The problem I have with this - looking at the abstract in haste - is that it seems to think that 'regular' bedtimes equate to 'early' bedtimes.
I have two ds, ds2 has never gone to sleep early, even as a baby he never went to sleep after his 7pm feed, whereas ds1 did.
Even now, ds2 tends to go to sleep later than ds1. He is good at recognising when he is tired and happy to climb into bed at that point. He has a regular bedtime, he has his bath, story, toothbrushing etc routine - but it's later in the evening than for most children of 5 years old I'm guessing (typically upstairs at 8pm, he'll be in bed by 9 and he'll usually fall asleep between 9 and 11 at some point).
But the abstract starts by talking about early bedtimes - and then goes on to talk about the effects of having a regular - or not - bedtime. DS2 can actually end up worse off if he does have a night when he goes to sleep early (ie irregular from his normal late but regular time) as it throws him off kilter in the morning, he wakes up too early and tired and he's out of sync for the rest of the next day.
It would also be interesting to see what time of day the children were tested on for their maths and english etc scores. I'm not a morning person - and wouldn't do very well on a test first thing in the morning. Give it to me later on though and I'd fly through it no problem. DH is the opposite - he'd do really well in the morning - give it to him at 6 in the evening and he'd struggle as he's really tired at that point. If you are testing all the children at the same time of day then it stands to reason that if some children are owls and others are larks as for adults, then they have different times of the day when they are at their peak mentally.
If you always test at a good time for larks then chances are the owls might not be at their peak and therefore their scores lower. But if you test at a different time of day then you might find the scores differ. You certainly shouldn't be hypothesizing that one group is better than the other without checking for the testing time factor - it's a bit like the old psychology experiments that 'showed' that children in Africa were 'less intelligent' than their european counterparts when the same IQ tests were used. But actually all it was doing was showing that the experimenters were the ones with a problem - the tests had a huge cultural bias - when the tests were re-designed around what the subjects actually knew (eg - if you had to split up apples, pears, lemons and oranges into two groups then most european kids could do it as they are familiar with all the fruit. If you've only ever eaten peaches, mangoes and melons, then apples, pears, lemons and oranges all fall into the 'fruit I don't know' category - and the european kids would have just the same problems categorising peaches, mangoes and melons. It's not measuring intelligence but knowledge. (if the details of this are slightly wrong then it's because it's over 20 years since I studied it and I can't remember the specifics - however the general implications are right).
It also seems to be equating bed times to amount of sleep you get in hours, which again, particularly for younger children who might not have to get up for a school run - not necessarily follow that just because you go to sleep late you get less sleep. It also doesn't mention anything to do with whether or not the child wakes up naturally themself or whether they need to be woken up. My ds2 who goes to sleep late is also the first of the household to wake up most of the time - whereas ds1 who has gone to sleep earlier often needs to be woken up, even though he has had more sleep.
Basically - I think there are huge number of holes in the study that suggests they either have not thought about the problem and gathered the data in enough detail to analyse it fully, and that they knew what outcome they wanted so they just (implicitly or explicitly) got enough data to prove what they wanted to prove. Or they have written it up very badly if they have looked at the data and have been able to take more factors into account other than early/regular bedtimes. Oh and it's very sloppy of them to equate early bedtimes with regular bedtimes without explaining what they mean by both terms as they are completely different things!