OBJECTIVE: Presleep activities have been implicated in the declining sleep duration of young people. A use-of-time approach may be used to describe the presleep period. The study aims were to describe the activities undertaken 90 minutes before sleep onset and to examine the association between activities and time of sleep onset in New Zealand young people.
METHODS: Participants (N = 2017; 5–18 years) self-reported their time use as part of a national survey. All activities reported in the 90 minutes before sleep were extracted. The top 20 activities were grouped into 3 behavioral sets: screen sedentary time, nonscreen sedentary time, and self-care. An adjusted regression model was used to estimate presleep time spent in each behavioral set for 4 distinct categories of sleep onset (very early, early, late, or very late), and the differences between sleep onset categories were tested.
RESULTS: In the entire sample, television watching was the most commonly reported activity, and screen sedentary time accounted for ∼30 minutes of the 90-minute presleep period. Participants with a later sleep onset had significantly greater engagement in screen time than those with an earlier sleep onset. Conversely, those with an earlier sleep onset spent significantly greater time in nonscreen sedentary activities and self-care.
CONCLUSIONS: Screen sedentary time dominated the presleep period in this sample and was associated with a later sleep onset. The development of interventions to reduce screen-based behaviors in the presleep period may promote earlier sleep onset and ultimately improved sleep duration in young people.
- CAPI —
- computer-assisted personal interview
- CATI —
- computer-assisted telephone interview
- CI —
- 95% confidence interval
- MARCA —
- Multimedia Activity Recall for Children and Adults
What’s Known on This Subject:
Presleep activities (eg, television watching) have been implicated in the declining sleep duration of young people. However, previous research reported on selected presleep activities, raising the possibility that important activities in this period are not accounted for.
What This Study Adds:
This is the first study in youth to construct the presleep period by using a use-of-time approach. Twin trajectories of higher screen time and lower nonscreen sedentary time/self-care were evident in late sleepers, with the opposite pattern occurring in early sleepers.
During the past 100 years, a rapid decline in the sleep duration of young people of >1 hour per night is evident.1 Inadequate sleep has been associated with a range of behavioral and health disturbances in young people, including poor concentration and academic performance,2 lack of coordination,3 increased aggression,4,5 hyperactivity,5 metabolic dysfunction,6 and obesity.7 Reduced total sleep time is thought to be due to later bedtimes rather than to early waking,8 and therefore, presleep activities may be implicated through disrupting or displacing sleep. These presleep activities include the use of electronic screen–based media9 (eg, television, computers, and video games) and other nonscreen activities, such as homework.10
The association between screen time and sleep onset or duration has been examined in a number of studies in young people. In Saudi 6- to 13-year-olds (N = 1012), late evening television viewing or computer game play was associated with reduced total sleep time (−0.63 hour, 95% confidence interval [CI] −7.9 to −0.47, P < .001).11 Similarly, television watching (β = −.041, P < .05), computer games (β = −.064, P < .001), and Internet use (β = −.119, P < .001) were all associated with reduced total sleep time during weekdays in Flemish adolescents (N = 2546).12 Data from a sample of Australian adolescents (N = 2200) indicated 79% were engaged in screen-based activities (television, video games, and computer) in the 120 minutes before sleep onset, with screen time accounting for more than half of this presleep period. Total daily screen time of adolescents with very late or late sleep onset was significantly greater than for those with very early or early sleep onset (247 vs 208 min/d; P < .0001).13 In addition to these traditional screen activities, presleep use of mobile phones (eg, text messaging) may also disrupt sleep. In the same Flemish cohort described earlier, the odds ratio of being “very tired” 1 year later increased to 3.3 (95% CI 1.9–5.7) and 5.1 (95% CI 2.5–10.4) in participants who used a mobile phone once per week or more than once per week after the lights are turned out, respectively, compared with those who did not do this.14
Presleep electronic media use is hypothesized to affect sleep patterns in 3 ways: time displacement, depression of melatonin, and cognitive arousal. Time displacement of sleep has been shown to increase when a media device, such as a television, is present in the bedroom.15 In addition, the blue light emitted by screens attenuates melatonin concentrations in children,16 which disrupts the circadian rhythm and delays sleep onset.17 Finally, “thriller” or action-oriented electronic games may stimulate wakefulness through heightened cognitive processes, such as fear or excitement,9,18,19 which may be reflected in somatic outcomes such as elevated heart rate and perspiration.20 These 3 factors have potentially additive and adverse effects on the viewer’s sleep duration.
Other nonscreen activities have also been associated with inadequate sleep in young people, most likely primarily through time displacement. For example, in a study of 5- to 19-year-old Americans (N = 2454), academic and religious activities were significant predictors of lower sleep duration.21 One other study examined the association between extracurricular activities, extracurricular employment, and sleep in American adolescents aged 12 to 19 years (N = 3094). A gradient effect on sleep duration was found: those who reported low levels of both extracurricular activity and work reported the highest sleep duration (459 min/night); those who reported high levels of extracurricular activities or high levels of work had a shorter sleep duration (448 and 429 min/night, respectively); and those who reported high levels of both extracurricular activities and work had the shortest sleep duration (408 min/night).22
Because of their association with sleep onset and duration, presleep activities are a logical intervention target to improve sleep in young people. However, there is a paucity of information regarding what young people do in the presleep period, how this influences sleep onset, and what is the ultimate impact on sleep duration. Previous research has reported on a priori selected presleep activities (primarily screen time), raising the possibility that important activities in this period are not accounted for. A “use-of-time” approach may be used to gain a holistic understanding of what young people are doing in the time preceding sleep. This approach involves the complete construction of all activities performed within a defined and bounded time period, accounting for all of the time in that period. In this context, this relates to the time period immediately before sleep onset. The use-of-time approach may be used to identify individual activities (eg, television or reading) or broader behavioral sets of interest (eg, non–screen-based sedentary behavior) occurring in the time period of interest. Therefore, the aims of this study were to examine the period of time 90 minutes before sleep onset to (1) provide a descriptive account of the most popular activities and (2) explore the relationship between presleep behaviors and sleep onset in a nationally representative sample of New Zealand young people aged 5 to 18 years.
A nationally representative cross-sectional survey of 2503 New Zealand children and young people aged 5 to 24 years was conducted between September 2008 and May 2009. The survey was conducted according to the ethical principles outlined in the Declaration of Helsinki and was covered by Statistics New Zealand Tier 1 ethical approval. Written consent was obtained from all participants or their parent, depending on the age of the participant. The survey design and methodology have been reported in detail elsewhere.23 Presleep data for participants aged 5 to 18 years are reported here.
Design and Participants
A complex survey design involving stratified multistage sampling was used. The primary sampling unit was a mesh block, which is a defined geographic area, varying in size from part of a city block to large areas of rural land. Within each mesh block, eligible households were identified and asked to participate in the survey. One child or young person was randomly chosen from each eligible household. The response rate was 55% and a total of 2503 households were surveyed. The race/ethnicity of the sample was 18.8% Māori (indigenous population), 9.6% Pacific, 12.9% Asian, and 71.4% New Zealand European. This is representative of the ethnic composition of the general New Zealand population.24
Data were collected during a face-to-face home visit (computer-assisted personal interview [CAPI]) and a subsequent telephone interview (computer-assisted telephone interview [CATI]) conducted 7 to 14 days after the CAPI. The CAPI collected self-report data on sociodemographic characteristics and 1 to 2 days of use-of-time data (which included the assessment of activities undertaken in the presleep period). The subsequent CATI collected an additional 2 days of self-reported use-of-time data.
Home visits were conducted mainly on weekend days to maximize the chance of participants being present at home for recruitment. To reduce the bias associated with collecting all time use data on the weekend, the return CATI was made on a weekday.
Self-reported time use, including sleep, physical activity, and sedentary behavior, was measured by using the Multimedia Activity Recall for Children and Adults (MARCA).25 The MARCA is a computerized use-of-time tool. All daily activities (including sleep) are retrospectively recalled in sequential time segments of 5 minutes or more for 24 hours of the previous day (midnight to midnight). Participants choose from a list of ∼250 activities. Each activity is linked to an energy cost taken from existing child26 and adult27 compendia. Metabolic equivalents28 are used to describe the intensity of activities. The MARCA has been shown to have adequate psychometric properties.25,29 For the current survey, up to 4 days of recall were completed. Participants recalled the 2 previous days of activity at each of the 2 data collection periods. Because of the limited cognitive ability of young children to recall time accurately,30 parents of participants aged 5 to 9 years provided a proxy recall of their child’s activities when they were directly supervising the child, including the presleep period.
For each eligible participant, 1 MARCA profile was randomly selected for analysis to avoid problems associated with intraindividual clustering. For each profile, the self-reported time of evening sleep onset was identified. Profiles with a reported sleep onset earlier than 6:00 pm were regarded as invalid and excluded. All activities reported in the 90 minutes before sleep onset were extracted for each profile, accounting for the entire presleep period. The top 20 most popular activities (by frequency of reporting) were identified and grouped into 3 intuitive behavioral sets: screen sedentary time (television, computer, and video games), nonscreen sedentary time (eg, reading, eating, and talking), and self-care (eg, showering, brushing teeth, and getting ready for bed). Presleep time spent in each behavioral set was calculated for each profile. If a profile reported none of the activities comprising a particular behavioral set, the time in that set was considered to be zero.
According to time of sleep onset, profiles were classified into 4 mutually exclusive categories (very early, early, late, and very late) by using quartiles of the residuals obtained from adjusted regression analysis. For each profile, sleep onset was estimated in the model adjusting for participant age and gender, as well as day type (school day or not school day). The residuals therefore represented sleep onset for each participant relative to other young people of the same age and gender on the same type of day.
All statistical analyses were performed by using SAS version 9.2 (SAS Institute, Inc, Cary, NC). Descriptive information on time spent in each behavioral set (screen sedentary time, nonscreen sedentary time, and self-care), for each of the 4 sleep onset categories (very early, early, late, and very late), was presented by age group (5–12 and 13–18 years) and gender (male and female). Regression analysis was conducted to estimate the time in each behavioral set for different categories of sleep onset, adjusting for age, gender, and day type. Differences between sleep onset categories were tested statistically, with the α set at .05.
A total of 2017 survey participants aged 5 to 18 years were included in the current analysis. The sample consisted of 951 (47.1%) girls, and the mean age was 11.6 years (SD 3.7 years).
Average time of sleep onset is presented in Table 1 for each age group and gender. In general, younger participants had an earlier sleep onset than older participants. There were no clear differences between male and female participants. The top 20 most popular activities (by frequency of reporting) in the 90 minutes before sleep onset are presented in Table 2. These activities were grouped into 3 behavioral sets. The activities performed in the presleep period were predominantly of low intensity, in the range of 1 to 2 metabolic equivalents. The 3 most commonly reported activities were watching television, dressing/undressing, and brushing teeth. Overall, these top 20 activities accounted for ∼80% of time in the 90-minute presleep period.
Descriptive data on time (minutes) spent in each of the 3 behavioral sets for each of the 4 sleep onset categories (very early, early, late, and very late) are presented in Table 3. Screen sedentary time accounted for the most time in the 90 minutes before sleep onset, ∼30 minutes or one-third of this period. Consistent with overall daily patterns of screen use in New Zealand31 and Australia,32 older participants and male participants engaged in more screen time in the presleep period than did younger participants and female participants, respectively. The opposite trajectory was found for nonscreen sedentary time and self-care; younger participants and female participants spent more time in these activities in the presleep period than did older participants and male participants, respectively. The descriptive data also indicated that screen time tended to be higher in those with a later sleep onset.
The model-adjusted means and differences in time spent in each of the behavioral sets between sleep onset categories are presented in Table 4. An early sleep onset was associated with significantly less time in screen-based sedentary activities compared with a later sleep onset. The difference between earlier and later sleep onset categories ranged between 4 and 13 minutes (eg, those in the late group spent 13 minutes more of the presleep period in screen time than did those in the very early group). Conversely, an early sleep onset was associated with significantly more time spent in nonscreen sedentary behavior, though the difference between early and late categories was less pronounced (5–8 minutes). Finally, for self-care, there were significant differences between early and late groups (higher in the early group), but these differences were even smaller (2–5 minutes).
The aims of this study were to describe the activities of 5- to 18-year-old New Zealand young people during the 90 minutes before sleep onset and to investigate the association between these activities and time of sleep onset. This is the first study to date to use a use-of-time approach to construct the entire presleep period, rather than reporting on selected individual activities.
In the entire sample, screen sedentary time (in particular, television watching) dominated the presleep period, by both frequency of reporting and duration. Nearly half the sample reported sitting and watching television, and screen time accounted for ∼30 minutes of the 90-minute period. In New Zealand, a maximum of 2 hours of screen time per day is recommended for young people33; therefore, participants in this study accumulated one quarter of this recommendation in the 90 minutes before sleep alone. In addition to these findings, those with a later sleep onset reported up to 13 more minutes of screen time in the presleep period than did those with an earlier sleep onset. The largest mean time differences between those of early and late sleep onset were for screen time, which suggests that this set of activities may be an appropriate target for interventions to promote earlier sleep onset and subsequently improve sleep duration in young people.
Causality, and the direction of causality, cannot be inferred from a cross-sectional design; therefore, it is necessary to consider a number of explanations when assessing the relationship between screen time and sleep onset. One explanation is that screen time may cause later sleep onset via somatic arousal, attenuation of melatonin, or sleep displacement, as proposed by other researchers. Alternatively, young people with later sleep onset may have a greater amount of discretionary time during which they can engage in screen time. A young person who arrives home from school at 4:00 pm and goes to sleep at 9:00 pm has 1 additional hour of discretionary time than a young person who arrives home at the same time but goes to sleep at 8:00 pm. If essential tasks, such as homework and dinner, take up a proportion of after-school time, those with a greater overall amount of after-school time, due to later sleep onset, may use this extra time for recreational screen-based activities. The sequencing of events may also be important. Those with later sleep onset may complete their self-care activities earlier before engaging in screen-time later in the evening, with the opposite pattern occurring in those with earlier sleep onset. In this way, screen time may not be keeping late onset sleepers awake but is reflective of differential time-patterning resulting in higher engagement in screen time in the 90 minutes immediately before sleep.
Use-of-time research operates within a bounded time period; therefore, it necessarily follows that increased engagement in one behavioral set results in reduced time spent in other behaviors. In this study, twin trajectories of higher screen time and lower nonscreen sedentary time and self-care were evident in those with a later sleep onset, with the opposite pattern occurring in those with an earlier sleep onset. Although statistically significant, the differences between early and late sleepers for nonscreen sedentary time (5–8 minutes) and self-care (2–5 minutes) were small and unlikely to be of clinical significance.
The results of this study are consistent with reports involving Saudi,11 Flemish,12 and Australian13 young people that indicated that screen-based activities are associated with lower sleep duration or later sleep onset. However, the use of mobile phones in the presleep period was not commonly reported by New Zealand young people, in contrast to previous research in Flemish adolescents.14 Previous research10,21,22 examining the effect of nonscreen activities on sleep (eg, extracurricular activities or employment) is not directly comparable to this study. Earlier studies assessed all nonschool time, whereas the current study assessed the 90 minutes before sleep onset only, when it is unlikely that these activities would occur. Less than 10% of participants reported engaging in study or homework in the 90 minutes before sleep onset. In the current study, we considered all presleep activities, most of which were categorized as self-care. Very few active pursuits were reported in the presleep period, but they may also influence sleep behavior.
The strengths of this study include the use of a large, nationally representative sample and the holistic approach toward the construction of the entire presleep period. However, several limitations should be acknowledged. First, there was a 45% nonparticipation rate, and nonparticipants may have differed from participants in important ways. However, the sample was representative of the New Zealand population by ethnicity, age, and geography. As discussed, the cross-sectional nature of the study does not allow for inferences of causality to be made. In addition, self-report tools such as the MARCA are associated with error due to memory or social desirability bias, although the MARCA has been shown to have sound psychometric properties,25,29 and in the current study the proportion of valid data were high (99.6%). Using data from a subset of participants (5–18 years) was considered the most appropriate approach to compare young people at a similar stage of development and life circumstances (preadolescent or adolescent school students). The current study examined time of sleep onset but not total sleep duration, which may be an avenue for future research. One final limitation is it was unknown where the activity occurred, although it may be assumed it was primarily in the home. For example, having a television in the child’s bedroom has been linked with sleep disturbances in previous research.9
Screen time accounted for one third of the 90 minutes before sleep onset in New Zealand young people aged 5 to 18 years. Higher engagement was evident in participants with a later sleep onset, suggesting that reducing screen sedentary time may be an appropriate intervention for promoting earlier sleep onset in young people.
The authors wish to thank all participants in the national survey.
- Accepted October 1, 2012.
- Address correspondence to Louise S. Foley, PhD, c/o Ralph Maddison, National Institute for Health Innovation, University of Auckland, Private Bag 92019, Auckland Mail Centre, Auckland 1142, New Zealand. E-mail:
Drs Maddison, Jiang, Olds, and Ridley were involved in the conception, design, and analysis of the original national survey; Dr Foley extracted the data and drafted the article; Dr Jiang conducted the analysis; and all authors were involved in the design of the presleep analysis reported, contributed to the interpretation of data, were responsible for revising it critically for important intellectual content, and gave final approval of the version to be published.
FINANCIAL DISCLOSURE: Dr Foley was supported by a Tertiary Education Commission Bright Futures Doctoral Scholarship and is currently supported by a Heart Foundation of New Zealand Postdoctoral Fellowship; Dr Maddison was supported by a Heart Foundation of New Zealand Fellowship and is currently supported by a Health Research Council of New Zealand Sir Charles Hercus Health Research Fellowship; Dr Marsh was supported by a University of Auckland Doctoral scholarship; the other authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: The national survey was funded by Sport and Recreation New Zealand (now Sport New Zealand), the Ministry of Health, the Ministry of Education and the Ministry of Youth Development.
- Kuriyama K,
- Stickgold R,
- Walker MP
- Spruyt K,
- Molfese DL,
- Gozal D
- Liu X,
- Liu L,
- Owens JA,
- Kaplan DL
- Owens J,
- Maxim R,
- McGuinn M,
- Nobile C,
- Msall M,
- Alario A
- Galloway MK,
- Pope D
- Olds T,
- Ridley K,
- Blunden S
- ↵Must A, Parisi SM. Sedentary behavior and sleep: paradoxical effects in association with childhood obesity. Int J Obes. 2009;33(Suppl 1): S82–S86
- Carskadon MA
- ↵Clinical Trials Research Unit, Synovate. A National Survey of Children and Young People's Physical Activity and Dietary Behaviours in New Zealand: 2008/09–Key Findings. Wellington, New Zealand: Ministry of Health; 2010
- ↵Statistics New Zealand. 2006 Census Data–QuickStats About Culture and Identity. Wellington, New Zealand: Statistics New Zealand; 2006
- ↵Sport and Recreation New Zealand. New Zealand physical activity guidelines for children and young people (5-18 years old): Available at: www.sparc.org.nz/getting-active/activity-guidelines
- Copyright © 2013 by the American Academy of Pediatrics