« Career vs. beer? | Main | Looking for junk DNA? »

Science Fair Secrets 2: Repeat Your Experiment

This is part of a series (copyright R Ford Denison) on the secrets of winning science fair projects. Click "science fairs" under Categories (at right) for more.

Like the first secret, this one is spelled out in the instructions to judges, but it seems to be a secret from the students. Don't tell anyone I let it out!

Here it is: do your experiment more than once. Two of the projects I judged last week had surprising and interesting results, but both got low scores because they only did the experiment once. Getting an interesting result once is almost meaningless. You think you're comparing apples and oranges, but you're really only comparing this apple to that orange. Maybe one had been in storage longer than the other, or something. Comparing three apples with three oranges is a little better, but if one apple was in storage longer, maybe they all were. So it would be better to come back to the store a month later, when you can be pretty sure they've gotten a new batch of each fruit.

Of course, this means you can't start your project the day before the fair. One guy complained that doing his experiment a second time would have taken a whole extra day. Poor baby! One of the winners worked on her project several hours a week all summer and into the fall. I know because she did the work in our lab.

Some experiments take months, because you're waiting for plants to grow or disposable diapers to decompose or something. In such cases, it may be very difficult to repeat the experiment, even if you start early. To some extent, you can substitute replication for repetition. For example, you could compare six organically grown apples to six conventional apples. But, as explained above, just comparing this batch (which happens to be organic) to that batch (which happens to be conventional) doesn't let you draw any general conclusions about organic vs. conventional apples, much less organic vs. conventional fruit in general. Ideally, you should compare organic Honeycrisp apples from farm A with conventional Honeycrisp apples from farm A, and also compare organic Keepsake apples from farm B with conventional Keepsake apples from farm B (or at least a neighboring farm), and so on. Then, if you get consistent results, you can be more confident there's a real difference between conventional and organic apples, not just a difference between specific farms or batches.

And what if there is no consistent difference? For example, what if organic Honeycrisp tasted better than conventional Honeycrisp, but conventional Keepsake tasted better than organic Keepsake? That's a perfectly valid and interesting result: there may or may not be a difference between conventional and organic apples, but other factors are also important. The important thing is to make sure your conclusions agree with your data. It doesn't matter whether they agree with your starting hypothesis.


Very interesting blog. There are some things in here that I would of never considered. I personally do notice a major taste difference when it comes to organic foods. The chicken has more flavor, the apples have more tart and the cookies even taste better. Perhaps some of this may be in my head, but I know that what I'm putting in my body is better for my health and for the environment.

Thank you

Great article!

Testing and retesting and then retesting some more is important and good scientific practice.

I am often amazed at how the set of data I collect in "one run" of an experiment is completely different in the next.

Very frustrating at times, however it serves to produce objective and hopefully reliable and credible scientific findings.

Considering the effect of random variables in experiments (even though we try to minimize these as much as possible) it pays to repeat experiments several times over.

Great post once again. Keep up the good work.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Type the characters you see in the picture above.