One of my pet peeves with science fairs is that they often confound experimental science with engineering projects--often to the disadvantage of both. I've seen some pretty cool engineering projects not recognized at all in science fair judging because there was no space for anything without a classically-testable hypothesis and repeated measures (I am, I should note, all for the empirical method and reproducibility). If you build a fabulous robot, you should get some credit for it, even if you can't easily make a graph or table with at least three trials for each condition. I think science fairs should have a science and an engineering category, judged separately. I've never seen this below the high school level.
And I agree, I like to see a project that exhibits the student's understanding of the scientific method.
