Expert: Patrick Juola - 9/24/2004

Question
-------------------------
Followup To
Question -
I am a high school student performing an experiment to see the effects of carbohydrates and proteins on cognitive performance. I was hoping you could suggest a simple cognition test that I could use with limited resources. Possibly something with pictures or puzzles. Thanks a lot.
Answer -
You've accidentally stepped into one of the tar pits of cognitive psychology; whether there really is a single measurable quantity called "cognition" or "intelligence" or something like that, and how it can be measured.  The literature on this particular discussions goes back nearly forever, and no resolution is in sight.  If you want a good discussion of how it works, look at the book "The Mismeasure of Man," by the late Stephen J. Gould.

But assuming that you're willing to cut directly to the chase, almost anything similar to an IQ test or an SAT test or any of the other intelligence tests you've taken so many of in high school would work.  There are free IQ tests available on line (for example, at intelligencetests.com), or you could simply photocopy any of the various analogies tests or "get into Mensa" tests, or even a page of brain teasers.

From a scientific point of view, the important thing is to show an average score for the control group, for the experimental group(s), and then how these scores differ.  If you can get any "cognitive task" to show differentiation in scores, there's a lot of literature that suggests that the correlation between different tasks is very high, so results on a mathematical test should be similar to results on a verbal test.  Not identical, but similar.  Which should be enough for a "simple" test as you specify.

This enough information for you?

    Thanks so much for all your input so far. The study that I am doing takes place over three days. The first day is a learning day for the subjects to get to know the test (whatever it winds up being). On the second day the subjects will eat a carbohydrate-rich, protein-poor diet and will take another test (but it should be a different iq test of the same level, right?). On the third day the subjects will eat a carbohydrate-poor, protein-rich diet and will take yet another test. The tests from all three days should be different but of the same level of difficulty. Would it work if I did it with three different IQ tests? Thank you!

Answer
It should work if you did it with three different IQ tests, but you have to make sure that the IQ scores are comparable from all three.  Ideally, you'd want perfect correlation between the different tests.

An easier way to get the same effect (if you've got enough subjects) would be to mix the tests up among the three days.  So you can get three different IQ tests (A, B, and C) and give 1/3 of your subject pool the first test on the first day, 1/3 the second test, and 1/3 the third test.  The second day, you give the people a different test (ideally randomly; so if you took test A on the first day, you have a 50/50 chance of taking test B or C on the second), and then on the third day, each subject takes the one test they haven't taken.

This way, you get to directly compare the scores from each test.  Furthermore, you've got a built-in replication; you can see if each of the subgroups shows similar effects, or if one IQ test seems to be more sensitive to diet effects than the others.....