Here's an interesting commentary on why we should NOT allow children's heads to go above the top of the seat shell (From the Yahoo group 'cpsplist', the writer is Kathleen Weber, one of the best known and published researchers in the field):
"To "pass" the 213 test, among other things, the target on the side of the
dummy's head (the center of gravity, approximately at the ears) must not
rise above the top of the shell. This is to protect against whiplash injury
from the head swinging over the top of the CR. To keep the target down
during the test, the harness needs to be threaded through slots below
the shoulders and it must be very snug.
If an actual child started out with the ears at the top of the shell, it would be an unusual parent who would tighten the harness enough to keep the center of gravity of the child's head from passing the top of the shell in a severe crash, and the child's shoulders would have to be unusually stiff. Since there will always be some ramping, and the heavier the child, the more ramping will occur, limiting the top of the head to the top of the shell, or a little below, is a good guideline. Other reasons exist as well, such a containing the head to minimize head contact injury or even neck compression."
(Ramping is what it sounds like: the head and body slide up and out of the seatback in a bad crash)
Sweden may have different recommendations because their seats are braced on dashboards (airbags turned off) which can prevent a lot of ramping, or they may have support feet to hold the seat up and the child in place better, but here (US and Canada) we have basically wobbly seats that move down a LOT in a crash, and you really want to be careful to keep your kiddos head contained in the seat shell to prevent whiplash injury or worse.
:thumbsup: