3”). Subjective sleep quality had a stronger effect relative to objective sleep duration (b = .12, p < .001), subjective sleep duration (b = .08, p = .001), objective sleep quality (b = .08, p = .01), sleep schedule (b = .06, p = .004), and sleep consistency (b = .07, p = .03).

Model 1.2B: Internalizing Symptoms.

The association between overall sleep and internalizing symptoms was modest and significant, r = −.22, 95% CI [−.26, −.18], p < .001. Each sleep parameter had a significant, weak to moderate effect (rs ranged from −.07 to −.32, ps < .02). The omnibus moderation analysis was significant (“figure s3”). The effect of subjective sleep quality was stronger than objective sleep duration (b = .23, p < .001), subjective sleep duration (b = .21, p < .001), objective sleep quality (b = .25, p < .001), sleep schedule (b = .19, p < .001), and sleep consistency (b = .20, p < .001).

Model 1.3: Associations Between Each Sleep Parameter and Mental Health

We examined whether each sleep parameter was more strongly related to externalizing versus internalizing symptoms.

Model 1.3A: Objective Sleep Duration.

Objective sleep duration had a weak, yet significant effect on mental health, r = −.05, 95% CI [−.08, −.01], p = .01. The effect did not vary across mental health domains and was weakly associated with both externalizing and internalizing symptoms (rs = −.05, ps ≤ .01).

Model 1.3B: Subjective Sleep Duration.

The effect of subjective sleep duration on mental health was weak, but significant, r = −.08, 95% CI [−.12, −.04], p < .001. This effect did not differ across mental health domains and was weakly associated with each (rs = −.09 [externalizing], −.08 [internalizing], ps < .001).

Model 1.3C: Objective Sleep Quality.

Objective sleep quality was unrelated to mental health, r = −.05, 95% CI [−.13, .02], p = .17, with no significant effects for either externalizing or internalizing symptoms, and no significant omnibus moderation.

Model 1.3D: Subjective Sleep Quality.

Subjective sleep quality showed a moderate, significant effect on mental health, r = −.28, 95% CI [−.33, −.23], p < .001. The omnibus moderation analysis was significant (“figure s3”). The effect was significant across both domains (rs = −.20 [externalizing], −.32 [internalizing], ps < .001) but was stronger for internalizing relative to externalizing (b = .11, p < .001).

Model 1.3E: Sleep Schedule/Chronobiology.

Sleep schedule had a small and significant effect on mental health, r = −.09, 95% CI [−.14, −.04], p < .001. The omnibus moderation analysis was significant (“figure s3”). Effects for each mental health domain were significant and weak to small (rs = −.07 [internalizing], −.11 [externalizing], ps < .02), but the effect for externalizing was stronger than internalizing (b = .04, p = .02).

Model 1.3F: Sleep Consistency.

Sleep consistency showed a small, but significant effect on mental health, r = −.09, 95% CI [−.12, −.06], p < .001. The omnibus moderation analysis was not significant—the effect was small and significant for both externalizing and internalizing symptoms (rs = −.09, ps < .001).

Model 2: Race-Ethnicity and SES as Moderators

Full results for race-ethnicity and SES moderation tests are provided in “table s6”.

Model 2.1: Variations Based on Sample Racial-Ethnic Composition

To test for variations based on sample racial-ethnic composition, the next set of models examined whether the effect between each sleep parameter and mental health was moderated by race-ethnicity. One marginal omnibus effect emerged: the association between subjective sleep duration and mental health weakened as the proportion of multi-racial, multi-ethnic, or other race-ethnicity youth increased, (b = .004, p = .08; “figure s4”).

Model 2.2: Variations Based on Sample SES

We also examined whether the effect between each sleep parameter and mental health was moderated by sample SES. Two significant moderation effects emerged (“figure s5”). The link between objective sleep quality and mental health was stronger in middle- relative to mixed-SES samples (b = .11, p = .01). The association between sleep consistency and mental health was stronger in low- relative to mixed-SES samples (b = .05, p = .002). Several marginal omnibus moderation effects also emerged, with the following significant pairwise comparisons (“figure s5”). The effect of objective sleep duration was weaker in low- than middle- (b = −.06, p = .04) and mixed-SES samples (b = −.07, p = .04). Subjective sleep duration showed a weaker effect in low- compared to mixed-SES samples (b = −.11, p = .03). Finally, the effect of subjective sleep quality was stronger among middle- relative to mixed-SES samples (b = .12, p = .01).

Supplemental Qualitative Interaction Patterns

A qualitative review on studies assessing interactions between SES and race-ethnicity was conducted to further contextualize findings. The complete qualitative review is provided in “file s1”. Briefly, findings were generally mixed, but greater support was found for SES and race-ethnicity as moderators of the association between sleep quality and mental health outcomes than sleep duration or sleep consistency; no studies examined moderation of sleep schedule/chronobiology. Significant interactions generally supported the role of poor sleep as a risk factor for worse mental health among historically marginalized youth.

Model 3: Sources of Heterogeneity Among Sleep Parameters and Mental Health

We examined moderation by participant and methodological characteristics (see “table s7” and “table s8” for full results). Significant moderation findings are detailed below. Participant and methodological characteristics did not moderate effects for objective sleep duration (Model 3A) or sleep consistency (Model 3F).

Model 3B: Subjective Sleep Duration

The weighted effect of subjective sleep duration was moderated by shared-informant variance, with stronger effects when the same informant reported on both sleep duration and mental health compared to different reporters (b = .06, p = .03; “figure s6”).

Model 3C: Objective Sleep Quality

The weighted effect of objective sleep quality was moderated by participant gender and outcome reporter. As the proportion of girls increased, the association with mental health strengthened (b = −.01, p = .002; “figure s7a”). The effect was also weaker when teachers versus youth (b = −.11, p = .01) and parents (b = −.15, p < .001) reported on mental health (“figure s6”).

Model 3D: Subjective Sleep Quality

The weighted effect of subjective sleep quality was moderated by time lag, study design, outcome informant, and shared-informant variance. As the time between measurement of subjective sleep quality and mental health increased, the weighted effect weakened (b = .03, p = .01; “figure s7b”). Likewise, the weighted effect was stronger for cross-sectional than longitudinal associations (b = .10, p = .002; “figure s6”). Additionally, the association was stronger when youth self-reported on mental health than when it was parent-reported (b = .15, p < .001), teacher-reported (b = .34, p < .001), objectively or clinician-rated (b = .25, p < .001), or based on multi-informant reports (b = .34, p = .03) (“figure s6”). Additionally, the weighted effect was stronger for parent than teacher reports (b = .19, p = .001). Finally, stronger effects emerged when the same informant reported on both sleep quality and mental health than different reporters (b = .24, p < .001; “figure s6”).

Model 3E: Sleep Schedule/Chronobiology

The weighted effect of sleep schedule/chronobiology was moderated by outcome reporter (“figure s6”): the association with mental health was weaker for objectively or clinician-rated mental health than self- (b = −.19, p = .03) or parent-reported mental health (b = −.23, p = .01).

Publication Bias

We assessed publication bias for each sleep parameter using three tests—record type moderation, MLMA Egger’s test, and contour-enhanced funnel plots. Overall, evidence of reporting bias was minimal. Full results are described in “file s1”.

Discussion

This meta-analysis is the first to examine: (a) race-ethnicity and SES as moderators of the link between sleep and mental health, and (b) specificity across sleep parameters and mental health domains. Broadly, less optimal sleep was associated with worse mental health, with variation by race-ethnicity, SES, sleep parameter, and mental health domain.

Sleep Disparities

The primary goal was to examine race-ethnicity and SES as moderators of the association between sleep and mental health. Prior reviews [4,5] highlight direct associations between sleep and race-ethnicity and SES, suggesting that sleep may play a pivotal role in the development of mental health disparities among youth [7,8]. In the current review, few moderating effects emerged for race-ethnicity and SES. Regarding race-ethnicity, one marginal effect suggested that the association between subjective sleep duration and mental health weakened as the proportion of multi-racial, multi-ethnic, or other race-ethnicity youth increased. Related to SES, objective sleep quality was associated with worse mental health in middle- but not mixed-SES samples. Sleep variability (i.e., lower consistency) was associated with worse mental health among low-SES samples relative to mixed-SES samples. Marginal findings also emerged for objective and subjective sleep duration, suggesting that effect sizes were significant for middle- or mixed-SES, but not low-SES samples. Additionally, a marginal finding showed that the effect of subjective sleep quality was stronger in middle-SES samples relative to mixed-SES samples.

Although the meta-analysis generally suggests that links between poor sleep and mental health are not stronger among historically marginalized than majority samples, this interpretation warrants caution. The limited and sometimes contradictory findings should not be taken as evidence that race-ethnicity and SES are irrelevant in the context of sleep and mental health [4,5]. For example, significant intercepts in categorical SES moderation analyses indicated that poor sleep was linked to worse mental health among socioeconomically disadvantaged youth (“table s6”). Likewise, our qualitative review suggests that effects of poor sleep on mental health are amplified among historically marginalized youth (“file s1”). Situating such findings within the context of broader structural barriers underscores the important role sleep may play in promoting mental health among historically marginalized youth, even when these effects are not fully captured in between-group comparisons.

The complexities of race-ethnicity and SES are difficult to synthesize across studies. For race-ethnicity, we synthesized the total proportion of multiple racial-ethnic groups across studies. Because minoritized groups are historically understudied and often reported as “other”, this approach may yield underpowered comparisons. By contrast, within-group designs offer a powerful tool for illuminating unique variability within specific groups [26]. Limited cell sizes precluded this approach in the current review. Regarding SES, we categorized samples as low-, middle-, high-, and mixed-SES by collapsing indicators of social prestige, financial resources, perceived stability, and community-level SES. These indicators capture distinct aspects of SES that may relate differently to sleep and mental health [56]. Small cell sizes prevented testing associations across SES levels within individual indicators. Moreover, SES varies across geographic regions, with its impact and associated experiences shaped by the local demographics and regional context [21,22].

The intertwined nature of race-ethnicity and SES in the US presents significant challenges in isolating their unique effects on sleep, as they contribute both shared and distinct variance. This complexity is often overlooked, potentially obscuring significant findings. Depending on how this covariance is handled, observed effects may not translate to meta-analytic models. For example, some meta-analyses use standardized beta coefficients to account for shared variance [57]; however, this approach presents certain challenges. First, variation in control variables across studies means that standardized beta coefficients do not consistently capture the same associations. Second, few studies have directly tested race-ethnicity and SES as moderators of the sleep-mental health association, limiting extractable betas and reducing statistical power.

Specificity in Associations Between Sleep and Mental Health

A second goal was to examine whether associations between sleep and mental health varied by sleep parameter and mental health domain. Subjective sleep quality consistently showed the strongest associations. Sleep schedule/chronobiology was stronger than objective sleep duration and quality, and sleep consistency was also stronger than objective sleep quality. These patterns largely replicated across externalizing and internalizing symptoms, with a few exceptions. First, objective sleep duration was unrelated to externalizing symptoms (Model 1.2A). Notably, when examining mental health domain as a moderator, a significant association between objective sleep duration and externalizing symptoms emerged (Model 1.3A), suggesting an inconsistent link that may vary depending on the broader sleep context. Further underscoring the potential influence of the broader sleep context on effects of individual sleep parameters, analyses testing mental health domain as a moderator of objective sleep quality revealed no significant associations with either outcome (Model 1.3C). Last, sleep schedule/chronobiology and sleep consistency were not stronger correlates of either externalizing (Model 1.2A) or internalizing (Model 1.2B) symptoms compared to objective sleep duration and objective sleep quality.

Extending prior meta-analyses linking sleep with mental health [1,3], our findings suggest that risk is not uniform across sleep parameters, underscoring the need to identify mechanisms underlying these differences. Subjective sleep quality consistently emerged as the strongest correlate, potentially due to greater study representation, shared-method or shared-informant bias artificially inflating effect sizes, and differences in what subjective assessments capture. Specifically, objective measures may offer more accurate estimates, but they do not capture subjective experiences including satisfaction and feeling rested. The definition of “good” sleep varies from person to person—individuals with similar objective sleep quality may still differ in how rested they feel [36]. These individual differences may explain divergent associations of objective and subjective sleep with mental health.

We also tested whether sleep parameters were more strongly associated with externalizing or internalizing symptoms. Overall, sleep was more strongly linked to internalizing symptoms. Objective sleep duration, subjective sleep duration, objective sleep quality, and sleep consistency showed similar associations across the two mental health domains. In contrast, sleep schedule/chronobiology was more strongly related to externalizing symptoms, whereas subjective sleep quality was more strongly linked to internalizing symptoms.

Subjective sleep problems are common symptoms of depression and anxiety [58,59]. Although some studies exclude sleep items when examining internalizing symptoms, this is not standard practice [60,61], which may inflate associations between subjective sleep quality and internalizing symptoms. In comparison, sleep schedule/chronobiology showed stronger associations with externalizing symptoms. Although links between sleep duration and quality and externalizing behaviors are well-established, in part due to impacts on cognitive functioning [62], our findings highlight the additional importance of sleep timing. Delayed sleep timing may be associated with reduced self-control, heightened sensation-seeking, and a lack of protective external factors (e.g., summer jobs that may promote earlier bedtimes), thereby increasing risk for externalizing symptoms [63]. Collectively, this meta-analysis highlights the value of treating sleep as a multi-faceted construct to better understand its relations with mental health.

Other Moderators and Risk of Bias

Weighted effects of objective sleep duration and sleep consistency were robust across participant and methodological characteristics. Generally, associations between sleep and mental health were robust across gender and age, with one exception: consistent with research indicating that girls may be more vulnerable to the negative effects of poor sleep [38,39,64,65], the effect of objective sleep quality strengthened as the proportion of girls increased.

Regarding methodological rigor, associations between sleep duration and mental health were consistent across objective and subjective assessments. Conversely, subjective sleep quality showed stronger effects than objective sleep quality and was more strongly associated with mental health when youth reported their symptoms. This may reflect shared-reporter and shared-method variance inflating effect sizes, greater accuracy in youth self-report, or age-related informant differences. Indeed, parent-report was more common in younger samples (sleep: Mage = 9.87 years; mental health: Mage = 12.41 years), whereas self-report predominated in older samples (sleep: Mage = 13.22 years; mental health: Mage = 14.20 years). Both objective and subjective sleep quality showed stronger effects when mental health was youth- and parent- versus teacher-reported, potentially reflecting contextual or situation-specific differences in behavior [66]. Finally, the effect of sleep schedule/chronobiology was stronger when mental health was assessed via self- or parent-report than when it was objectively or clinician-rated, again suggesting shared-informant and shared-method bias.

The association between subjective sleep quality and mental health was stronger in cross-sectional than longitudinal investigations, possibly indicating bias from random measurement error in concurrent assessments [44]. Likewise, this effect weakened with longer time lags, consistent with revisionist models suggesting that poor sleep quality influences long-term development indirectly and is superseded by more proximal experiences [45]. Time lag did not moderate effects for other sleep parameters, aligning with an enduring effects perspective suggesting stable, lasting consequences of poor sleep [45]. Minimal evidence of publication bias indicates that study findings were robust to study selection processes [49].

Limitations and Recommendations for Future Research

Contextualizing, Defining, and Measuring Sociocultural Concepts

We restricted our review to US samples to ensure consistency in sociocultural and political context. Social, economic, and cultural systems vary across countries shaping how race-ethnicity and SES are defined, measured, and experienced. Thus, their role as moderators in associations between sleep and mental health may not generalize globally. It would be extremely valuable for future research to examine these associations in international and multinational contexts to explore cross-cultural variation in how SES and race-ethnicity function as moderators.

We carefully selected our SES moderation approach to address challenges in quantifying SES across studies. Yet collapsing across indicators may have introduced information loss, as SES effects and social class experiences vary substantially by indicator (e.g., income vs. educational prestige) [21]. This highlights the inherent difficulty of meaningfully measuring and defining SES. Emerging recommendations suggest conceptualizing SES as an organizational construct composed of measurable structural features, akin to how race and sex are treated in demographic research [21]. SES should not be theorized monolithically; greater specificity is needed in articulating mechanisms and identifying which SES features (e.g., financial strain, educational disadvantage, occupational instability) confer risk. This precision enhances empirical utility and informs targeted intervention and prevention efforts.

Multiple studies were excluded due to missing SES or race-ethnicity data. Inconsistent sociodemographic reporting—compounded by the field’s multidisciplinary nature and lack of uniform reporting guidelines—remains a major limitation. Although a single reporting framework may be unrealistic, thorough sociodemographic reporting is essential for generating comprehensive insights. Relatedly, future research should not only improve representation of racial-ethnic groups but also adopt intersectional frameworks to better capture the complex interplay between race-ethnicity and SES, especially among individuals holding multiple marginalized identities [24,67].

Analytic Considerations

Few significant moderation effects emerged for race-ethnicity and SES. Meta-analytic frameworks emphasize between-sample variation, which may mask important within-sample (i.e., between-family) differences. Nevertheless, our findings provide a valuable starting point and highlight the critical need for more empirical tests of moderation within individual studies, particularly given that only 12 studies in our review tested these effects.

Included studies predominantly used cross-sectional and between-group designs. It is important to move beyond the inherently comparative nature of between-group designs, which emphasize group differences in terms of “risk” and “protection,” obscuring within-group variability. To capture how sleep–mental health associations and sociocultural influences shift developmentally [68], longitudinal, within-group designs are essential.

Relatedly, daily diary studies were treated as cross-sectional due to limited numbers and data aggregation processes. Though underused, daily diary methods offer valuable insights into day-to-day fluctuations in sleep, mental health, and experiences related to race-ethnicity and SES (e.g., discrimination) [30]. By capturing these processes in real time, they provide a powerful tool for understanding the daily mechanisms underlying long-term disparities and warrants further investigation.

Developing a More Comprehensive Understanding of Sleep Health

Subjective sleep quality was overrepresented in the included studies. Although it provides insight into perceived sleep satisfaction, two key points warrant consideration. First, sleep is multifaceted [18,68]. To advance a more comprehensive understanding of sleep health—and its developmental and clinical implications—broader research on other parameters, using both objective and subjective methods, is needed. Small cell sizes limited comparisons between objective and subjective assessments of sleep schedule/chronobiology and sleep consistency. Second, subjective sleep quality showed the strongest associations with mental health, especially in same-informant designs, suggesting possible inflation due to shared-method and shared-informant variance. Future research should prioritize multi-informant, multi-method designs to reduce such bias.

Conclusions

This meta-analysis is the first to synthesize evidence on whether sleep-mental health associations differ by race-ethnicity and SES. Although prior reviews underscore the importance of race-ethnicity and SES in shaping disparities [4,5,7,8], we found few significant omnibus moderation effects—possibly reflecting challenges in defining, measuring, and disentangling these sociodemographic factors. Additionally, by providing the first test of specificity across sleep parameters and mental health domains, this review underscores the value of a multifaceted conceptualization of sleep health and identifies actionable targets for intervention.

Supplementary Material

1
2
3

Practice Points.

  1. Poor sleep is associated with internalizing and externalizing symptoms and provides a viable therapeutic target for improving mental health among youth.

  2. Targeted intervention and prevention strategies are needed to rectify institutional and structural barriers underlying race-ethnicity- and SES-related sleep and mental health disparities.

  3. Subjective assessments of sleep quality capture meaningful individual differences in satisfaction with sleep that have important implications for mental health.

Research Agenda.

  1. Contextualize the intertwined nature of sociocultural factors to understand unique and shared contributions to sleep and mental health disparities.

  2. Adopt longitudinal, within-group designs to understand how unique risk and protective processes unfold over time within historically marginalized samples.

  3. Investigate explanatory mechanisms to understand how and why sleep parameters are differentially associated with mental health.

Acknowledgements:

This research was supported by Grants R01-HL136752 and R01-HL093246 from the National Heart, Lung, and Blood Institute and Grant R01-HD046795 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development awarded to Mona El-Sheikh. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institutes of Health. We wish to thank our research laboratory staff, particularly the undergraduate research assistants who assisted in title and abstract screening.

Abbreviations

AA

African American/Black

χ2

chi-squared

d

Cohen’s d

κ

Cohen’s kappa

CI

confidence interva

EA

European American/White

ICCs

intraclass correlation coefficients

MLMA

multilevel meta-analysis

n

number of observations

η2

partial eta squared

r

Pearson’s correlation coefficient

PRISMA

preferred reporting items for systematic reviews and meta-analyses

SES

socioeconomic status

US

United States

Footnotes

Conflict of Interests: The authors report there are no competing interests to declare.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

References marked with an asterisk indicate studies included in the meta-analysis. References for studies included in the meta-analysis but not cited in the main text are provided in “file s2”. “File s2” also includes references cited in “file s1”.

  • [1].Astill RG, Van der Heijden KB, Van Ijzendoorn MH, Van Someren EJW. Sleep, cognition, and behavioral problems in school-age children: a century of research meta-analyzed. Psychol Bull 2012;138(6):1109–38. [DOI] [PubMed] [Google Scholar]
  • [2].Jakobsson M, Sundin K, Högberg K, Josefsson K. “I want to sleep, but I can’t”: Adolescents’ lived experience of sleeping difficulties. J Sch Nurs 2020;38(5):449–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Bacaro V, Miletic K, Crocetti E. A meta-analysis of longitudinal studies on the interplay between sleep, mental health, and positive well-being in adolescents. Int J Clin Health Psychol 2023;24(1):100424. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Guglielmo D, Gazmararian JA, Chung J, Rogers AE, Hale L. Racial/ethnic sleep disparities in US school-aged children and adolescents: a review of the literature. Sleep Health 2018;4(1):68–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Etindele Sosso FA, Holmes SD, Weinstein AA. Influence of socioeconomic status on objective sleep measurement: a systematic review and meta-analysis of actigraphy studies. Sleep Health 2021;7(4):417–28. [DOI] [PubMed] [Google Scholar]
  • [6].Reiss F Socioeconomic inequalities and mental health problems in children and adolescents: a systematic review. Soc Sci Med 2013;90:24–31. [DOI] [PubMed] [Google Scholar]
  • [7].Buckhalt JA. Insufficient sleep and the socioeconomic status achievement gap: sleep and achievement. Child Dev Perspect 2011;5(1):59–65. [Google Scholar]
  • [8].El-Sheikh M, Gillis BT, Saini EK, Erath SA, Buckhalt JA. Sleep and disparities in child and adolescent development. Child Dev Perspect 2022;16(4):200–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Achenbach TM, Ivanova MY, Rescorla LA, Turner LV, Althoff RR. Internalizing/externalizing problems: review and recommendations for clinical and research applications. J Am Acad Child Adolesc Psychiatry 2016;55(8):647–656. [DOI] [PubMed] [Google Scholar]
  • [10].Heffernan ME, Macy ML. Trends in mental and physical health among youths. JAMA Pediatr 2025;179(6):683–685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Mojtabai R, Olfson M. Trends in mental disorders in children and adolescents receiving treatment in the state mental health system. J Am Acad Child Adolesc Psychiatry 2024;64(8):906–920 [DOI] [PubMed] [Google Scholar]
  • [12].Tkacz J, Brady BL. Increasing rate of diagnosed childhood mental illness in the United States: incidence, prevalence and costs. Public Health Pract 2021;2:100204. [Google Scholar]
  • [13].Boatswain-Jacques A-F, Dusablon C, Cimon-Paquet C, YuTong Guo É, Ménard R, Matte-Gagné C, et al. From early birds to night owls: a longitudinal study of actigraphy-assessed sleep trajectories during the transition from pre- to early adolescence. Sleep 2023;46(11). [Google Scholar]
  • [14].El-Sheikh M, Sadeh AI. Sleep and development: introduction to the monograph. Monogr Soc Res Child Dev 2015;80(1):1–14. [Google Scholar]
  • [15].Bei B, Wiley JF, Trinder J, Manber R. Beyond the mean: A systematic review on the correlates of daily intraindividual variability of sleep/wake patterns. Sleep Med Rev 2016;28:108–24. [DOI] [PubMed] [Google Scholar]
  • [16].Becker SP, Sidol CA, Van Dyk TR, Epstein JN, Beebe DW. Intraindividual variability of sleep/wake patterns in relation to child and adolescent functioning: a systematic review. Sleep Med Rev 2017;34:94–121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].*Thompson MJ, Gillis BT, Hinnant JB, Erath SA, Buckhalt JA, El-Sheikh M. Trajectories of actigraphy-derived sleep duration, quality, and variability from childhood to adolescence: downstream effects on mental health. Sleep 2024;47(8). [Google Scholar]
  • [18].Iii Sadeh A. Sleep assessment methods. Monogr Soc Res Child Dev 2015;80(1):33–48. [DOI] [PubMed] [Google Scholar]
  • [19].Conger RD, Conger KJ, Martin MJ. Socioeconomic status, family processes, and individual development. J Marriage Fam 2010;72(3):685–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Schenck-Fontaine A, Panico L. Many kinds of poverty: three dimensions of economic hardship, their combinations, and children’s behavior problems. Demography 2019;56(6):2279–305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Antonoplis S Studying socioeconomic status: conceptual problems and an alternative path forward. Perspect Psychol Sci 2022;18(2):275–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Easterbrook MJ, Kuppens T, Grigoryan L. Introduction to the special issue: nuances of social class and socioeconomic status (Introducción a este monográfico: los matices del concepto de clase social y del nivel socioeconómico ). Rev Psicol Soc 2023;38(3):493–507. [Google Scholar]
  • [23].Thompson MJ, Platts CR, Davies PT. Parent-child boundary dissolution and children’s psychological difficulties: a meta-analytic review. Psychol Bull 2024;150(7):873–919. [DOI] [PubMed] [Google Scholar]
  • [24].Ford CL, Harawa NT. A new conceptualization of ethnicity for social epidemiologic and health equity research. Soc Sci Med 2010;71(2):251–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].U.S. Census Bureau QuickFacts: United States [Internet]. Available from: https://www.census.gov/quickfacts/fact/table/US/PST045222
  • [26].Volpe VV, Smith NA, Skinner OD, Lozada FT, Hope EC, Del Toro J. Centering the heterogeneity of Black adolescents’ experiences: guidance for within-group designs among African diasporic communities. J Res Adolesc 2022;32(4):1298–311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Omenka O, Briggs A, Nunes J, Seixas A, Williams N, Jean-Louis G. Ethical and policy implications of racial and ethnic healthcare disparities in sleep health. J Racial Ethn Health Disparities 2024;11(4):2509–15. [DOI] [PubMed] [Google Scholar]
  • [28].Egede LE. Race, ethnicity, culture, and disparities in health care. J Gen Intern Med 2006;21(6):667–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Slopen N, Lewis TT, Williams DR. Discrimination and sleep: a systematic review. Sleep Med 2016;18:88–95 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Yip T, Xie M, Cham H, El Sheikh M. Linking ethnic/racial discrimination to adolescent mental health: sleep disturbances as an explanatory pathway. Child Dev 2022;93(4):973–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [31].*El-Sheikh M, Kelly RJ, Buckhalt JA, Hinnant JB. Children’s sleep and adjustment over time: the role of socioeconomic context. Child Dev 2010;81(3):870–83. [DOI] [PubMed] [Google Scholar]
  • [32].*Yip T. The effects of ethnic/racial discrimination and sleep quality on depressive symptoms and self-esteem trajectories among diverse adolescents. J Youth Adolesc 2015;44(2):419–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].*El-Sheikh M, Tu KM, Saini EK, Fuller-Rowell TE, Buckhalt JA. Perceived discrimination and youths’ adjustment: sleep as a moderator. J Sleep Res 2016;25(1):70–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Peverill M, Dirks MA, Narvaja T, Herts KL, Comer JS, McLaughlin KA. Socioeconomic status and child psychopathology in the United States: a meta-analysis of population-based studies. Clin Psychol Rev 2020;83:101933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Fuller-Rowell TE, Saini EK, El-Sheikh M. Social class discrimination during adolescence as a mediator of socioeconomic disparities in actigraphy-assessed and self-reported sleep. Sleep Med 2023;108:61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].*El-Sheikh M, Saini EK, Gillis BT, Kelly RJ. Interactions between sleep duration and quality as predictors of adolescents’ adjustment. Sleep Health 2019;5(2):180–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [37].Williamson AA, Zendarski N, Lange K, Quach J, Molloy C, Clifford SA, et al. Sleep problems, internalizing and externalizing symptoms, and domains of health-related quality of life: bidirectional associations from early childhood to early adolescence. Sleep 2021;44(1). [Google Scholar]
  • [38].*El-Sheikh M, Bub KL, Kelly RJ, Buckhalt JA. Children’s sleep and adjustment: a residualized change analysis. Dev Psychol 2013;49(8):1591–601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].*Shimizu M, Gillis BT, Buckhalt JA, El-Sheikh M. Linear and nonlinear associations between sleep and adjustment in adolescence. Behav Sleep Med 2020;18(5):690–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [40].Zink J, O’Connor SG, Blachman-Demner DR, Wolff-Hughes DL, Berrigan D. Examining the bidirectional associations between sleep duration, screen time, and internalizing symptoms in the ABCD study. J Adolesc Health 2024;74(3):496–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [41].*Shimizu M, Zeringue MM, Erath SA, Hinnant JB, El-Sheikh M. Trajectories of sleep problems in childhood: associations with mental health in adolescence. Sleep 2021;44(3). [Google Scholar]
  • [42].Fiese BH, Spagnola M. Narratives in and about families: an examination of coding schemes and a guide for family researchers. J Fam Psychol 2005;19(1):51–61. [DOI] [PubMed] [Google Scholar]
  • [43].Smith SR. Making sense of multiple informants in child and adolescent psychopathology: a guide for clinicians. J Psychoeduc Assess 2007;25(2):139–49. [Google Scholar]
  • [44].Maxwell SE, Cole DA. Bias in cross-sectional analyses of longitudinal mediation. Psychol Methods 2007;12(1):23–44. [DOI] [PubMed] [Google Scholar]
  • [45].Fraley RC, Roisman GI, Haltigan JD. The legacy of early experiences in development: formalizing alternative models of how early experiences are carried forward over time. Dev Psychol 2013;49(1):109–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [46].Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [47].McHugh ML. Interrater reliability: the kappa statistic. Biochem Med (Zagreb) 2012;22(3):276–82. [PMC free article] [PubMed] [Google Scholar]
  • [48].Syed M, Nelson SC. Guidelines for establishing reliability when coding narrative data. Emerg Adulthood 2015;3(6):375–87. [Google Scholar]
  • [49].Egger M, Smith GD. Bias in location and selection of studies. BMJ 1998;316(7124):61–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [50].Fernández-Castilla B, Aloe AM, Declercq L, Jamshidi L, Beretvas SN, Onghena P, et al. Estimating outcome-specific effects in meta-analyses of multiple outcomes: a simulation study. Behav Res Methods 2021;53(2):702–17. [DOI] [PubMed] [Google Scholar]
  • [51].Rodgers MA, Pustejovsky JE. Evaluating meta-analytic methods to detect selective reporting in the presence of dependent effect sizes. Psychol Methods 2021;26(2):141–60. [DOI] [PubMed] [Google Scholar]
  • [52].Johnson BT, Hennessy EA. Systematic reviews and meta-analyses in the health sciences: best practice methods for research syntheses. Soc Sci Med 2019;233:237–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [53].Viechtbauer W Conducting meta-analyses in R with the metafor Package. J Stat Softw 2010;36(3):1–48. [Google Scholar]
  • [54].Cohen J Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Erlbaum;1988. [Google Scholar]
  • [55].Cheung MW-L. Modeling dependent effect sizes with three-level meta-analyses: a structural equation modeling approach. Psychol Methods 2014;19(2):211–29. [DOI] [PubMed] [Google Scholar]
  • [56].Lupini F, Daniel LC, Mindell JA, Williamson AA. Variation in caregiver-reported child sleep patterns and problems by family socioeconomic indicators. J Dev Behav Pediatr 2023;44(8):e551–e558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [57].Slagt M, Dubas JS, Deković M, van Aken MAG. Differences in sensitivity to parenting depending on child temperament: a meta-analysis. Psychol Bull 2016;142(10):1068–110. [DOI] [PubMed] [Google Scholar]
  • [58].Kovacs M Children’s depression inventory. North Tonawanda, NY: Multi-Health Systems; 1992. [Google Scholar]
  • [59].Reynolds CR, Richmond BO. Revised Children’s Manifest Anxiety Scale (RCMAS-2) (2nd ed.). Los Angeles, CA: Western Psychological Services; 2008 [Google Scholar]
  • [60].*Benner AD, Fernandez CC, Límon KL Shifting and persisting in the face of life stressors: consequences for adolescent health. Appl Dev Sci 2024;28(1);22–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [61].*Kechter A, Leventhal AM Longitudinal association of sleep problems and distress tolerance during adolescence. Behav Med 2019;45(3):240–248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [62].Womack SD, Hook JN, Reyna SH, Ramos M. Sleep loss and risk-taking behavior: a review of the literature. Behav Sleep Med 2013;11(5):343–359. [DOI] [PubMed] [Google Scholar]
  • [63].McGlinchey EL, Harvey AG. Risk behaviors and negative health outcomes for adolescents with late bedtimes. J Youth Adolesc 2015;44(2):478–488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [64].Nunes S, Campbell MK, Klar N, Reid GJ, Stranges S. Relationships between sleep and internalizing problems in early adolescence: results from Canadian National Longitudinal Survey of Children and Youth. J Psychosom Res 2020;139:110279. [DOI] [PubMed] [Google Scholar]
  • [65].van Zundert RMP, van Roekel E, Engels RCME, Scholte RHJ. Reciprocal associations between adolescents’ night-time sleep and daytime affect and the role of gender and depressive symptoms. J Youth Adolesc 2015;44(2): 556–569. [DOI] [PubMed] [Google Scholar]
  • [66].Carneiro A, Soares I, Rescorla L, Dias P. Meta-analysis on parent–teacher agreement on preschoolers’ emotional and behavioural problems. Child Psychiatry Hum Dev 2021;52(4):609–618. [DOI] [PubMed] [Google Scholar]
  • [67].Lupini F, Williamson AA. Health disparities in pediatric sleep. Sleep Med Clin 2023;18(2):225–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [68].Meltzer LJ, Williamson AA, Mindell JA. Pediatric sleep health: it matters, and so does how we define it. Sleep Med Rev 2021;57:101425. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2126285-supplement-1.docx (33.9KB, docx)
2
NIHMS2126285-supplement-2.docx (30.1KB, docx)
3
NIHMS2126285-supplement-3.docx (617.4KB, docx)

RESOURCES