Recently, news reports are making the rounds suggesting that the lack of a set bedtime in children 3 to 7 is associated with greater behavioural problems. Parents are getting panicked if their child doesn’t go to sleep at the exact same time every night. Doctors read this and tell families they need to have a schedule (as opposed to a routine). This seemed a little odd to me as I’ve always been one who does fine so long as I get enough sleep, even if it comes at different times, so I decided to go to the source to see what I could find.
The research, which is based on the Millennium Cohort Study out of the UK, looked at over 10,000 children and their bedtimes and parent and teacher reported behavioural problems. The families are followed over time so the results look at children who responded for all three time points of interest herein: Ages 3, 5, and 7. The bedtime question was, “On weekdays during term-time, does your child go to bed at a regular time?” with answer options of always, usually, sometimes, and never. Answers to this were then divided into those who said always or usually and then those who said sometimes or never. At ages 5 and 7, if parents reported a regular bedtime, there was a further question about what time bedtime was with half hour intervals between 7:30 and 9pm or either earlier than 7:30pm or later than 9pm. Behavioural measures were mother report on the Strengths and Difficulties Questionnaire (SDQ) – a well-validated, commonly-used measure of a variety of behavioural and emotional areas – at all three ages and then teachers also completed the SDQ at age 7.
Analyses were conducted using regression. Confounds included mother’s age, birth order, family income, parental education, mother’s psychological well-being, maternal discipline tactics (focusing on harsher or negative discipline), maternal irritability with the child, maternal parenting competence (self-report), breastfed ever, child eats breakfast, daily help with academics, TV watching, computer use, parental employment, time spent with child by parents, breakfast club, child care used, reading stories daily to the child, rules for TV watching, overcrowding, bedwetting, child sleep disrupted by wheezing, and TV in bedroom. Notably, children with irregular bedtimes or late bedtimes (past 9pm) fell into a “socially disadvantaged” profile which included things like skipping breakfast, low SES, mothers with poorer mental health, etc.
What did they find? Importantly here, the results spoken of are ONLY for when the children were 7 years of age and in school. At age 7, children who had irregular bedtimes showed significantly greater behavioural problems, both maternal and teacher report. There was a dose-response in that children who had an irregular bedtime at any 1 of the 3 ages measured showed greater maternal report problems (but not teacher report) and this effect was greater when there were 2 of the 3 ages showing irregular bedtimes (and was significant for teacher report as well) and greatest for 3 of the 3 ages.
What can we conclude from this? Well, there are a few problems that the authors skim over or ignore. First, the authors suggest that there are two ways in which the effects they found may come about: Disruption of circadian rhythms or less sleep overall. At age 3, it’s difficult to talk about the disruption of the circadian rhythm as we know that it is still developing at that stage and as naps are still a normal part of many 3-year-olds schedule, their rhythm is still not as developed as that of an adult. Perhaps this does play a role, but my own guess is that the total sleep is more of an issue.
At all three ages tested, it seems fair to assume that the children have a regular rising time – especially at ages 5 and 7 whereby schooling becomes a factor. This means that irregular bedtime influences the amount of sleep a child gets, and as the authors of the article point out, irregular total sleep is predictive of behavioural problems. With this we have to question why later bedtimes weren’t predictive of problems, but I would argue we need more information about the bedtime. The authors included 9pm and later as one group, but a child who goes to bed at 9pm and up at 7am is getting 10 hours of sleep. A child going to sleep at midnight and getting up at 7am is only getting 7 hours, a stark difference. I would also hazard to guess that more children are going to sleep closer to the 9pm than midnight, thus reducing the main effect for that group.
A potential second issue is the third variable problem. Sometimes things like irregular bedtimes are reflective of more erratic parenting practices more generally. Although the authors tried to account for many things in the myriad variables they controlled for, they did not and simply could not account for this particular aspect of parenting or living. Importantly, in line with this, the effects of regular bedtimes on behaviour were drastically dampened (by 47% for teacher report behavioural problems and by 64% for maternal report behavioural problems) when the covariates were included in the model (though they both remained statistically and practically significant). This means that you may have children who have irregular bedtimes, but consistent, responsive parenting and the children get enough sleep. The question really becomes, for these kids, is there a link to behavioural problems?
A final consideration to do with what the authors have considered the “dosage effect”; that is, the fact that the effects gained strength the more time periods the children had an irregular bedtime. While important, I find it bizarre that the authors did not separate those who had an irregular bedtime at the earlier ages from those at 7 years of age given the outcome measure was at 7 years of age. One struggles to think how an irregular bedtime at age 3 (which was quite common in the sample as a whole) would influence behaviour at age 7 when a regular bedtime had been adopted for the interceding 3-4 years. I would have liked to see the effects broken down by these particular age groups to see the actual effect. I also would have liked to see the effects on behaviour at ages 3 and 5, which for some reason was not included.
In summary, the research should be seen as the beginning of looking at the influence of bedtime on behaviour. That said, it is hardly conclusive and leaves a lot to be desired. Clearly, bedtimes themselves seem not to matter, though I doubt that’s the case when the child needs to rise at a certain point and is chronically tired; however, if a child is getting enough sleep, the actual time seems not to matter. If it varies? Well, that’s what we don’t know. If the link is via circadian rhythm development, perhaps irregular bedtimes will matter even if the child gets enough sleep and is in a stable parenting environment. But this research simply can’t inform on that at all. And most importantly, this research says nothing about children’s sleep prior to age 3 or on the behavioural issues prior to age 7. The last thing parents should do is worry about getting their 6 month old on a schedule because of this research, and yet, that’s exactly what I fear will happen. A routine can be very beneficial to children (and adults), but when we stress about set times, we’re entering the realm of sleep training that has been found to have negative effects for the entire family. So please, don’t panic, don’t start looking at the clock over your child, but do consider that if you are worried about sleep, a routine may be a perfect thing for you and your family.
 Kelly Y, Kelly J, Sacker A. Changes in bedtime schedule and behavioral difficulties in 7 year old children. Pediatrics 2013; DOI: 10.1542/peds.2013-1906
 Watamura SE, Donzella B, Kertes DA, Gunnar MR. Developmental changes in baseline cortisol activity in early childhood: relations with napping and effortful control. Developmental Psychobiology 2004; 45: 125-33.
 Walker MP, Stickgold R. Sleep, memory, and plasticity. Annul Rev Psychol 2006; 57: 139-66.
 Bryant PA, Trinder J, Curtis N. Sick and tired: does sleep have a vital role in the immune system? Nat Rev Immunol 2004; 4: 457-67.