Moosehead Region Hunting Report: November 27, 2007

Last year about this time (actually a little prior to now), I was tasked with figuring out how many deer from WMD 14 would have to be examined to determine the “yearling frequency.” We examine the teeth up close to determine age.

Somehow, I knew the importance of that statistic from something I had read in the past; some biologist studying population dynamics had determined yearling frequency equals removal rate,. With removal rate, the population can be estimated. I took that assignment home, because statistical inference is a topic I’m not too conversant in. I was going to have to study this issue in a quiet atmosphere.

I skimmed thru an old text on statistics for the section that applied. I didn’t find it right away. Then I had a thought; couldn’t this situation be replicated using a table of random numbers such as odd = yearling, even = not yearling? Odd and even numbers presumably have an equal probability of being selected. And in some population ~50% of adults are yearlings. I worked with subsamples of 10, accumulating the result and studying the overall outcome. I determined this way that narrowly determining the percentage of yearlings in a population of unknown size was improbable without examining approximately 170 individuals drawn from that population. Regarding WMD 14, we could never expect to see that many shot by hunters, so the whole idea of trying was improbable. It wasn’t worth attempting.

After thinking more on this, I realized my trial was just one iteration, and that many more would be necessary to really determine what level of sampling would be necessary to generate an estimate in which you could put much confidence in (estimate +- say 15% and correct say 4 times out of 5). According to the textbook formula, the answer ranges between approximately 170 and 350, depending on how skewed the population is to the yearling age class. Who would guess such large numbers would have to be examined?
Science is very demanding.