Every so often I get a little tweaked over the hype associated with eBird. Yes, it's a useful endeavor. Yes, it's actual science - but it's science with a large error level in the data, and the error isn't trivial to estimate. It has the sort of error level I associate with a badly-controlled Sociology experiment. Speaking as a professional scientist in the biological sciences, I'd like to see the quality of some of those journals - the "peer reviewed publication" in Nature below is just a News article - not peer reviewed and just fluff. That's either sloppy or disingenuous. The P.N.A.S. article might be worth a look at - that's the second best journal in that list.
But "Bulletin of the American Society for Information Science and Technology". Since Monday Night Football is on in the background as I write this: Who are we kidding ? C'mon man. I made the Swiss Cheese comment: Ben C. and I were using eBird to look at Prothonotary records with a view to potential breeding populations north of NYC (there are some on the nw of NY State). Now I actually *know* at least some of the Prothonotary records in NYC. Many of these are simply not present in the database. Cornell doesn't mine this list or eBirdsNYC for those records, it relies on contributors which are a small subset of the birding population to report. NYSBirds and eBirdsNYC have the same thing - a small minority report sightings. However those lists don't purport to be a representative record of sightings. Other comments like: "robust verification" are largely meaningless if you have to say: "is it getting better every month". Robust verification would require actual checking of every reported sighting, not flagging the most error-prone observations using a very simple probability model. I wonder how many House Finch sightings are verified ? The actual counts ? So if we can skip over all the PR here - eBird *IS* currently not very useful for constructing an NYC checklist. It's currently a very incomplete record and fundamentally flawed as a result. I expect it will get better, and I certainly hope it will get better, but that doesn't mean that it already qualifies as "good". It only qualifies as "good" if you have very low standards indeed. Phil Jeffrey, D.Phil. On Mon, Nov 1, 2010 at 8:12 PM, Andrew Farnsworth <andrew.farnswo...@gmail.com> wrote: > Hi all, > I want to comment on some points in the checklist thread about eBird > and lists - the eBird team can speak to issues about how to generate > lists and give much more detail than I, but I want to discuss comments > relevant to science and eBird and what is an is not science. To speak -- NYSbirds-L List Info: http://www.NortheastBirding.com/NYSbirdsWELCOME http://www.NortheastBirding.com/NYSbirdsRULES ARCHIVES: 1) http://www.mail-archive.com/nysbirds-l@cornell.edu/maillist.html 2) http://birdingonthe.net/mailinglists/NYSB.html 3) http://www.surfbirds.com/birdingmail/Group/NYSBirds-L Please submit your observations to eBird: http://ebird.org/content/ebird/ --
