May I suggest a lower cost alternative that is likely going to be more stimulating for your students . . .
Paper Helicopters
Many DOE trainers have used these to demostrate methods with great success. I’ve found the catapults to be too costly (as mentioned by others), too hard to travel with, too prone to breakage, and too predictable (especially with a bunch of engineers who have learned the math behind them).
Do a Google search using “experiment paper helicopter” and you’ll find a bunch of information. With that, and a little creativity of your own, I think you will find it useful.
Personally I find it usefull to start with a little two sample comparison (t-test), then move to a 3 or more sample comparison (ANOVA), then on to a factorial design. Eventually you can move all the way to a response surface design, and even engineers find it difficult to guess the interactions between types of paper, blade width, blade length, etc…
.
