[texbirds] Re: eBird, science, and carrots (a bit long)

From: "John Arvin" <jarvin@xxxxxxxx>
To: <mtheindel@xxxxxxx>, <texbirds@xxxxxxxxxxxxx>
Date: Sat, 30 Mar 2013 11:21:59 -0500
Matt,
You are right on all counts. I have always felt that the "data" generated by 
CBCs only had value by virtue of its long run so that statistic analysis could 
be used to smooth out the bumps. All data bases are only as valuable as the 
date input into them. Far too often that is shaky at best. I agree the eBird 
reviewers are minor saints. Not that they are infallible but at least they do 
represent an effort to verify the data of the system. Likewise for NAB 
compilers, etc. Still we get a few mid-March Northern Waterthrushes and similar 
reports from observers whose knowledge of status and distribution is feeble at 
best. Once these things get into print (whether real or literal) they are 
really hard to erase.
jca
John C. Arvin
Research Associate
Gulf Coast Bird Observatory
103 West Hwy 332
Lake Jackson, TX 77566
jarvin@xxxxxxxx
www.gcbo.org

Austin, Texas 

----------------------------------------
 From: "Matt Heindel" <mtheindel@xxxxxxx>
Sent: Friday, March 29, 2013 12:36 PM
To: texbirds@xxxxxxxxxxxxx
Subject: [texbirds] eBird, science, and carrots (a bit long)

I find some of the posts regarding eBirdinteresting, as this medium represents 
different things to different people,and many of the posts are written as if 
eBird can or should only be as the writerframes the issue. But, we see this 
differently. Indeed, there are some of uswho continue to be concerned over the 
quality of the data and as the mass increases(more people submitting- even with 
the best of intentions), the more thisproblem grows. 

If it were not for incredibly diligentreviewers, eBird would be a flaming mess. 
These folks, unpaid, have to spend anamazing amount of time dealing with this 
ever-increasing number of submissions,many of which are submitted by folks who 
need to sharpen their identificationskills, better understand status and 
distribution, and embrace the concept of aclean database. Of course, they do 
not think they are making mistakes and arelikely being earnest in their 
efforts, but â¦..And by the way, if the reviewerscould just toss out 
sightings, I think being unpaid volunteers is right, but Iknow many of these 
reviewers who go to a lot of effort to first coax moreinformation, coach people 
through identification issues, etc, and to do this canbe monumental. It might 
only be 3,4, or 5 emails for one submitter, but it ishours and hours for the 
reviewer if they are trying to help people along. And,if they are not trying to 
help people, the observers can get surly in theirattitude about the p
rocess. (Actually, I have had several examples forwarded tome where the 
submitters were plenty surly even with a helpful reviewer.)

I was particularly struck by this  recent eBird commentary:
There needs to be a carrot to encourage submittingdata, and that carrot is
seeing your own data appear on the map.  Ebird is a scientific database,
not a listers tool or a piece of software run bythe American Birding
Association. It appears to me that some very important data is not being
gathered.

First, eBird is not science: just becausepeople say it is, does not make it so. 
Let's not confuse the point. The databasehas value (and great potential), but I 
think calling this science is a problem.(In this post it is just the database 
being called scientific, but there aremany cases where people call eBird itself 
science). Perhaps that is a goodgoal, but I think it is wiser to change the 
verbiage (and the debate) to focuson the value of the data, the shortfalls, and 
how these shortfalls might beaddressed. If some think this is science, ok, it 
is your right to use anamorphous definition, but we won't get far on that 
discussion.

As with every bird sighting repository, forexample North American Birds and its 
earlier forms, there are mistakes and anadvantage of eBird is the ability to 
correct them at some point, much harder todo in the old print version. But, as 
the increasing number of sightings areentered, the borderline identifications, 
or borderline early and late dates areeasily obfuscated. I have previously 
complained (in an area roughly San Antonio and pointseast) over fall 
Orange-crowned Warblers (for nominate orestera, generally lateSeptember at the 
early end) and Black-chinned Hummingbirds (ad males usuallygone in Aug; most 
gone by mid-Sep) but have been told by others that"there are a lot of records 
in eBird, so what is the problem?"  Yup, if they were there, then that must 
bewhat really happenedâ¦.

In the above clip, I see that the databaseis said to be scientific and yet see 
that the very normal and human emotionenters into the fray (gotta have that 
carrot, etc). It is the emotional aspect,a typically (and understandable) human 
trait, that is not the friend ofscientific data. 

When I talk to the key people involved witheBird (all of whom are good people, 
good birders, and well intended), they maketwo points that are correct but miss 
my criticisms. First, we can always goback and edit out things we find to be 
mistakes. True. Got it. But, it is notthe obvious mistakes that concern me- 
it's the little ones as those are harderto differentiate. Second, we can sort 
data any way we like to get a"real" picture. The problem with this solution is 
you have to knowthe answer before you can ask the question. As an example, back 
to myOrange-crowned Warbler: I could only include OCWA after 25 Sep, and look 
intoall earlier claims to decide if they should be included in some 
fictionalwrite-up I was doing. But, that is because I know orestera should not 
yet behere. For others, they will look at the graph and see that there are a 
lot ofreports and so that greenish thing they had here on Labor Day must be an 
early OCWA. Even well-intendedbirders who want to see if the
ir suspected sighting is unusual might use eBirdas a gauge, but if not governed 
by a good reviewer, they could be reading baddataâ¦â¦And, in this area people 
trying to be helpful often post their eBird liststo Texbirds so even if a 
reviewer has nixed a sighting after the fact, otherbirders will not necessarily 
know that and could then more likely make asimilar mistake. If we had great 
reviewers everywhere, we could better scrubthe data but a) this is not yet 
possible, and b) it takes a lot of time bythese volunteers. 

We have different objectives with regard toeBird and only the opinions of those 
in charge really count. A simpledelineation in these objectives might be speed 
and tranquility (happysubmitters) on the one side versus quality on the other 
side. It is quite frustrating forsome of us to see an unconstrained eBird being 
pushed by many people that emphasizeseeing their data fast (why aren't my 
sightings showing up right away?), aswell as the ever-present battle over 
whether someone did or did not have whatthey claimed. It is human nature. I get 
it. And eBird is with us to stay. I getthat, too. But, as we waddle down this 
road, there are opportunities to shiftthe value of the database and, for those 
most interested in the quality of thedata, we need to find ways to push emotion 
and other human frailties aside.This problem is not unique to eBird either- 
just ask any NAB compiler aboutgrief they have received from birders not having 
a record included. Thehuman/emotional side of this p
assion is a real challenge, but if we can placethe emotion into proper 
perspective, we might one day approach something that might,maybe, resemble 
something scientific.

Matt Heindel
Fair Oaks Ranch, TX

Edit your Freelists account settings for TEXBIRDS at 
//www.freelists.org/list/texbirds

Reposting of traffic from TEXBIRDS is prohibited without seeking permission 
from the List Owner



Edit your Freelists account settings for TEXBIRDS at 
//www.freelists.org/list/texbirds

Reposting of traffic from TEXBIRDS is prohibited without seeking permission 
from the List Owner
[texbirds] Re: eBird, science, and carrots (a bit long)

Other related posts:
