Social Jetlag Emerges in Preadolescent Children Despite Adequate Sleep Duration: Evidence for Child–Mother Circadian Misalignment Associations

ABSTRACT

Social jetlag, the misalignment between behaviourally expressed circadian timing and social schedules, has been extensively studied in adults but remains poorly characterized in preadolescent children. This cross‐sectional survey examined sleep–wake patterns in 972 Israeli children (mean age: 5.8 ± 1.6 years, range: 4–10) and their mothers (mean age: 37.1 ± 5.7 years, range: 22–49), who reported sleep behaviours using adapted Munich Chronotype Questionnaire items and an item about evening screen exposure between 16:00 and sleep onset. Children's mean sleep duration was 10.45 h, with 57.1% showing 30 min or less variation between free and school days. Surprisingly, children exhibited higher social jetlag than their mothers (64.9 ± 47.7 vs. 46.6 ± 52.9 min, p < 0.001), increasing from 54.5 min at age 4 to 107.1 min at age 10. Positive correlations existed between mother–child chronotypes (ρ = 0.224, p < 0.001) and social jetlag (ρ = 0.222, p < 0.001). Evening screen exposure averaged 84 ± 66 min and correlated with age (ρ = 0.240, p < 0.001). Later chronotype was the strongest predictor of child social jetlag; higher body mass index was associated with greater social jetlag, whereas older age, secular lifestyle, longer sleep, more consistent sleep duration and screen exposure were associated with lower social jetlag. Maternal social jetlag independently predicted greater child social jetlag. Young children experience substantial social jetlag, challenging assumptions about sleep timing regularity in preadolescence. Adequate sleep duration coexisting with high social jetlag, a marker linked to adverse metabolic and behavioural outcomes, suggests paediatric guidelines may need to incorporate weekly sleep timing consistency. Child–mother sleep interconnections highlight the necessity of exploring family‐based interventions beyond focus on sleep quantity.

1. Introduction

To most parents, a good night's sleep is synonymous with providing sufficient sleep duration for their child. Indeed, sufficient sleep duration is a critical basic component of physical and mental health (Chaput et al. 2020; Matricciani et al. 2012). However, emerging evidence suggests that sleep timing regularity may be equally important for development and well‐being as sleep duration (Sletten et al. 2023; Windred et al. 2024). The concept of social jetlag, which describes the misalignment between an individual's internal circadian rhythm and socially imposed schedules (Roenneberg et al. 2007), has gained increasing attention in adults and adolescents, but is rarely studied in preadolescent children.

1.1. Social Jetlag: A Growing Concern

Social jetlag, quantified as the difference between sleep timing on work/school days compared to free days, has emerged as an important marker of circadian misalignment associated with metabolic dysfunction, mood disorders, and cognitive impairment (Li et al. 2024; Roenneberg et al. 2012; Wong et al. 2015). Experimental studies have demonstrated direct adverse effects of circadian misalignment on glucose metabolism, insulin sensitivity and inflammatory markers, supporting plausible causal pathways independent of sleep duration (Buxton et al. 2012; Scheer et al. 2009). In children, social jetlag has been associated with obesity, behavioural problems, and temperament difficulties (Choubai et al. 2025; Doi et al. 2015; Giannoumis et al. 2022; Higgins et al. 2021; Stoner et al. 2018), with a reported breakpoint at age 5 years in a large cross‐sectional sample of 18,323 children (Randler et al. 2017). Similar associations extend to early adolescence, including cardiometabolic risk and cognitive performance (Cespedes Feliciano et al. 2019; Yang et al. 2023).

Unlike adults, whose sleep schedules reflect personal choices and social time pressure (Korman et al. 2020), children's sleep patterns are heavily influenced by parental decisions and family routines, involving parental limit‐setting, screen time exposure, and family light exposure habits (Illingworth et al. 2025). A likely driver of paediatric social jetlag is parents' tendency to relax rules and routines on free days, a recognized factor in childhood insomnia and behavioural sleep problems (Buxton et al. 2015; Kang and Kim 2021; Knappe et al. 2020). Social jetlag is conceptually and mathematically linked to chronotype rather than sleep duration per se: individuals with a later biological clock are more likely to experience a larger discrepancy between their preferred and socially imposed sleep timing (Roenneberg et al. 2007, 2019), making chronotype development central to understanding paediatric social jetlag.

1.2. Chronotype and Sleep Duration

Chronotype is estimated from free‐day sleep timing corrected for sleep debt accumulated over the work week, yielding MSFsc as the standard behavioural estimate (Ghotbi et al. 2020; Roenneberg et al. 2007). In children, chronotype is assessed via parental proxy report (Doi et al. 2015; Randler et al. 2017; Zimmermann 2016). Current guidelines from the American Academy of Sleep Medicine recommend 10–13 h for preschoolers and 9–11 h for school‐age children, focusing on quantity rather than timing or regularity (Paruthi et al. 2016; Williams et al. 2013). As children mature, their delaying chronotype increasingly conflicts with fixed social schedules, potentially producing circadian misalignment (Boatswain‐Jacques et al. 2023; Tétreault et al. 2019; Zimmermann 2016), though the extent of delays in young children remains poorly understood.

Childhood sleep is embedded within family systems where parents and children mutually influence each other's sleep (Buxton et al. 2015; Fisher et al. 2012; O'Callaghan et al. 2021). Parents serve as primary regulators of children's bedtime routines and models of sleep behaviours (Flint Bretler et al. 2022), and enforcing bedtime and technology rules has been associated with longer school‐day sleep duration (Buxton et al. 2015). Conversely, children's sleep problems can significantly impact parental sleep and family functioning (Fisher et al. 2012; O'Callaghan et al. 2021; Turnbull et al. 2013). Evening screen exposure is a recognized factor in children's sleep onset and duration: screen light suppresses melatonin, with pre‐pubertal children showing twice the suppression of adults (Hartstein et al. 2025; Higuchi et al. 2014; Wang et al. 2025). The overall quality of evidence on sleep timing in children has been rated ‘very low’ by GRADE, and no sleep timing recommendations for children are currently proposed (Dutil et al. 2022).

The current study addressed these gaps by characterizing sleep–wake patterns in children aged 4–10 years and their mothers. We expected most children to meet age‐appropriate sleep duration recommendations with chronotypes earlier than their mothers and, assuming reduced social constraints in early childhood, lower social jetlag than their mothers. We also expected a high prevalence of evening screen exposure and significant mother–child correlations in sleep parameters, reflecting the interconnected nature of family sleep. By examining individual and dyadic patterns, this research provides a foundation for family‐centered approaches to promote sleep health in young children.

2. Methods

The study received approval from the Ethics Committee of the School of Health Sciences at Ariel (approval number: AU‐HEA‐YF‐20221025). This quantitative field study was conducted among mothers of children aged 4–10 years. An online questionnaire assessing sleep and daily behaviour was distributed via social media parenting forums to reach diverse populations. The questionnaire was completed by mothers anonymously after signing informed consent.

Between February 2022 and March 2023, 1951 Israeli mothers of 4–10‐year‐old children opened the survey link. Data collection was paused during major holidays and daylight saving transitions. Exclusions were applied for missing data, incomplete questionnaires, extreme sleep durations (< 3 or > 14 h), considered outside plausible behavioural ranges and likely reflecting data entry error or underlying sleep pathology (Paruthi et al. 2016). Additional exclusions applied for infants under 1 year in the household and severe chronic illness in mother or child, yielding a final analytic sample of 972 dyads.

2.1. Tools

The demographic module included questions on age, height, weight, number of children under 18, marital status, lifestyle, ethnicity, routine activities and work hours, health status, and the presence of chronic medical diagnoses. Height and weight were used to calculate the Body Mass Index (BMI) of the child and the mother as (kg/m²).

Daily behaviours were assessed using questions from the ultra‐short Munich Chronotype Questionnaire (μMCTQ) (Ghotbi et al. 2020). Mothers reported habitual sleep onset and offset times on school days and free days separately, referring to the last ‘typical’ weeks. The original μMCTQ items were modified to refer to the child in the third person and to reflect the school/kindergarten context. For example, the item asking about falling asleep time on workdays was adapted to: ‘When does your child usually fall asleep on weekdays (school or kindergarten days)?’ Sleep duration (SD), chronotype (MSFsc), social jetlag (SJL), and day‐type sleep duration inconsistency (∆SD) were calculated as previously described (Ghotbi et al. 2020) and illustrated in Figure 1.

FIGURE 1.

A hypothetical graphical display of the daily sleep–wake behaviour indexes used in this report, calculated separately per participant (mothers and children) from habitual sleep onset and offset times on workdays and free days. MSW (turquoise) and MSF (blue) represent mid‐sleep timing on workdays and free days, respectively (local time). MSFsc (yellow dot) is an estimate of chronotype, corrected for sleep debt accumulated over the work week: MSFsc = MSF−0.5 × (SDF − SDweek). Day‐type sleep duration inconsistency (∆SD = SDF − SDW) reflects the difference in sleep duration between free and school/work days; positive values indicate longer sleep on free days. Social jetlag (SJL = MSF − MSW) reflects the difference in sleep timing between free and work/school days; positive values indicate later sleep timing on free days. The X‐axis represents local time in hh:mm format.

Child's habitual screen time exposure between 16:00 and sleep onset was reported using an open question format (hh:mm), reflecting typical behaviour across the week without distinction between school and free days.

2.2. Sample

The analytic sample included 972 mother–child dyads. Questionnaires were completed by mothers from Jewish (96.4%) and Arabic (3.6%) communities. Mothers ranged in age from 22 to 49 years (mean: 37.1 ± 5.7 years, median: 37.0, IQR = 9), with a mean BMI of 25.1 ± 5.2. Most mothers (94.4%) were married and employed (93%). Most mothers (67.1%) reported a religious family lifestyle. The mean number of children in the family (ages: 1–18) was 3.14 ± 1.41, indicating that on average, children in the current sample had two siblings. Children ranged in age from 4 to 10 years (mean: 5.8 ± 1.6 years, median: 5.0, IQR = 3), with a mean BMI of 18.03 ± 4.8, comprising 47.7% girls and 52.3% boys. Mothers who had several children within the 4–10 year range were requested to report on their youngest child, resulting in decreasing sample sizes across age groups: 4 years—270, 5 years—225, 6 years—191, 7 years—120, 8 years—82, 9 years—49, 10 years—35. Overall, 50.9% of the children sample were preschoolers (ages: 4–5) and 49.1% were school‐age children (ages: 6–10).

2.3. Statistical Approach

The data were analysed using SPSS 29.0 (IBM Corp., Armonk, NY, USA) and R (version 4.0.5). Conservative threshold criteria for normality based on skewness (±1.5) were applied. Paired t‐tests were used to assess day‐type differences (free vs. school days) in sleep parameters for children and mothers. One‐way ANOVAs were used to examine differences in sleep timing across the seven age groups (4–10 years) separately for school days and free days. Where interaction effects with child age were of interest, repeated measures general linear model (rm‐GLM) analyses were performed. Each rm‐GLM included day‐type (free vs. school days) as the within‐subject factor and child age as a between‐subjects covariate, with sleep onset and offset times modelled separately as dependent variables.

Bivariate partial correlations, child‐age corrected, served as exploratory and descriptive analyses of child–mother associations. Spearman correlations (ρ _S) were used to describe associations between dependent measures and child age. Pearson correlations (ρ) were used to describe associations between normally distributed dependent measures. For interpretation, Bonferroni correction was applied within three pre‐specified construct families: (1) correlations with child social jetlag (SJL × age, SJL × BMI, SJL × ∆SD; n = 3, corrected threshold p < 0.017); (2) mother–child parameter correlations (SDweek, MSFsc, SJL; n = 3, corrected threshold p < 0.017) and (3) age correlations with sleep and behavioural parameters (age × SDweek, age × MSFsc, age × SJL, age × screen exposure; n = 4, corrected threshold p < 0.013). All reported correlations survived Bonferroni corrections.

Two linear regression models with simultaneous forced entry of all predictors were used to identify factors associated with child social jetlag. Model 1 included child age, sex, BMI, chronotype (MSFsc), average sleep duration (SDweek), day‐type sleep duration irregularity (∆SD), family religious lifestyle (religious/secular) and evening screen time exposure (between 16:00 and sleep onset). Model 2 expanded upon Model 1 by adding maternal variables: age, average sleep duration, chronotype and social jetlag. Collinearity diagnostics were performed for both models, with all VIF values remaining below 2. Statistical significance was set at p < 0.05.

3. Results

3.1. Sleep Duration, Chronotype and Social Jetlag of Children and Their Mothers

Before investigating their interrelationships, we first described the dependent measures for children and mothers (Table 1).

TABLE 1.

Descriptive statistics across dependent measures.

Measure	Children	Mothers
SDweek, min	627.2 ± 63.5	437.5 ± 56.8
Preschoolers	643.8 ± 43.5	439.3 ± 55.8
Schoolers	608.6 ± 47.0	436.0 ± 57.7
SDF, min	623.3 ± 78.6	498.3 ± 100.1
Preschoolers	634.4 ± 65.3	492.2 ± 93.4
Schoolers	610.6 ± 65.5	506.1 ± 104.8
SDW, min	628.7 ± 71.6	413.2 ± 61.6
Preschoolers	647.5 ± 48.8	418.1 ± 62.4
Schoolers	607.8 ± 52.5	407.9 ± 60.2
ΔSD, min	−5.1 ± 70.4	85.9 ± 106.3
Preschoolers	−12.8 ± 71.0	74.1 ± 102.9
Schoolers	2.8 ± 68.9	98.2 ± 108.4
MSFsc, local time	02:31 ± 01:04	03:12 ± 01:18
Preschoolers	02:12 ± 00:49	03:08 ± 01:09
Schoolers	02:49 ± 00:54	03:17 ± 01:27
MSF, local time	02:38 ± 01:00	03:40 ± 01:02
Preschoolers	02:19 ± 00:50	03:35 ± 01:00
Schoolers	02:59 ± 00:58	03:48 ± 01:04
MSW, local time	01:33 ± 00:43	02:54 ± 00:42
Preschoolers	01:23 ± 00:32	02:54 ± 00:42
Schoolers	01:44 ± 00:35	02:53 ± 00:42
SJL, min	64.9 ± 47.8	46.6 ± 52.9
Preschoolers	56.9 ± 40.8	39.4 ± 49.0
Schoolers	74.7 ± 45.0	54.6 ± 54.8

Note: Values represent mean ± SD for the total sample and by group: Preschoolers (ages: 4–5, n = 495) and schoolers (ages: 6–10, n = 477).

Abbreviations: ∆SD, day‐type sleep duration inconsistency; MSF, mid‐sleep time on free days; MSFsc, chronotype, phase of entrainment; MSW, mid‐sleep time on workdays; SDF, sleep duration on free days; SDW, sleep duration on school/work days; SDweek, average weekly sleep duration; SJL, social jetlag, day‐type sleep timing irregularity.

Mean sleep duration (SDweek) of children was 10.45 h (627.2 ± 63.5 min, skewness: −0.73), with no sex differences observed (Figure 2, left column). Age and SDweek correlated (ρ = −0.48, p < 0.001), indicating that older children presented progressively shorter SDweek: from 10.9 h in the 4‐year‐olds to 9.5 h in the 10‐year‐olds. The paired t‐test analysis showed that children's sleep duration across day‐type (free/school) was rather consistent; the average ∆SD was small (−5.14 ± 70.4 min, skewness −0.25), though significant (t = −2.276, p = 0.023, Cohen's d = −0.073) (Figure 2, right column). Specifically, 57.1% of the children presented a sleep duration difference of 30 min or less across day‐type, 22.2% slept > 30 min longer on school days, whereas 20.7% slept > 30 min longer on free days. Preschoolers slept on average 12.8 ± 71.0 min less on free days than on school days (t = −4.019, p < 0.001, Cohen's d = −0.181), whereas schoolers showed no significant day‐type difference in sleep duration (mean difference 2.8 ± 68.9 min, t = 0.897, p = 0.370). These findings indicate that most children in the sample sleep within the age‐appropriate ranges recommended by the American Academy of Sleep Medicine and present relatively high consistency in sleep duration across free and school days.

FIGURE 2.

Habitual sleep duration and sleep duration inconsistency of children (turquoise) and mothers (blue). Upper panel: Distributions of SDweek and ∆SD (% of group). Middle panel: Boxplots of SDweek, MSFsc, and SJL (box borders: 25th–75th percentiles; line: Median). Lower panel: Mother–child linear correlations (red: Regression line).

On average, SDweek of mothers was 7.2 h (mean: 437.5 ± 56.8 min, skewness: 0.95), on average 3.1 h shorter than SDweek of their children (Figure 2, left column). No differences were observed in SDweek of mothers of preschoolers as compared to schoolers (t = −0.917, p = 0.180). A negligible though significant correlation was observed between SDweek indexes of children and their mothers (ρ = 0.088, p = 0.006, age‐corrected) (Figure 2, left column). Maternal sleep duration irregularity across day‐type was robust; on average, mothers reported 85.9 ± 106.3 min longer sleep on free days (skewness: 1.41, t = −25.20, p < 0.001, Cohen's d = 0.808). Only 28.3% of mothers presented a ∆SD ≤ 30 min, 6.3% slept > 30 min longer on school days, whereas 65.4% slept > 30 min longer on free days, indicative of a robust catch‐up sleep phenomenon on free days. Mothers of preschoolers had a smaller ∆SD (t = −3.557, p < 0.001, Cohen's d = −0.228). Sleep duration irregularity indexes (∆SDs) of children and mothers did not correlate (ρ = −0.040, p = 0.215).

Generally, children presented a significantly earlier chronotype (MSFsc) than their mothers (mean: 02:31 ± 01:04, skewness: 0.27 and 03:12 ± 1:18, skewness: 0.10, respectively, t = −15.462, p < 0.001, Cohen's d = 0.496), with an average mother–child difference in chronotypes of ~41 min (Figure 2). In preschoolers this difference was larger (mean: 54.7 ± 76.5 min, t = −15.912, p < 0.001, Cohen's d = 0.715) than in schoolers (mean: 28.3 ± 89.6 min, t = −6.896, p < 0.001, Cohen's d = 0.315). No sex differences in children's MSFsc were observed (t = 1.141, p = 0.254). Chronotype and sleep duration irregularity index (∆SD) of children did not correlate (ρ = 0.053, p = 0.099). Child age and MSFsc correlated (ρ _S = 0.380, p < 0.001), indicating that older children present later chronotypes. A trend for earlier chronotype was observed for mothers of preschoolers as compared to schoolers (t = −1.889, p = 0.059, Cohen's d = 0.121). Importantly, chronotypes of mothers and their children significantly correlated (ρ = 0.224, p < 0.001, age‐corrected) (Figure 3, left column).

FIGURE 3.

Chronotype (MSFsc) and social jetlag of children (turquoise) and mothers (blue). Upper panel: Distributions of MSFsc and SJL (% of group). Middle panel: Boxplots of MSFsc and SJL (box borders: 25th–75th percentiles; line: Median). Lower panel: Mother–child linear correlations (red: Regression line).

A notable misalignment of children's sleep timing has been revealed, with the average child social jetlag (SJL) exceeding 1 h (mean: 64.9 ± 47.7 min, skewness: 0.64) (Figure 3, right column). Sex was insignificant (t = 1.175, p = 0.140). Age and child SJL moderately correlated (ρ _S = 0.240, p < 0.001), with 4‐year‐olds presenting a mean SJL of 54.5 ± 36.61 min and 10‐year‐olds presenting a mean SJL of 107.14 ± 56.55 min. Child SJL significantly correlated with their BMI (ρ = 0.139, p = 0.004). Sleep duration irregularity (∆SD) inversely correlated with child SJL (ρ = −0.268, p < 0.001), indicating that children who sleep less on free days relative to school days tend to have smaller SJL. Surprisingly, child SJL values were higher compared to the SJL of their mothers (mean: 46.6 ± 52.9 min, skewness: −0.09, t = 9.779, p < 0.001, Cohen's d = 0.314), with an average difference of 18.44 ± 58.81 min. Higher SJL values in children were found consistently across age groups: preschoolers (mean difference: 17.1 ± 56.1 min, t = 6.792, p < 0.001, Cohen's d = 0.305), schoolers (mean difference: 19.8 ± 61.5 min, t = 7.035, p < 0.001, Cohen's d = 0.322). Mothers of schoolers demonstrated a significantly larger SJL than mothers of preschoolers (mean difference: 15.11 ± 3.31 min, t(970) = −4.563, p < 0.001, Cohen's d = −0.293). Child SJL significantly correlated with the SJL of their mothers (ρ = 0.222, p < 0.001, age‐corrected), indicating that sleep timing irregularity is a common mother–child behavioural trend (Figure 3, right column).

3.2. Daily Behaviours: Sleep Onset and Offset Times of Children and Their Mothers

To explore the behavioural patterns subserving the surprising pattern of social jetlag in children, detailed analyses of sleep onset and offset timing during school and free days were performed. Figure 4 displays the distribution of sleep onset and offset times for children and mothers across age groups on school and free days; corresponding mean ± SD, median and IQR values are reported in Table 2.

FIGURE 4.

Sleep patterns of children and mothers by child age group. Sleep onset (blue—children/grey—mothers) and offset times (orange—children/white—mothers) on school and free days, respectively. See Table 2 for mean ± SD, median and IQR values by age group.

TABLE 2.

Sleep onset and offset times of children and mothers on school and free days by child age.

Age (years)	N	Child sleep onset	Mother sleep onset	Child sleep offset	Mother sleep offset
School days
4	270	19:51 ± 00:40 19:45 (00:45)	23:25 ± 01:05 23:30 (01:15)	06:47 ± 00:40 06:45 (00:30)	06:24 ± 00:38 06:30 (00:45)
5	225	20:08 ± 00:39 20:00 (00:45)	23:26 ± 01:02 23:30 (01:00)	06:45 ± 00:42 07:00 (00:30)	06:23 ± 00:40 06:30 (00:45)
6	191	20:21 ± 00:39 20:15 (00:30)	23:30 ± 01:03 23:30 (01:15)	06:48 ± 00:36 06:45 (00:30)	06:21 ± 00:36 06:30 (00:45)
7	120	20:35 ± 00:45 20:30 (01:00)	23:34 ± 01:02 23:22 (01:15)	06:48 ± 00:45 06:45 (00:30)	06:17 ± 00:37 06:15 (00:45)
8	82	20:54 ± 00:40 21:00 (00:45)	23:30 ± 01:00 23:30 (01:15)	06:43 ± 00:51 07:00 (00:30)	06:17 ± 00:34 06:30 (00:45)
9	49	21:06 ± 00:42 21:00 (00:45)	23:18 ± 00:59 23:15 (01:30)	06:50 ± 00:19 07:00 (00:30)	06:06 ± 00:50 06:00 (01:00)
10	35	21:31 ± 00:40 21:30 (01:00)	23:22 ± 01:03 23:15 (01:30)	06:53 ± 00:53 07:00 (00:15)	06:12 ± 00:55 06:15 (00:45)
Free days
4	270	20:55 ± 00:55 21:00 (01:15)	23:29 ± 01:22 23:30 (01:45)	07:32 ± 00:56 07:30 (01:00)	07:37 ± 01:10 07:30 (01:00)
5	225	21:11 ± 01:01 21:00 (01:22)	23:33 ± 01:12 23:30 (01:00)	07:42 ± 01:06 07:45 (01:30)	07:45 ± 01:14 07:30 (01:30)
6	191	21:31 ± 00:59 21:30 (01:00)	23:30 ± 01:32 23:30 (01:15)	07:49 ± 01:00 08:00 (01:30)	08:01 ± 01:16 08:00 (01:30)
7	120	21:49 ± 00:55 21:52 (01:30)	23:48 ± 01:21 23:45 (01:41)	08:03 ± 01:03 08:00 (01:22)	08:03 ± 01:06 08:00 (01:45)
8	82	22:01 ± 00:53 22:00 (00:49)	23:46 ± 01:18 23:52 (01:30)	08:01 ± 01:15 08:00 (01:30)	08:00 ± 01:06 08:00 (01:45)
9	49	22:25 ± 00:54 22:30 (01:00)	23:39 ± 01:19 23:30 (01:52)	08:29 ± 01:11 08:30 (01:30)	07:47 ± 01:22 07:45 (02:08)
10	35	23:03 ± 01:08 23:00 (01:30)	00:09 ± 02:29 00:00 (02:00)	08:56 ± 01:34 09:00 (02:00)	08:19 ± 02:08 08:15 (02:30)

Note: Top line per cell: Mean ± SD. Bottom line per cell (italics): Median (IQR).

Age‐dependent delays in children's sleep onset times were observed on both school and free days across all age groups (Figure 4, blue—child sleep onset, orange—child sleep offset): on school days from ~20:00 in the 4‐year‐olds to ~21:30 in the 10‐year‐olds, and on free days from ~21:00 in the 4‐year‐olds to ~23:00 in the 10‐year‐olds. The magnitude of sleep onset delay between free and school days increased progressively with age, from 63.9 ± 48.8 min in the 4‐year‐olds to 91.7 ± 57.1 min in the 10‐year‐olds, with all age groups showing significant day‐type differences (Table 3). The rm‐GLM confirmed that sleep onset occurred consistently later on free than school days across the full sample (within‐subjects day‐type effect: F(1,965) = 1358.747, p < 0.001, η ² p = 0.585), that the between‐subjects effect of child age on sleep onset timing was significant (F(6,965) = 71.031, p < 0.001, η ² p = 0.306), and that older children showed larger delays in sleep onset between school and free days (day‐type × age interaction: F(6,965) = 3.147, p = 0.005, η ² p = 0.019).

TABLE 3.

Day‐type differences in sleep onset and offset times across child age groups.

		Delta [free‐school]
		Child sleep onset			Child sleep offset
Age (years)	N	Mean difference (min)	t	p	Mean difference (min)	t	p
4	270	63.93 ± 48.80	21.510	< 0.001	45.11 ± 50.98	14.538	< 0.001
5	225	62.53 ± 48.47	19.349	< 0.001	56.86 ± 67.03	12.725	< 0.001
6	191	70.20 ± 49.52	19.591	< 0.001	61.17 ± 47.80	17.686	< 0.001
7	120	74.62 ± 42.89	19.060	< 0.001	75.25 ± 63.19	13.045	< 0.001
8	82	66.95 ± 41.64	14.558	< 0.001	78.65 ± 79.14	9.000	< 0.001
9	49	79.28 ± 49.56	11.198	< 0.001	98.87 ± 64.65	10.705	< 0.001
10	35	91.71 ± 57.14	9.494	< 0.001	122.57 ± 81.71	8.874	< 0.001

Note: Values represent mean difference ± SD (free minus school days) in minutes, with paired t‐test results per age group. Positive values indicate later sleep onset or offset on free days relative to school days.

In contrast, sleep offset times on school days were remarkably consistent across children's age groups, with a sample mean of 06:47 ± 00:41 (one‐way ANOVA: F(6,965) = 0.424, p = 0.863, η ² p = 0.003). However, on free days, robust age‐dependent delays in sleep offset times were observed, from ~07:30 in the 4‐year‐olds to ~09:00 in the 10‐year‐olds (F(6,965) = 14.352, p < 0.001, η ² p = 0.082). Sleep offset delays between free and school days increased progressively with age, from 45.1 ± 51.0 min in the 4‐year‐olds to 122.6 ± 81.7 min in the 10‐year‐olds, with all age groups showing significant day‐type differences (Table 3). The rm‐GLM confirmed that sleep offset occurred consistently later on free than school days across the full sample (within‐subjects day‐type effect: F(1,965) = 950.933, p < 0.001, η ² p = 0.496), that the between‐subjects effect of child age on sleep offset timing was significant (F(6,965) = 8.342, p < 0.001, η ² p = 0.049), and that older children showed larger delays in sleep offset between school and free days (day‐type × age interaction: F(6,965) = 14.537, p < 0.001, η ² p = 0.083).

On free days, mothers showed systematic delays, with both sleep onset and offset occurring later than on workdays (Figure 4, grey and white boxes). The rm‐GLM for maternal sleep onset confirmed a significant within‐subjects day‐type effect (F(1,965) = 20.769, p < 0.001, η ² p = 0.021), with no significant between‐subjects effect of child age (F(6,965) = 1.216, p = 0.295, η ² p = 0.008). A significant interaction between day‐type and child age was observed for sleep onset (F(6,965) = 2.154, p = 0.045, η ² p = 0.013), indicating that mothers of older children show larger delays in sleep onset on free days relative to workdays. The rm‐GLM for maternal wakeup time confirmed a large within‐subjects day‐type effect (F(1,965) = 1328.317, p < 0.001, η ² p = 0.579), with no significant between‐subjects effect of child age (F(6,965) = 1.610, p = 0.141, η ² p = 0.010). A significant interaction between day‐type and child age was also observed for wakeup time (F(6,965) = 8.189, p < 0.001, η ² p = 0.048), indicating that mothers of older children show larger delays in wakeup time on free days compared to workdays relative to mothers of younger children.

As expected, mothers consistently reported falling asleep later than their children on both workdays (mean difference = 188.3 min, t(971) = −80.377, p < 0.001, Cohen's d = −2.578) and free days (mean difference = 128.8 min, t(971) = 42.541, p < 0.001, Cohen's d = 1.365). On workdays, mothers woke up earlier than their children (mean difference = 26.6 min, t(971) = 16.749, p < 0.001, Cohen's d = 0.537). Interestingly, on free days, mothers' wake‐up times were not significantly different from their children's (mean difference = 0.5 min, t(971) = −0.188, p = 0.851, Cohen's d = −0.006), indicating that the combination of maternal catch‐up sleep and child sleep‐in on free days brings wake‐up times into alignment across the dyad.

3.3. Evening Screen Exposure

On average, between 16:00 and sleep onset, children were habitually exposed to screens for 84 ± 66 min. Sex was insignificant (t = 0.166, p = 0.868). A significant correlation between age and screen exposure was found (ρ _S = 0.240, p < 0.001), with 4‐year‐olds' screen time lasting on average 55.6 ± 55.2 min and 10‐year‐olds' exposure lasting 176.4 ± 86.4 min.

3.4. Child Social Jetlag Predictors

To comprehensively assess the factors associated with child social jetlag (SJL), two linear regression models were estimated. Model 1 included child theoretical predictors: age, sex, BMI, chronotype (MSFsc), sleep duration (SDweek), family religious lifestyle (religious/secular), day‐type sleep duration irregularity (∆SD), and evening screen time exposure (between 16:00 and sleep onset). This model accounted for 60.8% of the variance in child SJL (Adjusted R ² = 0.604, p < 0.001) (see Table 4). Model 2 expanded upon Model 1 by incorporating maternal variables: age, BMI, sleep duration, chronotype and maternal social jetlag. This expanded model accounted for 62.3% of the variance in child SJL (Adjusted R ² = 0.618), representing a significant improvement over Model 1 (ΔR ² = 0.016, F‐change (5946) = 7.937, p < 0.001). A later child chronotype remained the strongest positive predictor of child SJL (β = 0.754, p < 0.001). Consistent with Model 1, greater ∆SD (β = −0.253, p < 0.001), older child age (β = −0.081, p = 0.002), longer sleep duration (β = −0.147, p < 0.001), and greater evening screen time exposure (β = −0.101, p < 0.001) were independently associated with reduced child SJL. Child BMI (β = 0.050, p = 0.019) and religious family lifestyle (β = 0.074, p = 0.001) remained significant positive predictors. Child sex remained insignificant (β = −0.021, p = 0.292). Among the newly added maternal variables, maternal SJL was significantly associated with increased child SJL (β = 0.112, p < 0.001), and older maternal age was significantly associated with reduced child SJL (β = −0.098, p < 0.001). Maternal BMI (β = 0.011, p = 0.593), sleep duration (β = −0.010, p = 0.636), and chronotype (β = −0.037, p = 0.105) were not significant predictors. Full statistics are presented in Table 4.

TABLE 4.

Linear regressions predicting child social jetlag (SJL, min).

Predictor	B	SE	β	t	p	95% CI for B	Tol.	VIF
Model 1, R 2 = 0.608, adj. R 2 = 0.604, F(8,951) = 184.104, p < 0.001
Child age	−2.804	0.665	−0.107	−4.218	< 0.001	[−4.108, −1.499]	0.642	1.557
Child sex	−1.801	1.791	−0.020	−1.006	0.315	[−5.315, 1.712]	0.995	1.005
BMI (child)	0.553	0.193	0.061	2.871	0.004	[0.175, 0.932]	0.920	1.087
∆SD (child)	−0.157	0.013	−0.252	−12.136	< 0.001	[−0.182, −0.132]	0.956	1.046
MSFsc (child)	0.607	0.018	0.757	32.996	< 0.001	[0.571, 0.644]	0.784	1.275
Religious lifestyle	2.697	0.895	0.065	3.014	0.003	[0.941, 4.453]	0.879	1.137
SDweek (child)	−0.117	0.021	−0.129	−5.437	< 0.001	[−0.159, −0.075]	0.735	1.360
Screen exposure	−3.829	0.929	−0.099	−4.121	< 0.001	[−5.652, −2.005]	0.711	1.407
Model 2, R 2 = 0.623, adj. R 2 = 0.618, F(13,946) = 120.479, p < 0.001
Child age	−2.133	0.681	−0.081	−3.131	0.002	[−3.470, −0.796]	0.590	1.696
Child sex	−1.856	1.760	−0.021	−1.055	0.292	[−5.309, 1.597]	0.994	1.006
BMI (child)	0.453	0.192	0.050	2.355	0.019	[0.075, 0.830]	0.893	1.120
∆SD (child)	−0.158	0.013	−0.253	−12.375	< 0.001	[−0.183, −0.133]	0.950	1.053
MSFsc (child)	0.605	0.019	0.754	32.212	< 0.001	[0.568, 0.642]	0.727	1.376
Religious lifestyle	3.039	0.923	0.074	3.293	0.001	[1.228, 4.850]	0.797	1.254
SDweek (child)	−0.133	0.021	−0.147	−6.206	< 0.001	[−0.175, −0.091]	0.709	1.409
Screen exposure	−3.882	0.924	−0.101	−4.200	< 0.001	[−5.695, −2.068]	0.693	1.443
Mother age	−0.755	0.184	−0.098	−4.096	< 0.001	[−1.117, −0.393]	0.690	1.449
BMI (mother)	0.092	0.172	0.011	0.534	0.593	[−0.246, 0.430]	0.941	1.062
MSFsc (mother)	−0.022	0.013	−0.037	−1.620	0.105	[−0.048, 0.005]	0.768	1.303
SDweek (mother)	−0.007	0.016	−0.010	−0.473	0.636	[−0.039, 0.024]	0.974	1.027
SJL (mother)	0.096	0.020	0.112	4.787	< 0.001	[0.057, 0.136]	0.722	1.384

Note: All VIF values < 2. p‐values are two‐tailed.

Abbreviations: ∆SD, free‐minus‐school day sleep duration discrepancy; B, unstandardized coefficient; CI, confidence interval; MSFsc, chronotype; SDweek, average weekly sleep duration; SE, standard error; Tol., tolerance; β, standardized coefficient.

4. Discussion

4.1. Child Social Jetlag: Early Development and Key Characteristics

Our study, conducted in a sample of 972 young Israeli children, demonstrated that social jetlag is already substantial in preadolescent children as young as 4. The progressive increase in social jetlag with age, from ~54 min at age 4 to ~107 min at age 10, is consistent with several recent studies indicating that behavioural circadian misalignment may develop during early childhood (Boatswain‐Jacques et al. 2023; Randler et al. 2017; Tétreault et al. 2019). Child chronotype was the strongest predictor of social jetlag, supporting the fundamental role of individual circadian timing in determining sleep timing irregularity (Zimmermann 2016). Longer sleep duration was associated with reduced social jetlag, consistent with evidence that children who achieve their nighttime sleep duration through early and consistent bedtimes are less likely to resort to catch‐up sleep, a key driver of social jetlag (Hasegawa et al. 2024). Sleep duration irregularity (∆SD) emerged as a significant negative predictor of social jetlag. The inverse association is consistent with two plausible interpretations that cannot be disentangled in the current cross‐sectional design: extended free‐day sleep may be a consequence of social jetlag through compensatory catch‐up sleep (Roenneberg et al. 2019), or a contributor through more permissive parental free‐day routines (Buxton et al. 2015; Illingworth et al. 2025). The present data do not support causal inference in either direction, and these findings should not be interpreted as recommending free‐day sleep restriction, as this could promote sleep debt.

Child BMI was positively associated with social jetlag, aligning with emerging evidence linking circadian misalignment to adiposity risk in children and adults (Higgins et al. 2021; Roenneberg et al. 2012; Stoner et al. 2018), indicating that child SJL is an earlier developmental risk factor than previously recognized. Child sex was not a significant predictor, differing from some previous literature reporting larger social jetlag in girls (Tétreault et al. 2019). The association of secular family lifestyle with lower social jetlag suggests that socio‐cultural factors may have specific impacts on children's sleep timing regularity across free/school day‐type. A negative association of evening screen exposure with child SJL contradicted prevailing reports of increased behavioural circadian misalignment associated with screen use in the circadian interference window (Hale et al. 2018; Hartstein et al. 2025). This may reflect culture‐specific parental restriction strategies whereby children with greater social jetlag are limited in screen use (Hale et al. 2018), reverse causation, or parental reporting biases. Future studies should account for the context, timing, and content of screen use and parenting practices surrounding it.

Most children demonstrated age‐appropriate sleep duration and high day‐type consistency (mean 10.45 h; 57.1% showing ≤ 30 min day‐type variation), meeting AASM recommendations (Paruthi et al. 2016; Williams et al. 2013). This contrasted with the substantial social jetlag observed, indicating that child SJL was not driven primarily by sleep debt compensation.

4.2. Familial Influences on Sleep Timing Patterns

Although the prevalence studies of social jetlag in young children are already robust and rapidly accumulating (Doi et al. 2015; Giannoumis et al. 2022; Higgins et al. 2021; Stoner et al. 2018; Wong et al. 2022), the comparative extent of child SJL relative to maternal SJL has not been systematically examined. Mother–child correlations in chronotype and SJL underscored the family‐embedded nature of paediatric sleep behaviours, likely reflecting genetic influences on circadian timing and environmental factors including light exposure, parental modelling, house routines, and culture (Buxton et al. 2015; Fisher et al. 2012; Higuchi et al. 2014; Toomey et al. 2015; Wong et al. 2022; Zimmermann 2016).

Children consistently experienced greater SJL than their mothers (Figure 3), suggesting differences in the mechanisms underlying paediatric and adult social jetlag. In adults, social jetlag typically reflects compensatory sleep extension combined with a later‐than‐social biological clock (Roenneberg et al. 2019). In contrast, children's social jetlag reflects a more complex interplay of developmental, biological and family‐driven factors (Buxton et al. 2015; Zimmermann 2016). The increase in social jetlag with age, from 54.5 min at age 4 to 107.1 min at age 10 and the invariable early wake times across age groups on school days, combined with progressive delays in both sleep onset and offset times on free days, supported the classical bio‐developmental perspective where children's naturally delaying chronotype increasingly conflicts with school‐imposed social schedules (Boatswain‐Jacques et al. 2023; Illingworth et al. 2025). Additionally, relaxed parental routines on free days likely contribute (Buxton et al. 2015). These mechanisms are not mutually exclusive and likely operate simultaneously, mirroring the perfect storm model proposed for adolescence (Carskadon 2011; Crowley et al. 2018). The family system presumably serves as both a constraint (through day‐type dependent routines) and an enabler (through free‐day flexibility) of children's circadian traits, where parents and children influence each other's sleep patterns. Model 2 provided empirical support for this familial perspective: maternal social jetlag and older maternal age independently predicted child social jetlag, whereas maternal chronotype, sleep duration, and BMI were not significant, suggesting familial transmission operates through specific behavioural pathways. These findings support the necessity of family‐level interventions (Wong et al. 2022), regardless of whether the underlying mechanism is primarily biological or contextual.

4.3. Limitations

Several limitations must be considered when interpreting current findings. First, the cross‐sectional survey study design precludes causal inferences about relationships among variables. Second, sleep parameters were assessed using maternal report rather than self‐report or objective measures. Although the Munich Chronotype Questionnaire has demonstrated validity in adult populations (Ghotbi et al. 2020), its application to paediatric populations through parental proxy reporting introduces potential measurement errors and biases. Third, the sample was predominantly composed of Jewish Israeli mothers (96.4%) with religious lifestyles (67.1%), limiting generalizability to other cultural and socioeconomic contexts. Additionally, the request for mothers to report on their youngest child in the 4–10‐year age range resulted in uneven age distribution, with fewer participants in older age groups. Although the use of non‐random (convenience) sampling limits the generalizability of our findings, the large and diverse sample, including families from various ethnic, secular, and religious backgrounds, as well as both working and non‐working mothers, broadens the range of perspectives represented. Importantly, the magnitude and developmental trajectory of social jetlag observed in our sample are broadly consistent with other large international paediatric cohorts reporting mean SJL values of 50–70 min in preschool and school‐age children (Choubai et al. 2025; Giannoumis et al. 2022; Higgins et al. 2021; Illingworth et al. 2025; Stoner et al. 2018). Finally, the study did not assess other significant factors influencing children's sleep patterns, including second parents and siblings, school start times, caffeine consumption, or clinical sleep disorders.

5. Conclusions, Clinical Implications and Future Research

Social jetlag is prevalent in preadolescent children and increases substantially with age. The divergence between adequate sleep duration and high social jetlag suggests current paediatric sleep guidelines, focused primarily on quantity, may be insufficient. Chronotype‐driven misalignment may exist even in children with sufficient sleep duration. Children's higher SJL relative to mothers highlights the underrecognized public health significance of paediatric behavioural circadian misalignment. Effective interventions should simultaneously address bio‐developmental and family‐system factors, accommodating delaying chronotypes while targeting parental routines and limit‐setting. Incorporating multiple informant reports may improve research reliability (Liu et al. 2018). Future studies should examine the longitudinal health and functional consequences of early social jetlag.

Author Contributions

The conception and design of the study were done by M.K. and L.H.‐H. Acquisition of data was done by M.K., L.H.‐H., and L.M.E., V.T. and M.K. did the principal data analysis and data visualization. Drafting the manuscript was done by M.K. M.K., V.T., L.M.E. and L.H.‐H. contributed to revising the article, and approved the final version as submitted.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

We included all the data needed for the evaluation of the conclusions in the results section. Additional data related to this article may be requested from the corresponding author. The current ethics approval does not include public repository deposition of the dataset.