Tag Archives: blinded trials

5. How to Think Like a Scientist – Controlling the Chocolate Chip Experiment

phineas_scientist001So I learned how to make my mother’s delicious chocolate chip cookies. But after a few batches I got bored, and like a scientist, started to experiment. My mom didn’t enjoy walnuts and rarely baked with them, but I love them. I decided to change her recipe and add walnuts to my mix. Success! My roommates and I thought they tasted great (except for my friend Tony, who is allergic to nuts).

But were the cookies better than the original recipe? It was hard to know. It was hard to remember exactly how the originals tasted. So, thinking like a scientist, the next time I divided the cookie dough into two bowls and added walnuts only to one of them. I baked both batches and we did a taste test. The walnut chocolate chip cookies were clearly superior. And Tony was happy.

Talking like a scientist, the batch of original recipe cookies was the control group and the walnut batch was test group. The original ingredients are called dependent variables and the walnuts the independent variable.

The mark of a good experiment is that it is a controlled experiment with only one independent variable. You only want to change one thing at a time so you know what you are testing. The control allows you to compare the results of the change.

It is fine to do an uncontrolled experiment (like I did with my first attempt at walnut cookies). But if the results are good, it should be followed by a controlled experiment to verify them.

Scientists use uncontrolled experiments to generate ideas for new hypotheses. Hey, what if I throw in some jalapeño peppers into the cookie dough?

Fun Phineas Fact
In medical research, controlled experiments are called clinical trials. Over the years, it has been shown that the results are more accurate if the patients don’t know whether they are in the control group or the test group. We call these blinded trials.

Think about doing a soda pop taste test with your friends. Then try it again putting a blindfold on them before they drink. Blinding prevents personal opinions from altering results.

It has been found that results are even more accurate when the doctors and nurses running the experiment don’t know which group is the control. We call these double blinded trials. A double blinded controlled trial is difficult and expensive to run, but is the gold standard of medical research.

The US Food & Drug Administration requires double blinded clinical trials be run before approving new medications.