In 1995 the Journal of the American Medical Association published a review of 23 comparatively rigorous studies conducted between 1982 and 1994. These were your classic controlled double-blind affairs: two groups of kids, one fed a bunch of sugar, the other given a placebo (i.e., artificial sweetener), everyone kept sufficiently in the dark as to who'd gotten what, etc.
The results? No discernible relationship between sugar ingested and how the kids acted. It didn't matter how old they were, how much sugar they got, what their diets were like otherwise — nothing. The JAMA authors stopped shy of drawing any definitive conclusions, but if there were a legitimate sugar-high effect out there, you'd like to see it turn up in the lab every so often.