actually...
it's best to hand it in early...
then, when you realise that you've done something wrong you can hand in corrections later...
Thats what I did at uni when I completly ignored scaling at the start of a graph I drew, and cam to a comclusion that the results were exponential rather than linear...
a couple of days later I handed in a rather rushed addendum that explain that I'd basically had a brain fart, and this is what the results and conclusions were meant to be, (when they were interperated properly!)... and that dragged my mark up from some 30% to around 80%.
I felt like I was the mutts nuts for handing it in early as well... thining that I could get on with other work, only to have to pretty much re-do everything except for the introduction and such...