If you pay close attention to this series on using gummi bears to understand design of experiments, you noticed that in my last post I mentioned pressure as a variable for the first time. Pressure wasn’t on the fishbone diagram that I used when planning variables, even though it’s just as obvious as temperature and humidity.
I’ve been referring to the fishbone diagram quite a bit, but I didn’t spend much time creating it nor did I gather anyone else’s input. Therefore, there are probably lots of unconsidered variables that could make a difference in how far the gummi bears traveled. I can think of several quickly: rubber band thickness, rubber band circumference, popsicle stick thickness, popsicle stick length, and popsicle stick width. Some of these variables seem like they could matter a lot. Will they ruin the experiment if we don’t include them?”
One of the most important things I can do in design of experiments is to randomize the order of data collection. Randomization gives me the best chance that the order of data collection doesn’t follow a pattern for a variable I’m not measuring. If all of my most tightly wound rubber bands also corresponded to my widest popsicle sticks, then I might conclude that the number of rubber band windings has an important effect, when that’s only part of the story.
The easiest way to avoid a pattern is to use randomization. It’s still possible that, by chance, we’ll end up following a pattern anyway. However, if we run a repeat of the experiment in a different order, then the experimental runs won’t follow the same pattern—the randomization lets us distinguish the variables that we haven’t kept track of from the ones that we have.
So how do we randomize in Minitab? The nice part is that Minitab will do that for you automatically when you design an experiment. We’ll do it soon to plan the gummi bear experiment. If you have a few minutes and want to jump ahead, check out how Eston Martz used design of experiments to plan a better pierogi!