Date: 10/30/20 1:42 pm
From: Robert O'Brien <baro...>
Subject: [obol] Re: eBird clarification
Hear, Hear!
To which I would just add, if one has literally a zillion eBird reports,
and 90% are discarded for a particular study.
that is still a VERY good average, because 10% of a zillion is still A VERY

On Fri, Oct 30, 2020 at 1:14 PM W. Douglas Robinson <
<w.douglas.robinson...> wrote:

> I have been off OBOL for about a year now, for a variety of reasons.
> However, Lars recently perpetuated a false statement about our research
> that I need to correct.
> We NEVER said 90% of eBird were worthless. In fact, a large part of my
> program is designed to learn more about the contributions of birders to
> eBird and finding ways to maximize what we can learn from those data. It’s
> a fascinating and intellectually stimulating part of our work.
> What we DID SAY was that we had to toss 90% of the eBird data that we had
> available at the time to answer one of our SPECIFIC QUESTIONS.
> That question was indeed part of Tyler Hallman’s thesis. The question was
> essentially: How do numbers of each species relate to the habitat at the
> location where observers were present? In other words, if I am at a marsh
> in June in Benton County, how many species and how many individuals of each
> do I expect I should find?
> Because most eBird checklists are Traveling counts, one cannot discern
> exactly where birds detected, so one cannot understand how habitat around a
> location influences the numbers of each bird species detected.
> So, those Traveling counts cannot be used to answer OUR SPECIFIC QUESTION.
> The issue of hotspots is related to this. Hotspots in eBird are currently
> point locations. If birders are not clearly located at exactly the pin
> dropped for the hotspot, then there’s too much uncertainty about where they
> were exactly to help us answer our specific question.
> BUT, the eBird data we did not use to address our question are certainly
> valuable for dozens or hundreds of other interesting questions.
> I won’t go on more because it should be obvious by now that we do NOT
> think eBird data are mostly worthless.
> If they were, I certainly would not have spent an immense amount of time
> and energy studying those data nor would I continue to invest heavily in
> contributing to it myself.
> So, please, let’s not keep stating this ridiculous conclusion about our
> research and our views of eBird.
> eBird is the best thing that has happened to birding since binoculars.
> Doug
> *******
> OBOL archives:
> Contact moderator: <obol-moderators...>

Join us on Facebook!