Be careful when you generalize

One of the things we learned at the Xenophobia meeting we went to a few months ago at Arizona State University, was how quick humans are to generalize. We can learn to be careful when we apply those generalizations to individual human members of various groups.  But here I’m interested in how we generalize in science. When is it justified and when is it not justified?

At the most simple, I suppose you could use simple logic. If any single Dictyostelium clone does a certain behavior, or has a certain gene, then some must have it. But do they all? What can we really say from one clone? How can we have a whole meeting on a single clone? What if that clone is not the same, but has gone on a random walk as it has been left to evolve on the bench of hundreds of labs? Rob Kay did a genome analysis of bench evolution, to astonishing results. I am happy to say that the Dictyostelium community is now looking routinely at other clones and even other species.
If I say yeast is a certain way, is that based on one clone of one species? Does the Saccharomyces cerevisiae research community use one or many clones? Yeast actually means simply a single-celled fungus. My point is simply that you should be careful to know what you know and what you do not know. If you see something in one cell at one time, you will do a replicate. The replicate can be your attempt to do exactly the same thing, with the same materials. It can be an attempt to do the same functional thing with different cells, even from a different individual, so you can generalize to those individuals.

The very first step in statistics is based on what your case is. You design your experiments according to what you want to know at the most general. If you want to know what all mice do, then you must have multiple species and multiple samples. If you want to know what your dog Spot will do when it sees a cat, you only need to observe Spot repeatedly. You will need multiple cats, unless you just want to say what Spot will do when it sees Fuzzy, your cat. Just don’t assume you know what all dogs do to all cats. Think about this every time you do an experiment. Think about this every time you tell about someone else’s experiment.
Here’s an example also from Rob Kay’s lab. They say that Dictyostelium discoideum has 3 sexes, and know the genes to show it. I suppose there is really also a fourth that is homothallic, or self-fertilizing. If there were two sexes, I would be easily satisfied. But, I wonder, how many clones do you have to look at to be sure there are three sexes? Whatever they used is enough to say the number is not two. But how many clones do they need to sequence to say it is not four, or ten? I bet they can’t say that. I’m especially puzzled because one sex is more common than the others and we know a lot about sex ratio evolution.

So, be careful, be critical, always think about what the question is and what the sample size should be, but remember to be kind.

About Joan E. Strassmann

Evolutionary biologist, studies social behavior in insects & microbes, interested in education, travel, birds, tropics, nature, food; biology professor at Washington University in St. Louis
This entry was posted in Experimental design, Scientific methods and pitfalls and tagged , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.