30 July 2009

Pardon me, are you going to use those beans...?

"Hunger is a powerful incentive to introspection."
. Willa Cather, "The Burglar's Christmas," Home Monthly, Dec. 1896

Today was 'mark-recapture' theory day in my Ecology II course. First-year students, and a preparation for our trip to Waterberg where we'll be doing game counts and a mark-recapture exercise with small mammals. 'Mark-recapture' describes the process of capturing, marking, releasing, and recapturing animals to obtain data used to estimate population size. Animals can be marked with tags, paint, bands, etc.

The traditional way to explain the notion of sampling a population with mark-recapture is to use dried beans. Students take a handful of beans (unknown number) and count out 20 beans. The students use a marker to place a 'mark' on the beans. Then the 20 beans are thrown back in to the original handful, mixed up (like animals milling around) and another sample of 20 beans are selected. In this second sample, some proportion of beans will be marked (say, 30%). If 30% of that sample are marked, it infers that 30% of the entire population was marked the first time and you can estimate the population size. For those of you playing along at home, you should try it. The formula is: Total number of beans = (number marked first time*total number sampled second time)/number of second sample that are marked. It actually works best if you take bigger samples than 20--try 40 or 50 each time. Same reason biologists struggle to get large sample sizes--more precise estimates.

Today, the exercise started out a little differently. One of the students, Theofilus, asked me, "Sir, why are you wasting these beans?"

"I'm sacrificing them to teach you a lesson," I told him, winking. Did I mention I really like being called "Sir" by students?

"They taste like nuts," he said, as he munched on one of the dried, white kidney beans.

"I think you are supposed to cook them for a couple hours, Theofilus," I said. "And, be careful, you're causing mortality to your population before we even get started!"

The exercise went very well, as usual. But, when it was over, I didn't get any of the beans back.

1 comment:

Tim Hiller said...

Can you show me how to use beans in a Bayesian sense?