Advertisement
Research Article| Volume 6, ISSUE 6, P758-766, December 2020

Download started.

Ok

Associations of time spent on homework or studying with nocturnal sleep behavior and depression symptoms in adolescents from Singapore

  • Sing Chen Yeo
    Affiliations
    Center for Cognitive Neuroscience, Program in Neuroscience and Behavioral Disorders, Duke-NUS Medical School, Singapore
    Search for articles by this author
  • Jacinda Tan
    Affiliations
    Center for Cognitive Neuroscience, Program in Neuroscience and Behavioral Disorders, Duke-NUS Medical School, Singapore
    Search for articles by this author
  • June C. Lo
    Affiliations
    Center for Cognitive Neuroscience, Program in Neuroscience and Behavioral Disorders, Duke-NUS Medical School, Singapore

    Center for Sleep and Cognition, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
    Search for articles by this author
  • Michael W.L. Chee
    Affiliations
    Center for Cognitive Neuroscience, Program in Neuroscience and Behavioral Disorders, Duke-NUS Medical School, Singapore

    Center for Sleep and Cognition, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
    Search for articles by this author
  • Joshua J. Gooley
    Correspondence
    Corresponding author: Joshua J. Gooley, Center for Cognitive Neuroscience, Neuroscience and Behavioral Disorders Program, Duke-NUS Medical School Singapore, 8 College Road, Singapore 117549, Singapore
    Affiliations
    Center for Cognitive Neuroscience, Program in Neuroscience and Behavioral Disorders, Duke-NUS Medical School, Singapore
    Search for articles by this author
Open AccessPublished:June 12, 2020DOI:https://doi.org/10.1016/j.sleh.2020.04.011

      Abstract

      Objective

      To investigate associations of adolescents’ time spent on homework/studying with nocturnal time for sleep and depression symptoms, in a competitive academic environment.

      Design

      Cross-sectional, anonymous survey of sleep habits, school life, and health-related measures.

      Setting

      Eight schools in Singapore.

      Participants

      Total 1225 adolescents aged 13–19 years.

      Measurements

      Self-reported sleep behavior and time use data were collected separately for school days and weekends. Multiple regression models were used to test covariation of time spent on homework/studying with other activities, and associations of homework/studying duration with depression symptoms.

      Results

      Time in bed for sleep and media use were inversely related with homework/studying duration on both school days and weekends, adjusting for time spent on other activities and demographic variables. Face-to-face family time and hanging out with friends were also reciprocally related with homework/studying duration on weekends. Depression scores were higher in adolescents who spent long hours on homework/studying. On school days, this was mediated by reduced time in bed for sleep. On weekends, homework/studying duration associated with depression symptoms, adjusting for time in bed and other covariates. Adolescents who spent ≥5 hours on homework/studying per day on weekends had greater symptoms of anhedonia and anxiety.

      Conclusions

      In a competitive academic setting, adolescents who spent more time on homework/studying spent less time on sleep, media use, and social activities. Independent of effects on sleep, long hours on homework/studying on weekends may be a risk factor for depression. Reducing adolescents’ workload outside of class may benefit their sleep, schoolwork-life balance, and mental well-being.

      Keywords

      Introduction

      Many adolescents have difficulty achieving adequate sleep. This is a major concern because short sleep is associated with reduced cognitive performance and academic achievement,
      • Dewald J.F.
      • Meijer A.M.
      • Oort F.J.
      • Kerkhof G.A.
      • Bogels S.M
      The influence of sleep quality, sleep duration and sleepiness on school performance in children and adolescents: a meta-analytic review.
      ,
      • Lo J.C.
      • Ong J.L.
      • Leong R.L.
      • Gooley J.J.
      • Chee M.W
      Cognitive performance, sleepiness, and mood in partially sleep deprived adolescents: the need for sleep study.
      as well as poorer physical and mental health.
      • Chaput J.P.
      • Gray C.E.
      • Poitras V.J.
      • et al.
      Systematic review of the relationships between sleep duration and health indicators in school-aged children and youth.
      There is a convergence of biological and sociocultural factors that can give rise to inadequate sleep in adolescents.
      • Crowley S.J.
      • Wolfson A.R.
      • Tarokh L.
      • Carskadon M.A
      An update on adolescent sleep: new evidence informing the perfect storm model.
      A delay in circadian rhythms,
      • Roenneberg T.
      • Kuehnle T.
      • Pramstaller P.P.
      • et al.
      A marker for the end of adolescence.
      slower build-up of homeostatic sleep pressure,
      • Jenni O.G.
      • Achermann P.
      • Carskadon M.A
      Homeostatic sleep regulation in adolescents.
      and reduced parental supervision of adolescents’ sleep behavior
      • Short M.A.
      • Gradisar M.
      • Wright H.
      • Lack L.C.
      • Dohnt H.
      • Carskadon M.A
      Time for bed: parent-set bedtimes associated with improved sleep and daytime functioning in adolescents.
      ,
      • Yeo S.C.
      • Jos A.M.
      • Erwin C.
      • et al.
      Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
      contribute to later bedtimes and reduced sleep. Early school start times also result in adolescents waking up earlier than desired, further curtailing their nocturnal sleep.
      • Watson N.F.
      • Martin J.L.
      • Wise M.S.
      • et al.
      Delaying middle school and high school start times promotes student health and performance: an American Academy of Sleep Medicine Position Statement.
      Additionally, social obligations, extracurricular activities, and school workload often intensify across adolescence and compete for students’ time. Consequently, many adolescents do not have enough time for sleep.
      • Wang S.M.
      • Zou J.L.
      • Gifford M.
      • Dalal K
      Young students' knowledge and perception of health and fitness: a study in Shanghai, China.
      Adolescents often cite homework as a barrier to getting enough sleep on school nights. While homework may positively associate with school achievement,
      • Cooper H.
      • Robinson J.C.
      • Patall E.A
      Does homework improve academic achievement? A synthesis of research, 1987–2003.
      too much homework can negatively influence students’ attitudes toward school and displace time spent on leisure, exercise/sports, extracurricular activities, and sleep. Previous studies have shown that adolescents who spend greater time on homework, private tutoring, and preparation for school entrance exams have later bedtimes and shorter nocturnal sleep.
      • Yoo C
      Sleep duration change and its associated factors during adolescence: a 6 year longitudinal study.
      • Street N.W.
      • McCormick M.C.
      • Austin S.B.
      • Slopen N.
      • Habre R.
      • Molnar B.E
      Examining family and neighborhood level predictors of sleep duration in urban youth.
      • Adam E.K.
      • Snell E.K.
      • Pendry P
      Sleep timing and quantity in ecological and family context: a nationally representative time-diary study.
      • Yang C.K.
      • Kim J.K.
      • Patel S.R.
      • Lee J.H
      Age-related changes in sleep/wake patterns among Korean teenagers.
      • Twenge J.M.
      • Krizan Z.
      • Hisler G
      Decreases in self-reported sleep duration among U.S. adolescents 2009–2015 and association with new media screen time.
      Adolescents who spend more time on homework also experience more school-related stress, and they are more likely to drop activities or hobbies in order to make way for schoolwork.
      • Galloway M.
      • Conner J.
      • Pope D
      Nonacademic effects of homework in privileged, high-performing high schools.
      ,
      • Galloway M.
      • Pope D
      Hazardous homework? The relationship between homework, goal orientation, and well-being in adolescence.
      The exchange of sleep and other positive activities for academic pursuits may be especially problematic in hard-driving Confucian-heritage cultures (e.g., Korea, Japan, China, Hong Kong, Taiwan, and Singapore), where working harder and longer is often viewed as necessary for achieving academic goals.
      • Ong X.L.
      • Cheung H.S
      Schools and the Class Divide: An Examination of Children's Self-Concept and Aspirations in Singapore.
      • Huang G.H.-.C.
      • Gove M
      Confucianism, Chinese families, and academic achievement: exploring how Confucianism and Asian descendant parenting practices influence children's academic achievement.
      • Stankov L
      Unforgiving Confucian culture: a breeding ground for high academic achievement, test anxiety and self-doubt?.
      In these highly competitive settings, there is strong emphasis on academic achievement and performance on high-stakes examinations that impact students’ future schooling and career options. These pressures may contribute to why adolescents in Asian countries go to bed later and sleep less than adolescents in Western cultures.
      • Yeo S.C.
      • Jos A.M.
      • Erwin C.
      • et al.
      Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
      ,
      Organisation for Economic Co-operation and Development
      PISA 2015 Results: Students' Well-Being (Volume III).
      The relative impact of homework/studying on sleep behavior and other activities in adolescents is still poorly understood, especially in cultures in which students have a heavy workload outside of school. While students may feel that they do not have enough time for sleep, there is little evidence on how adolescents allocate their time. Such information is important for developing behavioral interventions for improving adolescents’ sleep. Prior research has largely focused on the relationship between homework/studying and sleep behavior on school days, without considering how students’ workload on weekends may influence their sleep and mood. The present study aimed to address these gaps in knowledge by examining how students in a highly-competitive academic setting allocate their time on school days and weekends, and the degree to which homework/studying may displace sleep and affect mood. Our first objective was to evaluate students’ time use on school days and weekends. We hypothesized that students would extend their homework/studying duration and time for sleep on weekends. Our second objective was to assess the degree to which adolescents may reduce nocturnal sleep and other healthy activities (e.g., exercise and family time) for time spent on homework/studying. We hypothesized that homework/studying duration would be inversely associated with time in bed for sleep, and this association would be stronger on school days because adolescents have less time for sleep compared with weekends. Our third objective was to test whether increased time spent on homework/studying associated with greater severity of depression symptoms. We hypothesized that longer homework/studying durations on either school days or weekends would be associated with higher depression scores.

      Participants and methods

      Participants and data collection

      Adolescents aged 13–19 years from 8 schools in Singapore were recruited to take part in a cross-sectional, anonymous survey of sleep habits, school life, and health. School characteristics and recruitment procedures are described in detail in our previous work.
      • Yeo S.C.
      • Jos A.M.
      • Erwin C.
      • et al.
      Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
      In short, 74 schools were invited and 8 schools agreed to take part in the research (5 local schools and 3 international schools). The participating local schools ranked among the top 10% of all secondary schools in Singapore based on their cut-off national examination scores, and are considered elite in the Singapore context. The participating international schools follow overseas curricula and predominantly serve the expatriate community. Schools that declined to participate were either not interested or indicated that they did not have time. Recruitment of student participants was managed internally by the schools without direct involvement of the researchers. Recruitment materials (study description and permission forms) were provided by the researchers to school representatives, who distributed study information to students and informed parents about the study by email. Adolescents who obtained written parental permission were eligible to take part in the study. These students were invited to attend a one-time session during their morning assembly to complete the survey by pen and paper under supervision of the research team, with the exception of one school that completed an online version of the survey.
      • Lo J.C.
      • Lee S.M.
      • Lee X.K.
      • et al.
      Sustained benefits of delaying school start time on adolescent sleep and well-being.
      The survey took place at different times of the year across schools but was scheduled to avoid major examination periods.
      There were 2386 students who provided consent to take part in the research. Participants were excluded from the analysis if they were not 13–19 years old (n = 22), or if they had missing data (n = 592) for any survey question on bedtime, wake-up time, or time use on either school days or weekends. In participants with complete data, histograms were inspected for all variables in order to set thresholds for responses that were considered out of range. Participants who provided a response that was out of range for any time use variable were excluded from all analyses (n = 547). Thresholds for excluding time use data on school days were >6 hours on transportation, exercise/sports, co-curricular activities, extracurricular activities, part-time job, hanging out with friends, or face-to-face family time, or >12 hours on lessons/lectures/labs, homework/studying, or media use. Thresholds for excluding data on weekend days were >6 hours on transportation, lessons/lectures/labs, exercise/sports, co-curricular activities, extracurricular activities, or part-time job, >12 hours on hanging out with friends, or >16 hours on homework/studying, media use, or face-to-face family time. Based on Chi-square (χ2) tests and independent samples t-tests, adolescents with missing/ invalid data were similar to participants who were included in the analysis with respect to sex (44.7% versus 46.8% boys; χ2=1.1, P = 0.30), but were slightly younger (Mean ± SD: 15.7 ± 1.7 versus 16.2 ± 1.4 years; t = 7.8, P < 0.001) and less likely to be Chinese (63.3% versus 76.0%; χ2 = 71.9, P < 0.001). The final sample used in the present analysis included 1225 adolescents. Research procedures were approved by the Ministry of Education, Singapore, and the National University of Singapore Institutional Review Board.

      Assessment of sleep behavior and time use

      The survey comprised 40 items on sleep behavior, with most items taken or modified from the School Sleep Habits Survey.
      • Wolfson A.R.
      • Carskadon M.A
      Sleep schedules and daytime functioning in adolescents.
      Participants reported their bedtime and wake-up time separately for school days and weekends.
      • Yeo S.C.
      • Jos A.M.
      • Erwin C.
      • et al.
      Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
      Bedtime was assessed with the question “What time do you usually go to bed on school days / weekends?,” and wake-up time was assessed with the question “What time do you usually wake up on school days / weekends?” Time in bed for sleep was calculated as the duration of time elapsed from self-reported bedtime to wake-up time. In previous research that implemented the same questions for assessing sleep behavior in adolescents,
      • Wolfson A.R.
      • Carskadon M.A.
      • Acebo C.
      • et al.
      Evidence for the validity of a sleep habits survey for adolescents.
      time in bed was shown to correlate with an 8-day assessment of total nocturnal sleep using a daily sleep diary (Pearson's correlation analysis: school day, r = 0.61; weekend, r = 0.38), and with the nocturnal sleep period derived from actigraphy (school day, r = 0.53; weekend, r = 0.31), supporting the validity of the School Sleep Habits Survey for estimating sleep behavior.
      Adolescents’ daily time spent on waking activities was assessed using the following question: “On a typical weekday, state the average time you spend on each of the following activities. Put 0 hours if you do not spend time on that particular activity at all. The total number of hours spent on these activities can add up to more or less than 24 hours, and some options might not be mutually exclusive.” Activities included “Transportation,” “On lessons/lectures/lab,” “Doing homework/studying,” “Tuition,” “Exercise/sports,” “Co-curricular activities (e.g., drama club),” “Extra-curricular activities,” “Part-time job,” “Hanging out with friends outside of school,” “Media use (social media/ phone/ internet/ TV),” and “Time interacting face to face with family.” In our analyses, “Doing homework/studying” and “Tuition” (i.e., additional private schooling) were combined into a single category referred to as “homework/studying.” In a separate question, participants were asked to report their daily time spent on the same set of activities on weekends.
      Participants were asked to indicate the main reason for their bedtime, assessed separately for school days and weekends, by choosing from a list of options, including “I set my bedtime,” “My parents set my bedtime,” “I feel tired or sleepy,” “I finish my homework,” “I finish my video games,” “My brother(s) or sister(s) go to bed,” “I finish socializing (including talking/ messaging on the phone),” “I log off from social media platforms (i.e., Facebook, Instagram, Twitter, Snapchat),” “I get home from my job,” and “Other:_____.

      Assessment of depression symptoms

      The 11-item Kutcher Adolescent Depression Scale (KADS) was used to assess depression symptoms over the past week.
      • Brooks S.J.
      • Krulewicz S.P.
      • Kutcher S
      The Kutcher Adolescent Depression Scale: assessment of its evaluative properties over the course of an 8-week pediatric pharmacotherapy trial.
      The KADS was developed for epidemiological studies of adolescent depression and for monitoring treatment effects over time. The KADS has been shown to have good internal consistency (Cronbach's α, range 0.80 to 0.87),
      • Brooks S.J.
      • Krulewicz S.P.
      • Kutcher S
      The Kutcher Adolescent Depression Scale: assessment of its evaluative properties over the course of an 8-week pediatric pharmacotherapy trial.
      ,
      • Zhou H.
      • Hao N.
      • Du Y.
      • et al.
      Reliability and validity of the eleven item Kutcher Adolescent Depression Scale, Chinese Version (KADS-11CV).
      test-retest reliability (Pearson's r = 0.77),
      • Zhou H.
      • Hao N.
      • Du Y.
      • et al.
      Reliability and validity of the eleven item Kutcher Adolescent Depression Scale, Chinese Version (KADS-11CV).
      concurrent validity with the Children's Depression Inventory (Pearson's r = 0.74),
      • Zhou H.
      • Hao N.
      • Du Y.
      • et al.
      Reliability and validity of the eleven item Kutcher Adolescent Depression Scale, Chinese Version (KADS-11CV).
      and is sensitive to clinician-rated changes in depression severity.
      • Brooks S.J.
      • Krulewicz S.P.
      • Kutcher S
      The Kutcher Adolescent Depression Scale: assessment of its evaluative properties over the course of an 8-week pediatric pharmacotherapy trial.
      Individual KADS items included descriptions relating to sadness, irritability, sleep difficulties, apathy, feelings of worthlessness, fatigue/ low motivation, lack of focus, anhedonia, anxiety, physical signs of anxiety, and thoughts of self-harm or suicide. Response options included “Hardly ever,” “Some of the time,” “Most of the time,” and “All the time,” and were assigned a score of 0–3 for each item. The sum of scores for individual items was used as an overall measure of depression symptom severity (i.e., the global KADS score, ranging from 0 to 33). There are no normative data available for the KADS in adolescents in Singapore. A recent study of Chinese adolescents suggested a cut-off KADS score of ≥9 for diagnosing depression (sensitivity, 89%; specificity, 90%),
      • Zhou H.
      • Hao N.
      • Du Y.
      • et al.
      Reliability and validity of the eleven item Kutcher Adolescent Depression Scale, Chinese Version (KADS-11CV).
      but the generalizability of these findings have not been tested in other clinical populations. Hence, we did not attempt to categorize adolescents into different depression severity groups and instead report the global score and results for individual KADS items.
      In item-by-item analyses of depression symptoms, adolescents were categorized as exhibiting high severity of a symptom if they responded “Most of the time” or “All the time.” Among the 1225 adolescents included in analyses of time use (i.e., time in bed and each of the 10 waking activities), there were 40 participants with missing responses on the KADS. These individuals were excluded from analyses of depression symptoms because a global KADS score could not be calculated. Hence, there were 1185 adolescents who were included in analyses of depression symptoms.

      Data analysis and statistics

      Daily time spent on different activities was compared between school days and weekends using paired-samples t-tests (threshold for significant difference, P < 0.05), with effect size measured using Cohen's d. For analyses that examined associations of homework/studying with other behaviors, participants were categorized by the duration of time they spent on homework/studying (<2 hours, 2 to <3 hours, 3 to <4 hours, 4 to <5 hours, and ≥5 hours). These categories were chosen based on the distribution of homework/studying duration in our dataset and ensured that the sample size in each category was >100 adolescents per group on school days (n = 236, n = 377, n = 312, n = 175, n = 125) and weekends (n = 128, n = 173, n = 194, n = 179, n = 551). Bivariate associations of homework/studying duration with nocturnal time in bed for sleep, bedtime reason, and KADS scores were tested using one-way analysis of variance (ANOVA) or χ2 tests, with effect size measured using partial eta squared (ηp2) values or Cramer's V.
      Multiple linear regression models were used to test covariation of homework/studying duration with other activities. Our approach was based on previous analyses of time use data,
      • Basner M.
      • Fomberstein K.M.
      • Razavi F.M.
      • et al.
      American time use survey: sleep time and its relationship to waking activities.
      and aimed to determine how adolescents with short and long homework/studying durations differed in their daily activities compared with adolescents with a typical homework/studying duration on school days. The primary a priori hypothesis was that longer homework/studying durations would be associated with shorter nocturnal time in bed for sleep. However, we also tested associations with all other time use variables in order to assess how time spent on homework/studying may impact students’ daily activities and time management. These analyses were planned but exploratory, i.e. we did not test a specific hypothesis. Collinearity was examined for every pair of time use variables by performing Pearson's correlation analysis and calculating variance inflation factor (VIF) values. There was minimal evidence of collinearity based on the range of Pearson's r values and VIF values for time use variables on school days (r = −0.26 to 0.16; VIF = 1.01–1.29) and on weekends (r = −0.25 to 0.15; VIF = 1.01–1.42).
      In the multiple regression models, homework/studying duration was divided into 5 categories: <2 hours, 2 to <3 hours, 3 to <4 hours, 4 to <5 hours, and ≥5 hours. The 2 to <3 hours homework/studying duration category was chosen as the reference because the highest number of participants fell into this category on school days. For school days and weekends, we constructed 10 different multiple linear regression models, in which homework/studying duration was the independent variable of primary interest, and one of the other 10 daily activities served as the dependent variable. Each model was adjusted for age (in years), sex, ethnicity (Chinese, Indian, Malay, Other), school start time (before 0800, 0800 or later), and school type (local, international), with reference categories of 16 years-old, male, Chinese ethnicity, early school start time, and local school type. In each model, the coefficient for a given homework/studying duration category indicated how much participants' activity time for the dependent variable deviated from that observed in individuals with 2 to <3 hours of time spent on homework/studying, adjusted for each of the other daily activities and demographic variables. Unstandardized coefficients were expressed in minutes and represent the effect size of homework/studying duration on each of the time use variables. The threshold for statistical significance for regression coefficients was adjusted using a Bonferroni correction for multiple comparisons (P < 0.005; α = 0.05/10 models) because each model tested the association between homework/studying duration and 1 of the other 10-time use variables.
      Multiple linear regression was used to test the association between time spent on homework/studying and the global KADS score. The difference in depression scores relative to the reference group (i.e., the effect size) was expressed using both unstandardized (B values) and standardized (β values) regression coefficients. The primary a priori hypothesis was that longer homework/studying durations would be associated with a higher global KADS score. In a secondary analysis, we tested the hypothesis that time spent on homework/studying on school days has an indirect effect on depression score by reducing time in bed for sleep. A mediation analysis was conducted in which homework/studying was the causal variable, depression score was the outcome variable, and time in bed was the mediating variable. The theoretical basis for this model is that day-to-day variation in homework/studying influences the amount of sleep that adolescents obtain on a given night,
      • Fuligni A.J.
      • Hardway C
      Daily variation in adolescents' sleep, activities, and psychological well-being.
      and experimentally reducing sleep in adolescents worsens mood and emotional regulation.
      • Lo J.C.
      • Ong J.L.
      • Leong R.L.
      • Gooley J.J.
      • Chee M.W
      Cognitive performance, sleepiness, and mood in partially sleep deprived adolescents: the need for sleep study.
      ,
      • Baum K.T.
      • Desai A.
      • Field J.
      • Miller L.E.
      • Rausch J.
      • Beebe D.W
      Sleep restriction worsens mood and emotion regulation in adolescents.
      In the mediation model,
      • Preacher K.J.
      • Hayes A.F
      Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models.
      multiple regression analyses were performed to test correlations between homework/studying duration and time in bed for sleep (path a), time in bed for sleep and depression score (path b), and homework/studying duration and depression score (path c), adjusting for demographic variables. Path c (the total effect) represents the sum of the direct effect (path c’) of homework/studying duration on depression score and the indirect effect (path ab) mediated through time in bed for sleep. The significance of the indirect effect was tested using bootstrapping procedures.
      • Preacher K.J.
      • Hayes A.F
      Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models.
      Unstandardized indirect effects were computed for 5000 bootstrapped samples and the 95% confidence interval (CI) was determined. A nonzero indirect effect was considered statistically significant.
      In follow-up exploratory analyses (i.e., post hoc tests), we sought to determine which depression symptoms were higher among students who spent long hours on homework/studying. Multivariable logistic regression was used to model associations of homework/studying duration with individual depression symptoms (i.e., each of the 11 items on the KADS) reported most or all the time in the past week. The odds ratio (OR) and adjusted OR (ORadj) for each depression symptom were reported as a measure of effect size, with the threshold for statistical significance Bonferroni-corrected to P < 0.0045 (α = 0.05/11 models) to account for multiple comparisons between models. Models were adjusted for time spent on each of the other daily activities, age, sex, ethnicity, school start time, and school type. Statistical tests were carried out using SPSS software, version 25.0 (IBM Corporation).

      Results

      On school days, adolescents spent an average of about 6.5 hours each on nocturnal time in bed for sleep and classroom lessons, 3 hours for homework/studying, 2 hours for media use, and approximately 1 hour each for face-to-face family time, transportation, and co-curricular (school-based) activities (Table 1). As expected, students’ time use differed substantially on weekends, when little or no time was spent in school on classroom or co-curricular activities. Adolescents increased their nocturnal time in bed for sleep on weekends to about 9 hours, and also spent substantially more time on homework, media use, hanging out with friends, and face-to-face family time compared with school days (Table 1). Time spent on transportation, exercise/sports, extra-curricular activities, and part-time jobs showed little or no difference on school days versus weekends.
      Table 1Adolescents’ daily time spent on different activities.
      Time spent on activities (h)
      Daily activitiesSchool daysWeekendstPCohen's d
      Time in bed for sleep6.57 ± 1.238.93 ± 1.49−49.0<0.001−1.73
      Lessons/lectures/lab6.46 ± 1.110.07 ± 0.39194.9<0.0017.68
      Homework/studying2.87 ± 1.464.47 ± 2.45−30.0<0.001−0.79
      Media use2.06 ± 1.273.49 ± 2.09−32.4<0.001−0.83
      Transportation1.28 ± 0.650.98 ± 0.7411.4<0.0010.43
      Co-curricular activities1.22 ± 1.170.22 ± 0.6928.4<0.0011.04
      Family time, face-to-face1.23 ± 0.922.70 ± 1.95−32.5<0.001−0.97
      Exercise/sports0.86 ± 0.860.91 ± 0.97−2.20.031−0.06
      Hanging out with friends0.59 ± 0.771.24 ± 1.59−15.2<0.001−0.52
      Extracurricular activities0.32 ± 0.650.36 ± 0.88−1.90.057−0.06
      Part-time job0.01 ± 0.130.03 ± 0.22−2.40.014−0.08
      The daily duration of time spent on different activities is shown for 1225 adolescents. Time use is expressed in hours (mean ± SD). A student's paired t test was used to test for differences in activity time on school days and weekends.
      Taking into consideration time spent on classroom lessons, homework/studying, and co-curricular activities, adolescents in our study spent about 60 hours per week on schoolwork and school-related activities. The amount of time that students spent on homework/studying increased with age and was higher among girls on both school days and weekends (Table S1; Fig. S1). Adolescents who spent more time on homework/studying were also more likely to be Chinese and attended local schools (Table S1). By comparison, time in bed for sleep decreased in older adolescents and was about the same in boys and girls. Adolescents who attended international schools had greater time in bed compared with students in local schools, and spent less time on homework/studying on weekends (Fig. S1).
      Next, we tested covariation of homework/studying duration with time in bed for sleep and various waking activities. On school days, nocturnal time in bed decreased with greater time spent on homework/studying (Fig. 1A). Based on multiple linear regression analyses, time in bed for sleep and media use were the activities that showed the strongest inverse associations with homework/studying time (Fig. 1B; Table S2). Adolescents who spent long hours on homework/studying (≥5 hours) showed a reduction of about 30–35 minutes on their time in bed and media use relative to the reference group (2 to <3 hours), whereas those students with <2 hours of homework/studying per day had about 15–20 minutes more time for sleep and media use. Young adolescents (13–14 years) spent about an hour more on time in bed for sleep compared with older adolescents on school days, adjusting for other variables (Table S2).
      Fig 1
      Fig. 1Covariation of homework/studying duration with time in bed for sleep and different waking activities. (A) The duration of nocturnal time in bed is shown for adolescents grouped by their daily time spent on homework/studying on school days. (B) The average change in school day activity time is shown for different homework/studying durations, relative to the reference category. (C) Time in bed for sleep is shown for adolescents grouped by their homework/studying duration on weekends. (D) The average change in weekend activity time is shown relative to the reference category. In panels A and C, the mean ± 95% CIs are shown. In panels B and D, results are based on multiple linear regression, with each model adjusted for time spent on each of the other daily activities, age, sex, ethnicity, school start time, and school type. Asterisks indicate significant changes in activity time relative to the reference category (P < 0.005). Results are shown only for daily activities with at least a 30-minute change relative to the reference on either school days or weekends (See Table S2 for results of all time use and demographic variables).
      On weekends, nocturnal time in bed for sleep also decreased with greater time spent on homework/studying (Fig. 1C). Multiple linear regression analyses showed that weekend activities inversely associated with homework/studying time were nocturnal time in bed, media use, face-to-face family time, and hanging out with friends (Fig. 1D; Table S2). Adolescents who spent ≥5 hours on homework/studying per day spent about 40–60 minutes less on time in bed, family time, and hanging out with friends compared with those who spent 2 to <3 hours on homework/studying. Adolescents with long hours on studying/homework also spent about 90 minutes less on media use relative to the reference group.
      Next, we examined whether homework affected adolescents’ choice of bedtime. The most common reason for going to bed on school days was “I finish my homework” (Fig. 2A), which was reported more frequently in adolescents with long homework/studying durations (Fig. 2B). Adolescents whose reason for going to bed was “I finish my homework” had a bedtime that was about 30 minutes later than individuals who went to bed for any other reason (Mean ± SD: 00:03 ± 00:05 versus 23:29 ± 00:07; t = −7.6, P < 0.001), resulting in shorter nocturnal time in bed for sleep (Mean ± SD: 6.3 ± 1.2 versus 6.8 ± 1.2; t = −6.9, P < 0.001; Fig. 2C).
      Fig 2
      Fig. 2Relationships between homework/studying and bedtime. (A) Adolescents’ main reason for going to bed on school days is shown in the frequency plot. (B) The percentage of participants whose main reason for going to bed was “I finish my homework” is shown for adolescents grouped by their daily time spent on homework/studying on school days. (C) The time-in-bed interval (TIB) for sleep (gray bars) on school days is shown for adolescents whose main reason for going to bed was finishing homework versus any other reason. The mean and 95% CIs for bedtime and wake-up time are shown for each TIB interval. (D) Adolescents’ main reason for going to bed is shown on weekends. (E) The percentage of adolescents whose bedtime reason on weekends was finishing homework is shown for different homework/studying duration categories. (F) TIB is shown on weekends for students whose main reason for going to bed was finishing homework versus any other reason.
      On weekends, the most common reason for going to bed was “I feel tired or sleepy” (Fig. 2D). However, adolescents with ≥5 hours of homework/studying on weekends were more likely to cite “I finish my homework” as the main reason for their bedtime, as compared with participants with shorter homework/studying durations (Fig. 2E). In contrast to school days, there was no difference in bedtime (Mean ± SD: 00:16 ± 00:06 versus 00:24 ± 00:08; t = −1.5, P = 0.12) or nocturnal time in bed for sleep (Mean ± SD: 8.9 ± 1.5 versus 8.9 ± 1.5; t = 0.7, P = 0.48) in adolescents whose main reason for going to bed was “I finish my homework” versus any other reason (Fig. 2F).
      In bivariate analyses, global depression scores were higher in adolescents who spent more time on homework/studying on school days (Fig. 3A). In multiple regression models that adjusted for time spent on other activities (i.e., time in bed and each waking activity) and demographic variables, however, adolescents with long or short homework/studying durations on school days did not differ from the reference group (2 to <3 hours) in their global depression scores (Fig. 3B; Table S3) or individual symptoms on the KADS (Table S4). By comparison, depression scores were higher among girls, and in adolescents who spent less time in bed for sleep, less time on exercise/sports, or more time on media use on school days (Table S3). To explore the possibility that the bivariate association between homework/studying duration and global depression score was mediated by a reduction in time in bed, we repeated the multiple regression analysis with time in bed removed from the model. Doing so resulted in a significant association between long hours spent on homework/studying and depression score (≥5 hours versus 2 to <3 hours; B = +1.50, P < 0.005; Fig. 3B), whereas removing any of the other time use variables did not alter the results compared with the fully adjusted model (B ≤ 0.90 and P ≥ 0.08 for all comparisons). Mediation analysis was also performed (Fig. S2) to test whether time in bed for sleep mediated effects of homework/studying duration on depression score (unstandardized coefficients: path a, B=−0.15, P < 0.001; path b, B = −1.41, P < 0.001; path c, B = 0.36, P < 0.001; path c’, B = 0.15, P = 0.14). We found a significant indirect effect of homework/studying duration on depression score that was mediated by reduced time in bed for sleep (path ab indirect effect = 0.21, 95% CI 0.13–0.29).
      Fig 3
      Fig. 3Associations of homework/studying duration with depression score. (A) Global depression score on the Kutcher Adolescent Depression Scale (KADS) is shown for adolescents grouped by their daily time spent on homework/studying on school days. (B) The average change in KADS score for different homework/studying durations on school days is shown relative to the reference. (C) KADS scores are shown for different homework/studying durations on weekends. (D) The average change in KADS score for different homework/studying durations on weekends is shown relative to the reference. In panels A and C, the mean ± 95% CIs are shown. In panels B and D, results are based on multiple linear regression with each model adjusted for time spent on daily activities (time in bed for sleep, lessons/lectures/lab, media use, transportation, co-curricular activities, face-to-face family time, exercise/sports, hanging out with friends, extracurricular activities, part-time job), age, sex, ethnicity, school start time, and school type. See Table S3 for regression coefficients of all variables entered into the models. In panels B and D, the change in KADS scores is also shown when time in bed (TIB) for sleep is removed from the multiple regression model, showing that TIB influences the association of homework/studying duration on depression scores on school days, but not on weekends.
      Similar to results for school days, bivariate analyses showed that depression scores were higher among adolescents who spent long hours on homework/studying on weekends (Fig. 3C), and the association was most evident for very long homework/studying durations (i.e., extending to ≥7 hours; Fig. S3). In the multiple linear regression model that adjusted for time spent on other activities and demographic variables, increased time spent on homework/studying was associated with higher depression scores (Fig. 3D; Table S3). Depression scores were also higher in girls and associated with time spent on media use, but not with time in bed for sleep (Table S3). In multiple logistic regression analyses, adolescents with ≥5 hours of homework/studying on weekends had higher odds of feeling anhedonia (ORadj = 2.52, 95% CI 1.35–4.71, P = 0.004), and anxiety (ORadj = 2.72, 95% CI 1.53–4.83, P = 0.001) compared with the reference group, adjusting for time use and demographic variables (Table S4).

      Discussion

      The present study demonstrates that time spent on homework/studying may be a significant barrier to sleep in adolescents in a society with strong emphasis on academic achievement. Homework/studying duration was inversely associated with time in bed for sleep and media use on school days, adjusting for time spent on other activities. Additionally, going to bed later was often explained by having to first finish homework, suggesting that homework/studying has a major influence on how students choose their bedtime. On weekends, adolescents with long homework/studying durations (≥5 hours per day) spent less time on sleep, media use, and social activities, and had more depression symptoms compared with individuals with less time spent on homework/studying. Our results suggest that moderating adolescents’ workload outside of school may improve their time in bed for sleep. Additionally, reducing long hours spent on homework/studying, and associated drivers of this behavior, may confer benefits on students’ well-being.
      Our findings are consistent with previous work in school-aged children and adolescents showing that more time spent on homework associated with later bedtimes and less nocturnal sleep.
      • Yoo C
      Sleep duration change and its associated factors during adolescence: a 6 year longitudinal study.
      • Street N.W.
      • McCormick M.C.
      • Austin S.B.
      • Slopen N.
      • Habre R.
      • Molnar B.E
      Examining family and neighborhood level predictors of sleep duration in urban youth.
      • Adam E.K.
      • Snell E.K.
      • Pendry P
      Sleep timing and quantity in ecological and family context: a nationally representative time-diary study.
      ,
      • Twenge J.M.
      • Krizan Z.
      • Hisler G
      Decreases in self-reported sleep duration among U.S. adolescents 2009–2015 and association with new media screen time.
      ,
      • Fuligni A.J.
      • Hardway C
      Daily variation in adolescents' sleep, activities, and psychological well-being.
      ,
      • Sun W.Q.
      • Spruyt K.
      • Chen W.J.
      • et al.
      The relation among sleep duration, homework burden, and sleep hygiene in Chinese school-aged children.
      • Li S.
      • Yang Q.
      • Chen Z.
      • Jin X.
      • Jiang F.
      • Shen X
      Homework schedule: an important factor associated with shorter sleep duration among Chinese school-aged children.
      • Jiang X.
      • Hardy L.L.
      • Baur L.A.
      • Ding D.
      • Wang L.
      • Shi H
      Sleep duration, schedule and quality among urban Chinese children and adolescents: associations with routine after-school activities.
      The impact of homework/studying on sleep behavior may be greatest in cultures where more time is spent on homework, exam preparation, and private after-school classes. This is exemplified by Confucian-heritage societies such as China and Singapore, where adolescents spend more time on homework than their counterparts in most Western societies.
      Organisation for Economic Co-operation and Development
      PISA 2012 Results: What Makes Schools Successful (Volume IV).
      A typical adolescent in our study spent more than 20 hours per week on homework/studying, which is comparable to the amount of time spent on homework, tutoring, and after-school classes in Chinese (Shanghai) adolescents, and 2–3 times greater than the amount in most American, European, and Australasian countries/cities.
      Organisation for Economic Co-operation and Development
      PISA 2012 Results: What Makes Schools Successful (Volume IV).
      Even among high-performing schools in Western cultures, however, it is common for adolescents to spend more than 3 hours on schoolwork per night, and this behavior is associated with greater academic stress and less sleep.
      • Galloway M.
      • Conner J.
      • Pope D
      Nonacademic effects of homework in privileged, high-performing high schools.
      ,
      • Galloway M.
      • Pope D
      Hazardous homework? The relationship between homework, goal orientation, and well-being in adolescence.
      We extended prior work in competitive academic settings by demonstrating that adolescents with long homework/studying durations on weekends showed reduced time in bed for sleep and increased feelings of anxiety and anhedonia, despite extending their time in bed relative to school nights.
      We found that nearly one third of adolescents indicated their main reason for going to bed on school days was finishing their homework. While this result may have been influenced by the set of predetermined response options presented to participants (i.e., the question was not formulated as a free response), our findings are consistent with a previous study in Taiwanese junior high schools.
      • Gau S.F.
      • Soong W.T
      Sleep problems of junior high school students in Taipei.
      By comparison, in predominantly Caucasian countries (United States, Australia, and New Zealand) the percentage of adolescents whose main reason for going to bed is finishing homework was much lower, in the 7–13% range.
      • Short M.A.
      • Gradisar M.
      • Lack L.C.
      • et al.
      A cross-cultural comparison of sleep duration between US And Australian adolescents: the effect of school start time, parent-set bedtimes, and extracurricular load.
      • Foley L.S.
      • Maddison R.
      • Jiang Y.
      • Marsh S.
      • Olds T.
      • Ridley K
      Presleep activities and time of sleep onset in children.
      • Carskadon M.A
      Factors influencing sleep patterns of adolescents.
      Even within the same high schools, Asian-Australian students had later bedtimes, shorter sleep duration, and were more likely to go to bed after finishing homework compared with Caucasian-Australian students.
      • Lushington K.
      • Wilson A.
      • Biggs S.
      • Dollman J.
      • Martin J.
      • Kennedy D
      Culture, extracurricular activity, sleep habits, and mental health: a comparison of senior high school Asian-Australian and Caucasian-Australian adolescents.
      These results point toward sociocultural differences in practices before bedtime that influence sleep duration on school nights.
      While previous studies have shown that digital media use and screen time are inversely correlated with nocturnal sleep,
      • Twenge J.M.
      • Krizan Z.
      • Hisler G
      Decreases in self-reported sleep duration among U.S. adolescents 2009–2015 and association with new media screen time.
      ,
      • Hysing M.
      • Pallesen S.
      • Stormark K.M.
      • Jakobsen R.
      • Lundervold A.J.
      • Sivertsen B
      Sleep and use of electronic devices in adolescence: results from a large population-based study.
      adolescents in our study spent more time on homework/studying on both school day and weekends, and they were more likely to go to bed on school days after finishing homework compared with media use. Additionally, time spent on media use was reciprocally related with homework/studying duration on both school days and weekends, suggesting that adolescents may sacrifice screen time to pursue academic goals. These findings are consistent with previous research in Asian students, whose sleep duration associated more strongly with homework time compared with media use or sleep hygiene practices.
      • Yoo C
      Sleep duration change and its associated factors during adolescence: a 6 year longitudinal study.
      ,
      • Sun W.Q.
      • Spruyt K.
      • Chen W.J.
      • et al.
      The relation among sleep duration, homework burden, and sleep hygiene in Chinese school-aged children.
      ,
      • Jiang X.
      • Hardy L.L.
      • Baur L.A.
      • Ding D.
      • Wang L.
      • Shi H
      Sleep duration, schedule and quality among urban Chinese children and adolescents: associations with routine after-school activities.
      ,
      • Ren H.
      • Zhou Z.
      • Liu W.K.
      • Wang X.
      • Yin Z
      Excessive homework, inadequate sleep, physical inactivity and screen viewing time are major contributors to high paediatric obesity.
      It should be highlighted, however, that increased time spent on either homework/studying or media use on weekends was associated with higher depression scores. Additional research is needed to understand how the balance of time spent on these activities may influence students’ mood. In highly competitive academic settings, efforts to improve adolescents’ time in bed for sleep should target homework/studying behavior, in addition to media use and sleep hygiene.
      We found that homework/studying duration on weekends associated with adolescents’ depression symptoms, adjusting for time spent on other activities. The difference in depression scores for students who spent long hours on homework/studying was similar in magnitude to the difference found for girls versus boys and the association with media use (Table S3). Given the well-established associations of sex and sedentary screen time with depression symptoms,
      • Boers E.
      • Afzali M.H.
      • Newton N.
      • Conrod P
      Association of screen time and depression in adolescence.
      ,
      • Salk R.H.
      • Hyde J.S.
      • Abramson L.Y
      Gender differences in depression in representative national samples: meta-analyses of diagnoses and symptoms.
      our results raise the possibility that homework/studying duration on weekends may also be an indicator of depression risk; however, this hypothesis needs to be tested in other study populations including clinical samples. The difference in depression scores across homework/studying categories was also greater than the degree of improvement in depression scores that we found in a school that delayed its start time by 45 minutes,
      • Lo J.C.
      • Lee S.M.
      • Lee X.K.
      • et al.
      Sustained benefits of delaying school start time on adolescent sleep and well-being.
      suggesting that effect sizes reported in the present study are meaningful. We did not have access to data on clinical manifestations of depression in our sample, however, and normative data are currently not available for the KADS or other depression instruments for adolescents in Singapore.
      We provided evidence that long hours on homework/studying on school days may contribute indirectly to poorer mood by reducing nocturnal sleep duration.
      • Yeo S.C.
      • Jos A.M.
      • Erwin C.
      • et al.
      Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
      Consistent with this interpretation, homework/studying duration was associated with global depression score after removing time in bed from the multiple regression model, but not after removing any of the other time use variables. Additionally, a mediation analysis suggested that time spent on homework/studying had an indirect effect on depression score by reducing time in bed for sleep. On weekends, adolescents attempted to recover their sleep by spending more time in bed. However, they also increased substantially their time on homework/studying and continued to show an inverse association with time in bed, suggesting that they did not adequately recharge on weekends. Face-to-face family time, hanging out with friends, and media use were also inversely related with homework/studying duration on weekends, suggesting that adolescents who spend long hours on homework may be missing out on social interactions, and opportunities for rest and relaxation, that are important for emotional health and development.
      An important limitation of our cross-sectional study is that we cannot be certain of the direction of the association (i.e., the cause-and-effect relationship) between homework/studying duration and time in bed or depression symptoms. While long hours spent on homework/studying may decrease time in bed for sleep, it is also possible that those students who allocate more time for sleep are able to work more efficiently and do not need to spend as much time on homework/studying. Effects of workload versus work efficiency could not be distinguished because we did not collect information on the number of classes that adolescents were taking or the amount of homework that was assigned. On weekends, students who spent long hours on homework may have been more prone to expressing depression symptoms due to reduced time spent on leisure and social activities (i.e., life might not be fun). Alternatively, adolescents with greater self-perceived academic stress may have felt compelled to commit more time to homework/studying to achieve academic success, which may have affected their mood. Prospective studies are needed to determine the cause-and-effect relationships between students’ sleep and homework behaviors and associated changes in depression symptoms.
      The present study assessed time use by asking adolescents to report how much time they usually spend per day on different activities. A limitation of this approach is that information on the timing and sequence of activities was not captured, i.e. we did not specifically examine the duration of homework/studying immediately preceding bedtime. Other studies have addressed this issue by asking participants to provide a detailed 24-hour recall of activities,
      • Foley L.S.
      • Maddison R.
      • Jiang Y.
      • Marsh S.
      • Olds T.
      • Ridley K
      Presleep activities and time of sleep onset in children.
      ,
      • Basner M.
      • Spaeth A.M.
      • Dinges D.F
      Sociodemographic characteristics and waking activities and their role in the timing and duration of sleep.
      or by tracking behavior over weeks using a daily diary.
      • Fuligni A.J.
      • Hardway C
      Daily variation in adolescents' sleep, activities, and psychological well-being.
      ,
      • Gillen-O'Neel C.
      • Huynh V.W.
      • Fuligni A.J
      To study or to sleep? The academic costs of extra studying at the expense of sleep.
      Similar to our results, these approaches have shown that more time spent on homework was associated with reduced nocturnal sleep. Another limitation of our study is that we did not collect data on socioeconomic status or academic pressure, which have been shown to associate with differences in time spent on homework and sleep.
      • Galloway M.
      • Conner J.
      • Pope D
      Nonacademic effects of homework in privileged, high-performing high schools.
      ,
      • Galloway M.
      • Pope D
      Hazardous homework? The relationship between homework, goal orientation, and well-being in adolescence.
      ,
      Organisation for Economic Co-operation and Development
      PISA 2012 Results: What Makes Schools Successful (Volume IV).
      ,
      • Felden E.P.
      • Leite C.R.
      • Rebelatto C.F.
      • Andrade R.D.
      • Beltrame T.S
      Sleep in adolescents of different socioeconomic status: a systematic review.
      It is possible that these variables or other unexplored factors may have influenced results of the present study. We also did not collect daily time spent on napping, which may compensate for shorter nocturnal sleep.
      • Lo J.C.
      • Lee S.M.
      • Teo L.M.
      • Lim J.
      • Gooley J.J.
      • Chee M.W
      Neurobehavioral impact of successive cycles of sleep restriction with and without naps in adolescents.
      ,
      • Lo J.C.
      • Twan D.C.K.
      • Karamchedu S.
      • et al.
      Differential effects of split and continuous sleep on neurobehavioral function and glucose tolerance in sleep-restricted adolescents.
      Our study also had several potential sources of selection bias related to the types of schools (e.g., our sample included high-performing schools) and students who agreed to take part in the research.
      • Yeo S.C.
      • Jos A.M.
      • Erwin C.
      • et al.
      Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
      Finally, we cannot rule out the possibility that some participants may have interpreted bedtime as the time that they got into bed, rather than the time that they attempted to sleep. This would result in an overestimation of time in bed for sleep. Prior work that used the same survey questions in adolescents found a weak-to-moderate correlation with nocturnal sleep period determined by actigraphy monitoring.
      • Wolfson A.R.
      • Carskadon M.A.
      • Acebo C.
      • et al.
      Evidence for the validity of a sleep habits survey for adolescents.
      Hence, a general limitation of our study is that self-reported sleep characteristics may differ from objectively monitored sleep behavior.
      Future studies should examine the drivers of long hours spent on homework/studying, in order to develop interventions that can help students strike a healthy balance between schoolwork and personal life. This includes strategies for improving time management, and working with schools and parents to ensure a manageable workload for adolescents. While we have shown a link between time spent on homework/studying and depression symptoms, the potential clinical implications are unclear. Additional studies are needed to evaluate the relative impact of homework/studying on sleep habits and mental health in pediatric populations with depression or anxiety. There is also a need for longitudinal studies to assess relationships between time use, sleep, and well-being, in order to understand the temporal development and interaction of these factors. Similar to previous findings for paid work in adults,
      • Basner M.
      • Fomberstein K.M.
      • Razavi F.M.
      • et al.
      American time use survey: sleep time and its relationship to waking activities.
      we found that homework/studying duration was inversely associated with time for nocturnal sleep in adolescents. It is possible that behaviors set during childhood and adolescence are carried forward through adulthood, which may have long-term consequences for life satisfaction and health. In future work, it is therefore important to examine the role of homework/studying and other activities that compete for adolescents’ time on their sleep behavior and socio-emotional development.

      Conclusions

      In highly competitive educational settings, there is strong pressure on students to work long hours in the pursuit of academic excellence. Our results in adolescents indicate that spending a long amount of time on homework/studying was associated with reduced time for nocturnal sleep, media use, and social activities with friends or family. On school days, increased time spent on homework/studying may indirectly influence adolescents’ mood by reducing their opportunity for sleep. In contrast, homework/studying duration on weekends associated with greater depression symptom severity, even though adolescents showed marked sleep extension compared with school days. The belief that homework is necessary for improving academic achievement is entrenched in many societies, which makes it difficult to convince students, parents, and schools that reducing students’ workload outside of school should be endorsed in order to make more time for sleep and other activities such as leisure and face-to-face family time. Even small changes in students’ homework/studying habits could impact substantially on opportunities for sleep and other waking activities that are important for schoolwork-life balance and well-being. Future studies should examine interventions to reduce students’ workload and homework prior to bedtime, with the aim of helping adolescents to achieve better sleep and lowering academic stress.

      Conflict of interest

      The authors have no conflicts of interest to disclose.

      Acknowledgments

      We thank students, parents, and teachers for their participation and contribution to the research. We also thank research assistants from the Chronobiology and Sleep Laboratory and the Cognitive Neuroscience Laboratory (Duke-NUS Medical School) who assisted in data collection and analyses.
      This work was supported by the National Research Foundation, Singapore (NRF/2015/SOL001-004, NRF/2016/SOL002-001); the National Medical Research Council, Singapore (NMRC/StaR/015/2013); and the Far East Organization (unrestricted donation to the Center for Cognitive Neuroscience, Duke-NUS Medical School).

      Appendix. Supplementary materials

      References

        • Dewald J.F.
        • Meijer A.M.
        • Oort F.J.
        • Kerkhof G.A.
        • Bogels S.M
        The influence of sleep quality, sleep duration and sleepiness on school performance in children and adolescents: a meta-analytic review.
        Sleep Med Rev. 2010; 14: 179-189
        • Lo J.C.
        • Ong J.L.
        • Leong R.L.
        • Gooley J.J.
        • Chee M.W
        Cognitive performance, sleepiness, and mood in partially sleep deprived adolescents: the need for sleep study.
        Sleep. 2016; 39: 687-698
        • Chaput J.P.
        • Gray C.E.
        • Poitras V.J.
        • et al.
        Systematic review of the relationships between sleep duration and health indicators in school-aged children and youth.
        Appl Physiol Nutr Metab. 2016; 41: S266-S282
        • Crowley S.J.
        • Wolfson A.R.
        • Tarokh L.
        • Carskadon M.A
        An update on adolescent sleep: new evidence informing the perfect storm model.
        J Adolesc. 2018; 67: 55-65
        • Roenneberg T.
        • Kuehnle T.
        • Pramstaller P.P.
        • et al.
        A marker for the end of adolescence.
        Curr Biol. 2004; 14: R1038-R1039
        • Jenni O.G.
        • Achermann P.
        • Carskadon M.A
        Homeostatic sleep regulation in adolescents.
        Sleep. 2005; 28: 1446-1454
        • Short M.A.
        • Gradisar M.
        • Wright H.
        • Lack L.C.
        • Dohnt H.
        • Carskadon M.A
        Time for bed: parent-set bedtimes associated with improved sleep and daytime functioning in adolescents.
        Sleep. 2011; 34: 797-800
        • Yeo S.C.
        • Jos A.M.
        • Erwin C.
        • et al.
        Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: problems and possible solutions.
        Sleep Med. 2019; 60: 96-108
        • Watson N.F.
        • Martin J.L.
        • Wise M.S.
        • et al.
        Delaying middle school and high school start times promotes student health and performance: an American Academy of Sleep Medicine Position Statement.
        J Clin Sleep Med. 2017; 13: 623-625
        • Wang S.M.
        • Zou J.L.
        • Gifford M.
        • Dalal K
        Young students' knowledge and perception of health and fitness: a study in Shanghai, China.
        Health Educ J. 2014; 73: 20-27
        • Cooper H.
        • Robinson J.C.
        • Patall E.A
        Does homework improve academic achievement? A synthesis of research, 1987–2003.
        Rev Educ Res. 2006; 76: 1-62
        • Yoo C
        Sleep duration change and its associated factors during adolescence: a 6 year longitudinal study.
        Child Indic Res. 2020; 13: 573-590
        • Street N.W.
        • McCormick M.C.
        • Austin S.B.
        • Slopen N.
        • Habre R.
        • Molnar B.E
        Examining family and neighborhood level predictors of sleep duration in urban youth.
        Fam Syst Health. 2018; 36: 439-450
        • Adam E.K.
        • Snell E.K.
        • Pendry P
        Sleep timing and quantity in ecological and family context: a nationally representative time-diary study.
        J Fam Psychol. 2007; 21: 4-19
        • Yang C.K.
        • Kim J.K.
        • Patel S.R.
        • Lee J.H
        Age-related changes in sleep/wake patterns among Korean teenagers.
        Pediatrics. 2005; 115: 250-256
        • Twenge J.M.
        • Krizan Z.
        • Hisler G
        Decreases in self-reported sleep duration among U.S. adolescents 2009–2015 and association with new media screen time.
        Sleep Med. 2017; 39: 47-53
        • Galloway M.
        • Conner J.
        • Pope D
        Nonacademic effects of homework in privileged, high-performing high schools.
        J Exp Educ. 2013; 81: 490-510
        • Galloway M.
        • Pope D
        Hazardous homework? The relationship between homework, goal orientation, and well-being in adolescence.
        Encounter Educ Mean Soc Justice. 2007; 20: 25-31
        • Ong X.L.
        • Cheung H.S
        Schools and the Class Divide: An Examination of Children's Self-Concept and Aspirations in Singapore.
        Singapore Children's Society, 2016
        • Huang G.H.-.C.
        • Gove M
        Confucianism, Chinese families, and academic achievement: exploring how Confucianism and Asian descendant parenting practices influence children's academic achievement.
        in: Khine MS Science Education in East Asia: Pedagogical Innovations and Research-Informed Practices. Springer International Publishing, Cham2015: 41-66
        • Stankov L
        Unforgiving Confucian culture: a breeding ground for high academic achievement, test anxiety and self-doubt?.
        Learn Individ Differ. 2010; 20: 555-563
        • Organisation for Economic Co-operation and Development
        PISA 2015 Results: Students' Well-Being (Volume III).
        2017
        • Lo J.C.
        • Lee S.M.
        • Lee X.K.
        • et al.
        Sustained benefits of delaying school start time on adolescent sleep and well-being.
        Sleep. 2018; 41: zsy052
        • Wolfson A.R.
        • Carskadon M.A
        Sleep schedules and daytime functioning in adolescents.
        Child Dev. 1998; 69: 875-887
        • Wolfson A.R.
        • Carskadon M.A.
        • Acebo C.
        • et al.
        Evidence for the validity of a sleep habits survey for adolescents.
        Sleep. 2003; 26: 213-216
        • Brooks S.J.
        • Krulewicz S.P.
        • Kutcher S
        The Kutcher Adolescent Depression Scale: assessment of its evaluative properties over the course of an 8-week pediatric pharmacotherapy trial.
        J Child Adolesc Psychopharmacol. 2003; 13: 337-349
        • Zhou H.
        • Hao N.
        • Du Y.
        • et al.
        Reliability and validity of the eleven item Kutcher Adolescent Depression Scale, Chinese Version (KADS-11CV).
        J Child Adolesc Behav. 2016; 4: 1-5
        • Basner M.
        • Fomberstein K.M.
        • Razavi F.M.
        • et al.
        American time use survey: sleep time and its relationship to waking activities.
        Sleep. 2007; 30: 1085-1095
        • Fuligni A.J.
        • Hardway C
        Daily variation in adolescents' sleep, activities, and psychological well-being.
        J Res Adolesc. 2006; 16: 353-378
        • Baum K.T.
        • Desai A.
        • Field J.
        • Miller L.E.
        • Rausch J.
        • Beebe D.W
        Sleep restriction worsens mood and emotion regulation in adolescents.
        J Child Psychol Psychiatry. 2014; 55: 180-190
        • Preacher K.J.
        • Hayes A.F
        Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models.
        Behav Res Methods. 2008; 40: 879-891
        • Sun W.Q.
        • Spruyt K.
        • Chen W.J.
        • et al.
        The relation among sleep duration, homework burden, and sleep hygiene in Chinese school-aged children.
        Behav Sleep Med. 2014; 12: 398-411
        • Li S.
        • Yang Q.
        • Chen Z.
        • Jin X.
        • Jiang F.
        • Shen X
        Homework schedule: an important factor associated with shorter sleep duration among Chinese school-aged children.
        Behav Sleep Med. 2014; 12: 389-397
        • Jiang X.
        • Hardy L.L.
        • Baur L.A.
        • Ding D.
        • Wang L.
        • Shi H
        Sleep duration, schedule and quality among urban Chinese children and adolescents: associations with routine after-school activities.
        PLoS One. 2015; 10e0115326
        • Organisation for Economic Co-operation and Development
        PISA 2012 Results: What Makes Schools Successful (Volume IV).
        2013
        • Gau S.F.
        • Soong W.T
        Sleep problems of junior high school students in Taipei.
        Sleep. 1995; 18: 667-673
        • Short M.A.
        • Gradisar M.
        • Lack L.C.
        • et al.
        A cross-cultural comparison of sleep duration between US And Australian adolescents: the effect of school start time, parent-set bedtimes, and extracurricular load.
        Health Educ Behav. 2013; 40: 323-330
        • Foley L.S.
        • Maddison R.
        • Jiang Y.
        • Marsh S.
        • Olds T.
        • Ridley K
        Presleep activities and time of sleep onset in children.
        Pediatrics. 2013; 131: 276
        • Carskadon M.A
        Factors influencing sleep patterns of adolescents.
        in: Carskadon MA Adolescent Sleep Patterns: Biological, Social, and Psychological Influences. Cambridge University Press, Cambridge2002: 4-26
        • Lushington K.
        • Wilson A.
        • Biggs S.
        • Dollman J.
        • Martin J.
        • Kennedy D
        Culture, extracurricular activity, sleep habits, and mental health: a comparison of senior high school Asian-Australian and Caucasian-Australian adolescents.
        Int J Ment Health. 2015; 44: 139-157
        • Hysing M.
        • Pallesen S.
        • Stormark K.M.
        • Jakobsen R.
        • Lundervold A.J.
        • Sivertsen B
        Sleep and use of electronic devices in adolescence: results from a large population-based study.
        BMJ Open. 2015; 5e006748
        • Ren H.
        • Zhou Z.
        • Liu W.K.
        • Wang X.
        • Yin Z
        Excessive homework, inadequate sleep, physical inactivity and screen viewing time are major contributors to high paediatric obesity.
        Acta Paediatr. 2017; 106: 120-127
        • Boers E.
        • Afzali M.H.
        • Newton N.
        • Conrod P
        Association of screen time and depression in adolescence.
        JAMA Pediatr. 2019; 173: 853-859
        • Salk R.H.
        • Hyde J.S.
        • Abramson L.Y
        Gender differences in depression in representative national samples: meta-analyses of diagnoses and symptoms.
        Psychol Bull. 2017; 143: 783-822
        • Basner M.
        • Spaeth A.M.
        • Dinges D.F
        Sociodemographic characteristics and waking activities and their role in the timing and duration of sleep.
        Sleep. 2014; 37: 1889-1906
        • Gillen-O'Neel C.
        • Huynh V.W.
        • Fuligni A.J
        To study or to sleep? The academic costs of extra studying at the expense of sleep.
        Child Dev. 2013; 84: 133-142
        • Felden E.P.
        • Leite C.R.
        • Rebelatto C.F.
        • Andrade R.D.
        • Beltrame T.S
        Sleep in adolescents of different socioeconomic status: a systematic review.
        Rev Paul Pediatr. 2015; 33: 467-473
        • Lo J.C.
        • Lee S.M.
        • Teo L.M.
        • Lim J.
        • Gooley J.J.
        • Chee M.W
        Neurobehavioral impact of successive cycles of sleep restriction with and without naps in adolescents.
        Sleep. 2017; 40: zsw042
        • Lo J.C.
        • Twan D.C.K.
        • Karamchedu S.
        • et al.
        Differential effects of split and continuous sleep on neurobehavioral function and glucose tolerance in sleep-restricted adolescents.
        Sleep. 2019; 42: zsz037