October 9, 2019
Pine Siskins by Bob Vuxinic
Welcome to our newest blog series Methods Behind the Madness. In this series, we’re addressing some questions that delve into the more data-driven side of Project FeederWatch. We hope that these blog posts make our scientific methods easier to understand, and spark critical thinking about FeederWatch data. The first question we’d like to explore is:
Why shouldn’t I start a count when I see an exciting species?
If you see a new or uncommon bird species, it’s completely understandable to want to report it on a FeederWatch count. We get it—some birds knock your socks off! It’s great to notice rare birds, but reporting counts only when you see exciting species misrepresents the birds at your feeders and makes it seem as if rare species are more common than they really are.
Let’s say we have three FeederWatchers, Lucy, Peter, and Maria, all of whom had Pine Siskins visiting their feeders over the course of a month (see illustration below). Lucy and Peter rarely saw siskins, but Maria saw siskins frequently. Let’s see what their counts would look like if Lucy and Maria counted in the correct way (i.e. chose their count days irrespective of siskins being present) and Peter counted in an incorrect way (i.e. chose to count only on days when he saw siskins).
As the illustration shows, Maria’s counts correctly indicated that she saw siskins frequently, and Lucy’s counts correctly indicated that she saw siskins rarely. However, because Peter did his FeederWatch counts only when he saw siskins, his counts misleadingly indicated that he saw siskins as frequently as Maria. Peter counted in what is called a “biased” manner—he biased his counts toward over representing siskins, making it look like Pine Siskins were common at his feeders when they were actually as uncommon as at Lucy’s feeders. Peter wanted us to know every time he saw siskins, but counting this way makes it impossible for us to get an accurate picture of where siskins are common and where siskins are rare. Of course, there is always a chance that a species will be over or underrepresented in a count just by chance. That’s why getting lots of counts from lots of participants is so valuable. Having a large sample of counts means that a few anomalies won’t change the entire picture.
We know it may be boring to count even when there aren’t “exciting birds” around, but making sure your counts are unbiased is part of what makes Project FeederWatch data valuable. We want to know where and when the rare birds show up, but we need to know where and when they aren’t around too. Otherwise, we wouldn’t know they are rare!
The next time you see an exciting bird outside of your chosen count days, don’t despair. Make a note in your journal, take a photo, and enjoy the fantastic sighting. Keep to your normal FeederWatch schedule and trust that your counts are valuable whether you have rare visitors or not.