Date: 7/16/17 10:44 pm From: David Starrett <StarrettDA...> Subject: Re: Reporting accurately on eBird
Thank you for responding. Yes, we have birded together a couple of times, last year at Bradford farms we got some lifers with you. You also gave a good tip on finding both wills. With my dad's sightings, I had them in excel and formatted them for upload. All went in as historical. I uploaded for two reasons. One was to document his sightings and have an ongoing place for his life list. Second was contributing hsi 60 years of sightings to scientific knowledge. The first one is completely met, as you say, the unverified signings still count on his list. The latter was met as well, though I do think eBird is missing out by being selective about the data being contributed.
I would note that none of the flags came from any Missouri data, most of it was international sightings with a few US like the Florida Scrub Jay I illustrated my point with.
From: Robert Jacobs <robertbjacobs...>
Sent: Sunday, July 16, 2017 10:17 AM
To: David Starrett
Subject: Re: Reporting accurately on eBird
Hi Dave and MOBirders,
Bird observation data from years ago are extremely valuable, probably more than many eBird users know and even eBird reviewers have thought about enough to formulate a good response to folks like Dave Starrett. Dave, I have met you a couple of times and haven't welcomed you properly as you moved from Cape Girardeau to Columbia. I hope we can do some more birding together. Bill Eddleman told me about your move before it happened. Welcome and thanks for your eBird efforts.
I don't intend to throw anyone under the train because they "rejected" my checklist or just one of my observation. In fact, rejected is a word that is not used in the eBird review process. (http://help.ebird.org/customer/en/portal/articles/1055676-understanding-the-ebird-review-and-data-quality-process?b_id=1928) The eBird help button has an answer for just about any question that might think of. As records are flagged (you know you have been flagged if you can't get your app to send the checklist until you do something) and the checklist submitted, the flagged species will be seen on a separate list for the eBird reviewers and are marked "not-valid" or unreviewed" on the spreadsheet of flagged data records.Once a reviewer acts on the record it will will be considered reviewed. Reviewers can make the record as confirmed or unconfirmed. Unconfirmed may mean they will send an email to the observe with additional questions.
Reviewers have access to the filters of which there are 20 plus in Missouri and will likely be more information is supplied by observers. The comment sections by each flagged species is the observer's opportunity to help the reviewers with local information that is more accurate than the initial filter setting. This spring was a nightmare with shorebirds and ducks in abundance. I hope that some of the filters will reflect the new abundances for the Boone County filter. Yes, Boone County has its own filter and we all need to help the reviewers set it to the right level. A number of species will be set at Zero (0) as there is a need to learn more about the species, it may be a hard to identify species or just a casual visitor to the county.
With entering data from another person who has deceased, the eBird requirement of being able to speak to the observer kicks in. If the data were entered as traveling or stationary it wouldn't matter, without a live observer to talk to, one needs to enter data under Historical. Historical data often is missing some of the other requirements such as starting time, duration, and/or area or miles covered.
Historical data is extremely valuable and should be enter in a way that everyone know how to use it. I have been given a pile of field notes from several eastern Missouri counties from the 30s and 40s that I hope someone can enter. I still have many boxes of data from three continents that I need to enter. Fortunately I kept a journal that has most of the information needed for traveling or stationary and I am still around. But, I will have to enter under Historical for those data that are lacking all the required data such as duration, area or miles surveyed. It is very easy to change each checklist to Historical by going to My Checklists and clicking on the Edit Date and Effort and change to Historical.
At no point in time are the data that you enter rejected or deleted by eBird. Your data will always be accessible by your self. and you can make it available to someone if you want. Data that the reviewers consider invalid will not go into the science database but will always be in your personal checklists.
I hope this helps a little to get the cogs moving toward making the observer-reviewer interface work more smoothly. Having been a reviewer many years ago with the Great Backyard Bird Count. It was a massive job and often meant checking every record for house and purple finch errors, Pine Siskins and female purple finches, and many, many common mistakes made especially by birders in their first few years of observation.
Take photos help immensely in getting a record through the review process, especially when know one know your skill levels. I birded for three years in Arizona and submitted records for one of the statewide birding club newsletters. I don't think I ever saw one of my records appear in the newsletter until after I got a call one day asking me to lead a field trip in my area, which was remote from much of the rest of Arizona. It was a great time for me being able to meet about 25 of the top birders in the state that needed some new lifers that I had been reporting over the the three years. Fortunately, I was able to get four lifers within a half a square mile for most of the the people on the field trip. Harris's Sparrow, Black-capped Chickadee, Northern Shrike and American Tree Sparrow.
The next issue began posting the records I and a few others were submitting. I need to start entering my data from NE Arizona. But first I will figure out how to contact the reviewer for that region, Eric Hough, who apparently lives in western Missouri now.
So this brings up another issue with eBird. I have had this conversation with a number of ebird reviewers and none could satisfy me. My concern goes like this.
At some point, data started being submitted to eBird. Everything was "accepted". As the data filled up at some point a filter was turned on. New data that didn't match/fit was rejected or required proof. So, any data that went in originally was assumed accurate, but later data was filtered. There is no inherent reason the first data is any more accurate than latter, it is just the order of entry. I am oversimplifying, but to make a point.
I have had this manifest itself more than once. My father had detailed records of 8,600 sightings spread out over 50 years on all seven continents. He passed away 10 years ago. I painstakingly uploaded all his sightings to eBird. My dad was a field biologist. He knew how to take accurate and detailed notes. I trust his sightings to be more accurate than mine. I received dozens and dozens of ebird reviewer flags. They asked questions I couldn't answer since he saw the bird not me. So they rejected the submissions. Many times they were asking about data 20-40 years old. They were using current range maps and counts. When I asked if it was possible that when he saw the bird back then the range was different or the population density was higher, etc. 100% of the time the answer was no.
The classic case for me was Florida scrub jay. He had a sighting from 1945. It was not in the current range, but not far from it. And in eBird, a couple of the older sightings were at the edge of the range near my father's sighting. Submission was rejected. When I suggested to the reviewer that rather than assume my dad's sighting as in error consider that he might have just provided valuable historical data about the Jay's range. We had a back and forth about this and eventually the reviewer, somewhat rudely even, dismissed the submission as unreliable and let me keep it on my dad's list, but out of the actual eBird data. I am sure of the sighting and location. eBird has rejected data that could support a changing range. That sighting is 70 years old. We know bird ranges change, they are changing now. Why is the assumption the data is invalid rather than valuable.
I am bothered by this first in data closes the door behind them approach. I realize there is much more to it than this and I am oversimplifying, but it does seem that maybe the method of validating data has some flaws to it.
I went to Eagle Bluffs this morning determined to count every Indigo Bunting, Dickcissel and Common Yellowthroat. I ended up with 38 Indigo Buntings, 15 Dickcissels and 17 Common Yellowthroats. I'm sure I missed some, sometimes the brain just tunes out or several sing at once to confuse the brain, my brain at least.
And about 3/4 way through the refuge I noticed I only had 1 hr left before I had to leave and I hadn't been to Pool 14 or 15 yet, so I had to quit stopping to tally every one of those. My point is, I think there are many reasons why the counts are not perfectly accurate, but we do the best we can.
Kathleen Anderson, Columbia