Experimental design - Deciding what
to measure
Due: Wednesday, September 10
Write concisely - One side of a page.
Read the article by Jared Diamond, "Pearl Harbor and the Emperor's
Physiologists", Natural History, Dec 1991,
pp. 2-8.
1. Diamond discusses several experiments designed to
elucidate why individuals differ in how comfortable they feel in warm climates.
Although all the researchers wanted to study thermoregulation, they
didn't all choose the same variables to measure. List all the variables that were
measured to give some indication of people's ability to regulate
temperature. Write just a list, not an essay. For
example, if you were studying individual differences in eating behavior, you
might choose to measure one of these variables:
- number of meals per day
- number of calories per day
- number of calories of fat, protein,
carbohydrate
- number of different kinds of foods that are
eaten
- etc.
2. The decision of which variable to
measure will affect the conclusions that can be drawn from the data
collected. In the previous example, you may conclude that two people ate
the same amount (if you found that they both ate three meals a day) or very
different amounts (if you found that one ate 300 calories in each meal, and the
other ate 1000 calories in each meal). Many other aspects of the
experimental design can similarly determine what conclusions can be drawn from
the data. When scientists report results in primary research articles,
these details of experimental design are described in the Materials and Methods
section, but these details are usually omitted from popular science articles in
newspapers and magazines. When reading popular science like this, you
should always be thinking, "How did they do this experiment? What did
they measure?"
Consider the experiment described in the paragraph
beginning "Measurement of sweat's salt concentration". Diamond says
that these experiments showed that Japanese resorb less salt than do the other
groups. But since he hasn't presented all the details of the original research
design,
you may not be willing to accept this conclusion. What questions would you want to ask about how the experiment was
done in order to be certain of this conclusion? How might the answer affect the likelihood that the given conclusion is
true? List three questions you'd want to ask, and for each one explain why
the answer might affect your willingness to believe Diamond's conclusions.
Here's one example:
a) How old were the people they studied?
b) If the people in one country were older than the people in another country,
they might have really been measuring a difference related to age, rather than
one related to nationality, and we know that elderly people have more difficulty
thermoregulating than younger people.