If you’re unhappy with how your cookies turned out, don’t know which ingredient or which process variable caused the trouble, and don’t have the time, ingredients, or test eaters to vary each variable individually: statistics to the rescue! Done right, design of experiment allows an appreciable reduction in trials without losing the information on which variables have a significant effect on the process. Doing it with a cookie recipe provides an accessible illustration of how it’s done. In this case, the trial runs were reduced by at least a third, depending on how an individual variation experiment might be set up.
One of our friends in college always wanted to do a mini research project like this. We had another friend who was always forgetting what we considered critical ingredients in her cookies and they always came out good. And then there are the preferences – soft vs. crunchy cookies, etc. Fun that someone actually did it! And it does confirm our findings of the dream cookies, as Mom mentioned.
If any of you want to replicate the experiments, Minitab offers a 30-day trial of its software. Or you can tell me which variables you want to investigate (and which values of those variables) and I can clock out at work and do a quick DOE for you.