Birdingonthe.Net Recent Postings from Bird Records Committee Forum > Home > Mail > Alerts Updated on Friday, July 18 at 08:34 AM ET The most recently received Mail is at the top.	Tufted Puffin,©Shawneen Finnegan

18 Jul Re: technology and BRCs ["K. Dean Edwards" ] 17 Jul Re: technology and BRCs [William Rowe ] 17 Jul Re: technology and BRCs ["George L. Armistead" ] 17 Jul Re: technology and BRCs [Alan Wormington ] 17 Jul Re: technology and BRCs [Mark Stevenson ] 17 Jul technology and BRCs [Mark Lockwood ] 13 Jul Re: Question on protocol in records committee journal reports [Richard Heil ] 13 Jul Re: Question on protocol in records committee journal reports [Ted Floyd ] 3 Jul Re: Question on protocol in records committee journal reports [Richard Heil ] 3 Jul Re: Question on protocol in records committee journal reports ["Paul A. Guris" ] 3 Jul Re: Question on protocol in records committee journal reports [Ted Floyd ] 2 Jul Re: Question on protocol in records committee journal reports [Richard Heil ] 3 Jul Re: Question on protocol in records committee journal reports (RESEND) [laurent raty ] 2 Jul Re: Question on protocol in records committee journal reports (RESEND) [Ted Floyd ] 2 Jul Re: Question on protocol in records committee journal reports [Ted Floyd ] 2 Jul Black-bellied Whistling-Duck sighting request [Phil Davis ] 1 Jul Re: Question on protocol in records committee journal reports [Buford Myers ] 1 Jul Questions for Ted on Records Probability ["Paul A. Guris" ] 1 Jul Acceptance/non-acceptance of records, Type I & II errors [David Christie ] 1 Jul Re: Question on protocol in records committee journal reports [Laurent Raty ] 1 Jul Re: Question on protocol in records committee journal reports [Ted Floyd ] 1 Jul MD/DC Records Committee web page updates and Review List changes [Phil Davis ] 30 Jun Re: Question on protocol in records committee journal reports [Rick Fridell ] 30 Jun Re: Subject: Re: Question on protocol in records committee journal reports [Geoff Malosh ] 30 Jun Re: Question on protocol in records committee journal reports [Richard Heil ] 30 Jun Re: Question on protocol in records committee journal reports [Ted Floyd ] 30 Jun Re: Fwd: Re: Question on protocol in records committee journal reports [Alan Wormington ] 30 Jun Re: Question on protocol in records committee journal reports [Kimball Garrett ] 30 Jun Fwd: Re: Question on protocol in records committee journal reports [Phil Davis ] 30 Jun Re: Question on protocol in records committee journal reports ["Donna L. Dittmann, ddittma" ] 29 Jun Re: Question on protocol in records committee journal reports [Phil Davis ] 29 Jun Re: Question on protocol in records committee journal reports [Alan Wormington ] 28 Jun Question on protocol in records committee journal reports [Martin Meyers ] 15 May "Work in progress" news from the British Birds Rarities Committee [Phil Davis ] 12 May RFI: North American albatross references [Phil Davis ] 2 May Re: Extralimital Mottled Ducks [K Dean EDWARDS ] 2 May Re: Extralimital Mottled Ducks ["Spector, David (Biology)" ] 2 May Re: Extralimital Mottled Ducks ["Spector, David (Biology)" ] 2 May Re: Extralimital Mottled Ducks [William Rowe ] 2 May Extralimital Mottled Ducks [Alan Wormington ] 7 Apr Re: BRCs and "umbrella" organizations [Martin Meyers ] 6 Apr Re: BRCs and "unbrella" organizations [Joseph Morlan ] 6 Apr Re: BRCs and "umbrella" organizations [Phil Davis ] 3 Apr Re: BRCs and "unbrella" organizations [Alan Wormington ] 3 Apr Re: BRCs and "unbrella" organizations [Mark Lockwood ] 2 Apr Re: BRCs and "unbrella" organizations [William Rowe ] 2 Apr Re: BRCs and "unbrella" organizations [] 2 Apr Re: BRCs and "unbrella" organizations [Bill Sheehan ] 2 Apr umbrella organizations ["kratter AT flmnh.ufl.edu" ] 2 Apr BRCs and "unbrella" organizations [Steven Mlodinow ] 3 Mar Re: Compilation of North Atlantic albatross sightings - data call [Phil Davis ] 29 Feb Re: Compilation of North Atlantic albatross sightings? [Steven Mlodinow ] Subject: Re: technology and BRCs From: "K. Dean Edwards" <kde AT ANGST.ENGR.UTK.EDU> Date: Fri, 18 Jul 2008 09:34:36 -0400 We have actually had a few reports come in from folks that aren't really even casual birders who just saw an odd bird, grabbed their digital camera, went online and found the TOS website, and emailed the photo to us to ID. One such instance was an odd bird coming to a lady's feeder which turned out to be the first winter record of Painted Bunting for the state and first accepted record for East TN anytime of the year. This bird has now returned to the same area three winters in a row and has been seen by several folks. Just this morning we got a similar report (no photos this time) of a Swallow-tailed Kite (Review-list species) in Knoxville. Unfortunately, another trend we're seeing is report forms coming in with the only thing written under "Description:" being "See photo" ... sometimes even from folks that know better. Dean Edwards Chair/Secretary, Tennessee Bird Records Committee -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: technology and BRCs From: William Rowe <rowe AT TJS.ORG> Date: Thu, 17 Jul 2008 19:26:15 -0500 Mark et al.: I concur, although I'm not sure that the percentage of reports we are getting with photographs is quite as high as yours—still, it's higher than in the past for sure. Our two most recent additions to the state list—Mew Gull in April, Broad-tailed Hummingbird in June—both came in with a nice set of photographs that left no doubt. Echoing Alan's point, in the case of the Mew Gull the photos were taken by someone other than the observer, someone who might not have made it there had the report not been called in and posted on line quickly. I'm not sure that the average quality of written descriptions has gone down. It wasn't all that high to begin with. There has always been a disparity between the minority who really get how to describe a bird and expend the effort to do it well, and the majority who either don't quite get it or don't put in the effort. Again, in the case of the Mew Gull, we got a thorough, detailed writeup from the finder in addition to the good photos from someone else. Best of both worlds. Bill Rowe St. Louis, Missouri On Jul 17, 2008, at 3:35 PM, Mark Lockwood wrote: > Dear all, > > > > With all of this discussion about certainly and uncertainty, I do > know one thing with what I think is a 100% certainty. That is that > the advent of digital cameras have changed the way the Texas BRC > works in a very short time. I was looking at the uncirculated > records before the TBRC and to my amazement 28 out of the next 33 > records that are ready to go into circulation have at least one > associated photo. I am sure we can all recall when that number was > closer to, if not below, 50%. As time goes by I am also seeing an > increase in the overall quality of the images provided as well. > This has resulted in an upward shift in the percentage of accepted > records since for some records it shifts the burden away from the > reporter’s ability to provide detailed descriptions that eliminate > congerers and other similar species. It would be nice, but > probably too much to ask, for digital audio recorders and higher > quality microphones to become available at more reasonable process > as well. > > > > This is not anything revolutionary or anything, but it was kind of > a moment when I looked at folders and saw all of the photos lined up. > > > > Mark > > > > > > Mark Lockwood > > Secretary, Texas Bird Records Committee > > 402 E. Harriet Ave. > > Alpine, Texas 79830 > > mark.lockwood AT tpwd.state.tx.us > > > > Visit the Texas Bird Records Committee at http://texasbirds.org/tbrc/ > > TOS Handbook of Texas Birds at http://www.tamu.edu/upress/BOOKS/ > 2004/lockwood.htm > > TOS Occ. Publications at http://www.texasbirds.org/publications/ > occasional.html > > > > -------------------------------------------------- Bird Records > Committee Forum archives: http://listserv.indiana.edu/archives/brcf- > l.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: technology and BRCs From: "George L. Armistead" <georgearmistead AT HOTMAIL.COM> Date: Thu, 17 Jul 2008 18:39:53 -0400 All, I think it was Al Jaramillo who told me that some study showed that no new technological device has been so readily accepted and adopted by the American public as the digital camera. Much faster than say computers, cell phones, mp3 players, etc. (Hopefully folks will come around on recording equipment. I love my new Sony PCM D50 which runs about $600, and is real easy to use.) Alan is absolutely right that fewer written descriptions make it into committee member hands as used to be the case. Mostly, I don't think that is a bad thing. As he indicates committee efficiency (and perhaps veracity?) have been improved w/ the technological improvements. Certainly there are instances where submitters don't realize the value of a written description. We've all seen what are probably solid records go unaccepted b/c the photos aren't diagnostic or as supportive as the observer thinks they are. Especially with tough ID situations, with few or no words to back them up, they don't leave voting members with much to work with. Mostly though we are coming out ahead. I can think of several observers who can't write to save their lives, but they can sure take photos, and since they've gotten cameras they've become more valuable contributors. Prior to that nobody was quite sure what to make of their reports, given their writing. The digital age has allowed them to contribute more productively and solidify their reputations. Best, -George George L. ArmisteadPhiladelphia, PAgeorgearmistead AT hotmail.comhttp://www.flickr.com/photos/seabirder/Field Guides, Inc.Birding Tours Worldwidehttp://www.fieldguides.com/ ---------------------------------------- > Date: Thu, 17 Jul 2008 17:41:26 -0400 > From: wormington AT JUNO.COM > Subject: Re: technology and BRCs > To: BRCF-L AT LISTSERV.INDIANA.EDU > > Mark and all, > > I could not agree more. > > I have in front of me the soon-to-be-published 2007 Annual Report of the > Ontario Bird Records Committee. > > Of 122 records that are accepted, 77 records (63%) have photos as part of > the documentation. This equates to almost 2/3 of all records having > photos! > > For comparative purposes, I randomly pulled out the OBRC Annual Report > for 1986. In that report there are 132 accepted records, of which only > 37 (28%) include photographs. > > So Mark's observation is certainly real -- more people own cameras than > ever before. Despite this, I believe there is yet another variable as to > why more rarities are being photographed. And that reason is the > following: in this age of instant communication, word gets out much > faster about the presence of a rarity. The result is that a rarity is > likely to be observed by many more individuals than compared to the past > -- more individuals, more likely the bird will be photographed. I can > easily compare this modern scenario to the past, where many rarities were > seen by the finder only. After all, that person had to drive home to > make telephone calls, and there was no guarantee that others would even > be home to take the call. So a whole day might go by before anyone else > looks for the bird, assuming it is even still present. > > Despite the blessing of more records with photographs, in contrast I have > seen a different (but depressing) trend over the years. And that > pertains to written reports that are clearly of much poorer quality > compared to those of the past. Today a records committee is lucky to get > a single paragraph of description. In the past it was not unusual to get > multiple pages of written description to document a rarity. Why this is > happening could be debated, but I think the core reason is that most of > the newer, somewhat less-experienced birders nowadays just don't see the > big picture. Or they are more interested in bird listing rather than > bird study. Who knows. > > Cheers, > > Alan Wormington > Assistant to the OBRC Secretary > > > ************************** > > > On Thu, 17 Jul 2008 15:35:05 -0500 Mark Lockwood > writes: >> Dear all, >> >> >> >> With all of this discussion about certainly and uncertainty, I do >> know >> one thing with what I think is a 100% certainty. That is that the >> advent of digital cameras have changed the way the Texas BRC works >> in a >> very short time. I was looking at the uncirculated records before >> the >> TBRC and to my amazement 28 out of the next 33 records that are >> ready to >> go into circulation have at least one associated photo. I am sure >> we >> can all recall when that number was closer to, if not below, 50%. >> As >> time goes by I am also seeing an increase in the overall quality of >> the >> images provided as well. This has resulted in an upward shift in >> the >> percentage of accepted records since for some records it shifts the >> burden away from the reporter's ability to provide detailed >> descriptions >> that eliminate congerers and other similar species. It would be >> nice, >> but probably too much to ask, for digital audio recorders and higher >> quality microphones to become available at more reasonable process >> as >> well. >> >> >> >> This is not anything revolutionary or anything, but it was kind of a >> moment when I looked at folders and saw all of the photos lined up. >> >> >> >> Mark >> >> >> >> >> >> Mark Lockwood >> >> Secretary, Texas Bird Records Committee >> >> 402 E. Harriet Ave. >> >> Alpine, Texas 79830 >> >> mark.lockwood AT tpwd.state.tx.us >> >> >> >> Visit the Texas Bird Records Committee at >> >> http://texasbirds.org/tbrc/ >> >> TOS Handbook of Texas Birds at >> >> http://www.tamu.edu/upress/BOOKS/2004/lockwood.htm >> >> TOS Occ. Publications at >> http://www.texasbirds.org/publications/occasional.html >> >> >> >> >> >> -------------------------------------------------- >> Bird Records Committee Forum archives: >> http://listserv.indiana.edu/archives/brcf-l.html > > "The Early Worm Gets The Bird!" > --- Alan Wormington > > -------------------------------------------------- > Bird Records Committee Forum archives: > http://listserv.indiana.edu/archives/brcf-l.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: technology and BRCs From: Alan Wormington <wormington AT JUNO.COM> Date: Thu, 17 Jul 2008 17:41:26 -0400 Mark and all, I could not agree more. I have in front of me the soon-to-be-published 2007 Annual Report of the Ontario Bird Records Committee. Of 122 records that are accepted, 77 records (63%) have photos as part of the documentation. This equates to almost 2/3 of all records having photos! For comparative purposes, I randomly pulled out the OBRC Annual Report for 1986. In that report there are 132 accepted records, of which only 37 (28%) include photographs. So Mark's observation is certainly real -- more people own cameras than ever before. Despite this, I believe there is yet another variable as to why more rarities are being photographed. And that reason is the following: in this age of instant communication, word gets out much faster about the presence of a rarity. The result is that a rarity is likely to be observed by many more individuals than compared to the past -- more individuals, more likely the bird will be photographed. I can easily compare this modern scenario to the past, where many rarities were seen by the finder only. After all, that person had to drive home to make telephone calls, and there was no guarantee that others would even be home to take the call. So a whole day might go by before anyone else looks for the bird, assuming it is even still present. Despite the blessing of more records with photographs, in contrast I have seen a different (but depressing) trend over the years. And that pertains to written reports that are clearly of much poorer quality compared to those of the past. Today a records committee is lucky to get a single paragraph of description. In the past it was not unusual to get multiple pages of written description to document a rarity. Why this is happening could be debated, but I think the core reason is that most of the newer, somewhat less-experienced birders nowadays just don't see the big picture. Or they are more interested in bird listing rather than bird study. Who knows. Cheers, Alan Wormington Assistant to the OBRC Secretary ************************** On Thu, 17 Jul 2008 15:35:05 -0500 Mark Lockwood writes: > Dear all, > > > > With all of this discussion about certainly and uncertainty, I do > know > one thing with what I think is a 100% certainty. That is that the > advent of digital cameras have changed the way the Texas BRC works > in a > very short time. I was looking at the uncirculated records before > the > TBRC and to my amazement 28 out of the next 33 records that are > ready to > go into circulation have at least one associated photo. I am sure > we > can all recall when that number was closer to, if not below, 50%. > As > time goes by I am also seeing an increase in the overall quality of > the > images provided as well. This has resulted in an upward shift in > the > percentage of accepted records since for some records it shifts the > burden away from the reporter's ability to provide detailed > descriptions > that eliminate congerers and other similar species. It would be > nice, > but probably too much to ask, for digital audio recorders and higher > quality microphones to become available at more reasonable process > as > well. > > > > This is not anything revolutionary or anything, but it was kind of a > moment when I looked at folders and saw all of the photos lined up. > > > > Mark > > > > > > Mark Lockwood > > Secretary, Texas Bird Records Committee > > 402 E. Harriet Ave. > > Alpine, Texas 79830 > > mark.lockwood AT tpwd.state.tx.us > > > > Visit the Texas Bird Records Committee at > > http://texasbirds.org/tbrc/ > > TOS Handbook of Texas Birds at > > http://www.tamu.edu/upress/BOOKS/2004/lockwood.htm > > TOS Occ. Publications at > http://www.texasbirds.org/publications/occasional.html > > > > > > -------------------------------------------------- > Bird Records Committee Forum archives: > http://listserv.indiana.edu/archives/brcf-l.html "The Early Worm Gets The Bird!" --- Alan Wormington -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: technology and BRCs From: Mark Stevenson <drbrdr AT WORLDNET.ATT.NET> Date: Thu, 17 Jul 2008 14:04:20 -0700 Hi all, The other side of the coin is definitely sound. With the advent and popularity of MP3 players, there is probably a decreased number (and definitely a decreased percentage) of people carrying recording equipment into the field. Still, the rarities whose documentation most benefits from being recorded often get recorded. We are seeing the occasional recording made on a digital voice recorder like the Olympus DS-30 or made on a video recording where the image is marginal and the sound is useful. You can get a DS-30 (that interfaces with a computer) for a not-unreasonable price and make surprisingly good recordings without an expensive external mic if the sound level is not too low. Mark Stevenson Tucson, AZ ----- Original Message ----- From: Mark Lockwood To: BRCF-L AT LISTSERV.INDIANA.EDU Sent: Thursday, July 17, 2008 1:35 PM Subject: technology and BRCs Dear all, With all of this discussion about certainly and uncertainty, I do know one thing with what I think is a 100% certainty. That is that the advent of digital cameras have changed the way the Texas BRC works in a very short time. I was looking at the uncirculated records before the TBRC and to my amazement 28 out of the next 33 records that are ready to go into circulation have at least one associated photo. I am sure we can all recall when that number was closer to, if not below, 50%. As time goes by I am also seeing an increase in the overall quality of the images provided as well. This has resulted in an upward shift in the percentage of accepted records since for some records it shifts the burden away from the reporter's ability to provide detailed descriptions that eliminate congerers and other similar species. It would be nice, but probably too much to ask, for digital audio recorders and higher quality microphones to become available at more reasonable process as well. This is not anything revolutionary or anything, but it was kind of a moment when I looked at folders and saw all of the photos lined up. Mark Mark Lockwood Secretary, Texas Bird Records Committee 402 E. Harriet Ave. Alpine, Texas 79830 mark.lockwood AT tpwd.state.tx.us Visit the Texas Bird Records Committee at http://texasbirds.org/tbrc/ TOS Handbook of Texas Birds at http://www.tamu.edu/upress/BOOKS/2004/lockwood.htm TOS Occ. Publications at http://www.texasbirds.org/publications/occasional.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: technology and BRCs From: Mark Lockwood <Mark.Lockwood AT TPWD.STATE.TX.US> Date: Thu, 17 Jul 2008 15:35:05 -0500 Dear all, With all of this discussion about certainly and uncertainty, I do know one thing with what I think is a 100% certainty. That is that the advent of digital cameras have changed the way the Texas BRC works in a very short time. I was looking at the uncirculated records before the TBRC and to my amazement 28 out of the next 33 records that are ready to go into circulation have at least one associated photo. I am sure we can all recall when that number was closer to, if not below, 50%. As time goes by I am also seeing an increase in the overall quality of the images provided as well. This has resulted in an upward shift in the percentage of accepted records since for some records it shifts the burden away from the reporter's ability to provide detailed descriptions that eliminate congerers and other similar species. It would be nice, but probably too much to ask, for digital audio recorders and higher quality microphones to become available at more reasonable process as well. This is not anything revolutionary or anything, but it was kind of a moment when I looked at folders and saw all of the photos lined up. Mark Mark Lockwood Secretary, Texas Bird Records Committee 402 E. Harriet Ave. Alpine, Texas 79830 mark.lockwood AT tpwd.state.tx.us Visit the Texas Bird Records Committee at http://texasbirds.org/tbrc/ TOS Handbook of Texas Birds at http://www.tamu.edu/upress/BOOKS/2004/lockwood.htm TOS Occ. Publications at http://www.texasbirds.org/publications/occasional.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Richard Heil <rsheil AT COMCAST.NET> Date: Sun, 13 Jul 2008 23:25:11 -0400 Since Ted won't let it rest: At 07:06 PM 7/13/2008, Ted Floyd wrote: >Sorry to have been unclear. but you're missing my point. I was >stating, in an absolute sense, my assignment of a probability to >that particular flycatcher. Since there is no certainty, how can you possibly assign a probability to the unknown? > > Since both the actual observation and the assignment of > probability are being made by the observer, > >You're *definitely* missing my point! (And, again, sorry to have >been unclear.) I have been assuming, for the sake of argument, that >my assignment of probability is somehow absolute. Somehow, indeed! Again using a concept in the attempt to deny it. > I am treating the matter hypothetically. I'm a naturally > "subjunctive," conditional, hypothetical sort of person, who > frequently writes in a voice that I have recently learned is termed > "free indirect discourse." I have further been told that such > writing is inscrutable to engineers, doctors, airplane pilots, and > computer programmers. Hey, those are great people, and much more > valuable in this life than I am, but they often miss my point. As > far as I know, I rarely state anything literally. Where would be > the fun in that? The logic in all of this is so convoluted as to be incomprehensible. Gibberish, as I said in an earlier post. Hypothetically, if that there's a wall I'm speeding towards at 80 mph, well, then, hypothetically, I'd better start applying my brakes. Whadya think? >Meanwhile, Richard Heil writes: > > > If your main point in this entire discussion was merely > > that often there is uncertainty regarding records ... > >Yes, that is the basic gist of what I was saying. > > > then of course, obviously, that is true, and I agree. > > But you went much farther. You said that certainty > > never exists: "there is no certainty." > >We're definitely in the realm of semantics now, but, yes, I am >basically comfortable with the preceding. We used to live in a >yes/no, black/white, good/evil, right/wrong, us/them, accept/reject >world. That is the world of Plato. We do not live in that world >anymore. Check that: The vast majority of us cheerfully go about our >business living in that world! But science and philosophy have not >been there for close to 150 years. Accept/reject is a dichotomy that >has been rejected--heh!--by the scientific establishment for 150 >years, and, in many respects, for 400 years. Science and philosophy >assure us that we live in a world of uncertainty, that our world is >colored in shades of gray, that there is no certainty. In terms of >my own health and happiness, I'm very happy that engineers, doctors, >airplane pilots, and computer programmers have not heeded those >lessons. But I believe that those lessons apply well to BRCs. Ted, you've ignored my earlier points regarding the dichotomies and contradictions inherent in the idea of "absolute uncertainty" and have continued blithely on. Your points are so full of contradictions it's hard to know where to begin. You make declarative statements and write in certainties and absolutes all the while denying their existence. You state declaratively that the world is one of moral relativism and scientific subjectivism, a world with no truths, no good or evil, and no right or wrong (and I was accused of exaggerating your subjectivism earlier!). However, intellectually you aren't true to your ideas at all. You abandon your own stated philosophy faster than light ... as needed. You don't believe in truths and certainties, you claim, but you are happy that those that you actually depend upon for your happiness, safety, and life do!!! You live this contradiction because you really know your ideas to be false. Scientists, doctors, engineers, and pilots, contrary to your declarations, do indeed believe in a yes/no, black/white, right/wrong world, at least if their behavior is any measure of their beliefs. They have to. They make these sorts of decisions all day long. Lives depend upon it, including theirs. You take possible quantum and other particle uncertainties and inappropriately apply them to the macro-world. Big macro mistake. Somehow, particle uncertainty becomes "there is no yes/no or right/wrong or good/evil or accept/reject in your mind. I'm not getting this giant leap. I don't see the connection ... at all. To get back to bird records, as I stated earlier, but ignored, any quantum or other micro-uncertainties are IRRELEVANT to bird observations and records committees analysis of them. That's a certainty. Finally, if I were to adhere to your philosophy, how can I trust or even know that anything you say is true since there are no truths? Furthermore, how can you begin to make any kind of argument in the first place? This will be my last post on the subject. I've had enough of the nonsense, literally. Can't take no more, hypothetically speaking of course. Richard S. Heil S. Peabody, MA rsheil AT comcast.net Massachusetts Avian Records Committee P.S. Ted, you might consider moving to a lower elevation. -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Ted Floyd <tedfloyd57 AT HOTMAIL.COM> Date: Sun, 13 Jul 2008 16:06:14 -0700 Hi, all. Sorry for my delay in jumping back into this thread. We're kinda in the realm of semantics right now, and I suppose that's familar territory for a lot of BRC'ers, but here goes. First, Paul Guris says: > So you say your probability is very high, but less than > 100% so you didn't submit it. I said it was greater than 50%. I wouldn't call that "very high." > Looking at the above, do you wish to argue that the > probability of Ivory-billed still existing is much higher > than the probability that you saw an Alder Flycatcher? > After all, there are multiple observers who are more > certain than you are of their ID's. Sorry to have been unclear. but you're missing my point. I was stating, in an absolute sense, my assignment of a probability to that particular flycatcher. > If you wouldn't argue that point, my question would be "why not?" They're just not comparable. Apples and oranges. There are whopping differences, for example, in prior likelihood of occurrence. By the way, that word, "prior," presents the opportunity for me to encourage BRCs to consider using Bayesian methods of inference in evaluating records. Indeed, I would say that the whole Bayeseian paradigm is splendidly suited to the overall enterprise of committee work. > Since both the actual observation and the assignment of probability are being made by the observer, You're *definitely* missing my point! (And, again, sorry to have been unclear.) I have been assuming, for the sake of argument, that my assignment of probability is somehow absolute. I am treating the matter hypothetically. I'm a naturally "subjunctive," conditional, hypothetical sort of person, who frequently writes in a voice that I have recently learned is termed "free indirect discourse." I have further been told that such writing is inscrutable to engineers, doctors, airplane pilots, and computer programmers. Hey, those are great people, and much more valuable in this life than I am, but they often miss my point. As far as I know, I rarely state anything literally. Where would be the fun in that? > 2) Data SHOULD have a quality threshold. That's not a > bug, it's a feature. You are not confident enough to say > that your conclusion is correct, so you don't submit. > Isn't this how science is supposed to work in general? No doubt this will reveal my pettiness, but I didn't submit because, at some level, I didn't want to get rejected. But, more to the point, I didn't submit because of my belief that the accept/reject dichotomy just wouldn't work in the case of my presumptive Alder Flycatcher. As to how science works, no, I hope it doesn't work the way you're describing it! We report our results, we report our p-values, and that's that. > I'm not sure how other records committees work, but > I think "preponderance of evidence" is often used for > regularly occurring rarities, while "beyond a reasonable > doubt" would be used for extreme rarities, first state > records, birds assumed to be extinct, etc. In NJ, an > Ash-throated Flycatcher report in Cape May in November > would require a vastly lower threshold than any Eskimo > Curlew, and rightly so. I'm aware of this distinction, and I think it's improperly applied. Let's say I'm 85% certain that a 0.1-meter object will hit the earth; and let's say I'm also 85% certain that a 10-kilometer object will hit the earth. If my only criterion is accuracy, then I report both results with equal gusto. If I have practical concerns, then I ascribe greater importance to the 10-km object than to the 0.1-m object. Shouldn't it be the same with BRCs? If all we seek is accuracy (and that's fine, as human lives, honestly, aren't at stake), then we should have the same thresholds for megas as for regular vagrants. But if we see practical value in our work (and that's fine, too), then we should be especially careful about the danger of Type II errors with megas. Incorrectly rejecting a November Ash-throated Flycatcher is no big deal, but incorrectly rejecting an Eskimo Curlew record, anywhere, could be a big mistake. > The assumption that the record would not be accepted > is yours and yours alone, though perhaps not without > cause. Perhaps they would have felt that what you observed > was diagnostic and accepted the record. Good point, and I've wondered about this. Specifically, I have a question for anybody (Kimball?) out there on the California committee. If an observer indicates that he or she is not certain of the ID, how does that affect the committee's deliberations? I assume that it is a huge blow against the submission, akin to confessing "guilty" before a judge, but maybe I am wrong. (Note: If you missed it earlier, the California form has--or, at least, it used to have--a place for indicating whether you're certain of the identification of the species you're reporting.) Meanwhile, Richard Heil writes: > If your main point in this entire discussion was merely > that often there is uncertainty regarding records ... Yes, that is the basic gist of what I was saying. > then of course, obviously, that is true, and I agree. > But you went much farther. You said that certainty > never exists: "there is no certainty." We're definitely in the realm of semantics now, but, yes, I am basically comfortable with the preceding. We used to live in a yes/no, black/white, good/evil, right/wrong, us/them, accept/reject world. That is the world of Plato. We do not live in that world anymore. Check that: The vast majority of us cheerfully go about our business living in that world! But science and philosophy have not been there for close to 150 years. Accept/reject is a dichotomy that has been rejected--heh!--by the scientific establishment for 150 years, and, in many respects, for 400 years. Science and philosophy assure us that we live in a world of uncertainty, that our world is colored in shades of gray, that there is no certainty. In terms of my own health and happiness, I'm very happy that engineers, doctors, airplane pilots, and computer programmers have not heeded those lessons. But I believe that those lessons apply well to BRCs. Ted Floyd tedfloyd57 AT hotmail.com Lafayette, Boulder County, Colorado _________________________________________________________________ The i’m Talkaton. Can 30-days of conversation change the world? http://www.imtalkathon.com/?source=EML_WLH_Talkathon_ChangeWorld -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Richard Heil <rsheil AT COMCAST.NET> Date: Thu, 3 Jul 2008 17:15:12 -0400 Ted, If your main point in this entire discussion was merely that often there is uncertainty regarding records ... then of course, obviously, that is true, and I agree. But you went much farther. You said that certainty never exists: "there is no certainty." That's what I responded to, pointing out the contradictions and paradoxes associated with such a statement, and tried to explain why this and other micro-uncertainties that may exist are irrelevant in the real macro world that we all live in. Thanks, Richard S. Heil S. Peabody, MA rsheil AT comcast.net Massachusetts Avian Records Committee At 08:11 AM 7/3/2008, Ted Floyd wrote: >Hello, everybody. Just 2 quick points. 1. Richard Heil keeps >returning to quantum uncertainty, but that's not where I, >personally, have been trying to take this. Recent (post-1859) >scientific and philosophical ideas about uncertainty go well beyond >quantum uncertainty. Yes, I cited Schroedinger's famous 1927 >discovery, but I also cited other examples of general >indeterminacy/incompleteness/relativism/etc. Here's a nice >discussion of uncertainty that doesn't, so far as I can tell, in any >way appeal to quantum uncertainty: >http://sibleyguides.blogspot.com/search/label/Ivory-billed%20Woodpecker >2. Laurent Raty's broad take on Type I and Type II errors looks to >be spot-on. I was thinking only of BRC errors in the narrow context >of birds that are actually detected and reported to records >committees. In those instances, I believe it is still correct that a >shift from a standard of "beyond a reasonable doubt" to a >"preponderance of evidence" would result in a lower overall error >rate. Indeed, that is, by definition, the goal of the standard of a >"preponderance of evidence," namely, What is the most likely >explanation? Thanks, all. Fun conversation. All best, --Ted Ted >Floyd tedfloyd57 AT hotmail.com Lafayette, Boulder County, Colorado > >---------- >Don't get caught with egg on your face. >Play >chicktionary! -------------------------------------------------- >Bird Records Committee Forum archives: >http://listserv.indiana.edu/archives/brcf-l.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: "Paul A. Guris" <paul AT PAULAGICS.COM> Date: Thu, 3 Jul 2008 11:24:04 -0400 Ted- A number of points (sorry for the length): 1) It's interesting that you selected David Sibley's post that is highly related to the Ivory-billed Woodpecker. Let's look at that example and compare it to your Alder Flycatcher, and how it would work in your system: - Your report is from a single observer. Multiple people would swear on a stack of Bibles that they saw an Ivory-billed. - Part of your ID is call note. Multiple people swear that they have heard what they deem to be diagnostic calls and drums. - Your report is of a species that is one of the most difficult ID problems in North America. Ivory-billed is a slam-dunk. - Your report is out of range. Most Ivory-billed reports are well within their historic range. So you say your probability is very high, but less than 100% so you didn't submit it. Looking at the above, do you wish to argue that the probability of Ivory-billed still existing is much higher than the probability that you saw an Alder Flycatcher? After all, there are multiple observers who are more certain than you are of their ID's. If you wouldn't argue that point, my question would be "why not?" Are you simply better than them? [That last one is sooooo unfair and I don't mean it. I'm just hoping to get a snicker out of you. ;-) ] Since both the actual observation and the assignment of probability are being made by the observer, the probability portion of the data is pretty much useless. It's merely a subjective measure of "how confident am I?", but is based on nothing but each person's own interpretation which can be skewed by their abilities, what they saw, how much they want that bird for a state list, who they're birding with, etc. Actually, I can't imagine how you could create a data model that is more subjective. With the number of errors made by people who believe their sightings to be definite, I think any attempt to go to finer granularity is pretty much useless. As I said before, it gives the illusion of greater accuracy, but does not, in fact, deliver it. 2) Data SHOULD have a quality threshold. That's not a bug, it's a feature. You are not confident enough to say that your conclusion is correct, so you don't submit. Isn't this how science is supposed to work in general? If you have data points that are subjective, open to interpretation, and/or incomplete, do you publish a paper on them and say "I know my data is questionable, but here's my conclusion IF it's correct?" If we published every suspicion that every birder ever had, you wouldn't be able to find the wheat(ear) for the chaf(finch). Yes, different people have different personal quality thresholds for submitting a record, but adding even more subjectivity to the system is hardly the solution. 3) You state the following: > I believe it is still correct that a shift from a standard of > "beyond a reasonable doubt" to a "preponderance of evidence" > would result in a lower overall error rate. I'm not sure how other records committees work, but I think "preponderance of evidence" is often used for regularly occurring rarities, while "beyond a reasonable doubt" would be used for extreme rarities, first state records, birds assumed to be extinct, etc. In NJ, an Ash-throated Flycatcher report in Cape May in November would require a vastly lower threshold than any Eskimo Curlew, and rightly so. Difficulty of ID is also used as part of the decision making process, so Hammond's Flycatcher might require a higher threshold than a Swallow-tailed Kite. Again, features, not bugs. 4) You seem to be projecting a lot of personal opinions onto records committees. Looking at your example where you were not 100% confident so did not submit the Alder Flycatcher: - The decision not to submit was yours, so really does not have anything to do with the records committee. You could have submitted and noted your lack of assuredness and why, very valuable information. - The assumption that the record would not be accepted is yours and yours alone, though perhaps not without cause. Perhaps they would have felt that what you observed was diagnostic and accepted the record. - The view that not having your record accepted would render your data useless is yours and yours alone, unless the Colorado committee told you explicitly not to bother turning it in or throws away all unaccepted reports. Had you submitted, I assume anybody doing research on out of range Alder Flycatchers could have had access to your details and determined for themselves whether or not they felt it was a valid data point. The committee isn't the end all, be all in assessing data quality. Again, sorry for the long-winded response, but I find my disagreements to be with your base assumptions. For me, your process arguments are moot until the issues with your assumptions are addressed. -PAG Paul A. Guris Green Lane, PA www.paulagics.com paul AT paulagics.com -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Ted Floyd <tedfloyd57 AT HOTMAIL.COM> Date: Thu, 3 Jul 2008 05:11:11 -0700 Hello, everybody. Just 2 quick points. 1. Richard Heil keeps returning to quantum uncertainty, but that's not where I, personally, have been trying to take this. Recent (post-1859) scientific and philosophical ideas about uncertainty go well beyond quantum uncertainty. Yes, I cited Schroedinger's famous 1927 discovery, but I also cited other examples of general indeterminacy/incompleteness/relativism/etc. Here's a nice discussion of uncertainty that doesn't, so far as I can tell, in any way appeal to quantum uncertainty: http://sibleyguides.blogspot.com/search/label/Ivory-billed%20Woodpecker 2. Laurent Raty's broad take on Type I and Type II errors looks to be spot-on. I was thinking only of BRC errors in the narrow context of birds that are actually detected and reported to records committees. In those instances, I believe it is still correct that a shift from a standard of "beyond a reasonable doubt" to a "preponderance of evidence" would result in a lower overall error rate. Indeed, that is, by definition, the goal of the standard of a "preponderance of evidence," namely, What is the most likely explanation? Thanks, all. Fun conversation. All best, --Ted Ted Floyd tedfloyd57 AT hotmail.com Lafayette, Boulder County, Colorado _________________________________________________________________ Don't get caught with egg on your face. Play chicktionary! http://club.live.com/chicktionary.aspx?icid=chick_wlhmtextlink1_feb -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Richard Heil <rsheil AT COMCAST.NET> Date: Wed, 2 Jul 2008 20:43:14 -0400 At 08:18 AM 7/1/2008, Ted Floyd wrote: >And as to straight-out certainty, I'm just not buying it. Not since >1859. We now know that we live in a universe that is evolutionary >(1859), relativistic (1905), uncertain (1927), impermanent (1929), >incomplete (1931), chaotic (1961), entangled (1982), and otherwise >fuzzy, blurry, and indeterminate. That's not political correctness >or some "cult of uncertainty." That's 149 years of scientific >progress. Old ideas die hard. A lot of folks still cling to >Platonistic ideals of certainty, truth, and beauty. But not >scientists. Not since 1859. > >We now know that it is better to say p=0.35, or p=0.99, or even >p~1.00. But not p=1.00. Indeed, it is in the realm of those very >small (p~0.00) and very large (p~1.00) probabilities where some of >the most surprising of mathematical and philosophical issues play >out. A fine introduction, in my opinion, is Brian Greene's Fabric of >the Cosmos (Knopf, 2004). An important sub-theme--and it is one that >runs throughout Fabric of the Cosmos--is that it is a mistake to >equate near-certainty with certainty. The best we can say about the >universe around us is that it is a continuum of uncertainties; and >the best we can do is describe reality in probabilistic terms. For >sure, that is a complete repudiation of Plato. But it does seem to >be where modern science has taken us. Ted et al., I'm not sure that there is unanimous agreement among physicists or philosophers either on the nature of reality or to what extent quantum features undermine it. The universe may be an inseparable entity because of interactions but to a very close approximation it may be divided up into lots of quasi-autonomous little things whose separate identity is rarely questioned in daily life. An observed Chestnut-sided Warbler is only infinitesimally different from an unobserved Chestnut-sided Warbler. It's essential identifying characteristics, what makes it what it is, and differentiates it from all other things (a tree, a can opener, or a Bay-breasted Warbler) is unchanged. Epistemologically only its non-essential are altered very slightly. Consider that there is an inherent paradox in the statement "all things are uncertain" which has not been addressed. Absolute uncertainty is an oxymoron. "There are no certainties" is a contradiction. QT is about particles, not existence as a whole. The primacy of existence is the axiom that existence exists and that things have an identity. One cannot argue against existence or certainty (=identity) because you must use and therefore accept the concept in any attempt to deny it. The fact that things change helps confirm identity/certainty because change presupposes the concepts of what changes, from what and to what. Without the concept of identity/certainty, change is meaningless. Similarly paradoxical is the fact that quantum theory science uses classical concepts of science and experimental apparatus that are themselves discredited by quantum theory. Furthermore, if all things are imbued with absolute uncertainty, how exactly does one calculate the probability of the unknown? If there is no certain thing as an Alder Flycatcher, how do calculate the probability that your sighting is of an Alder Flycatcher? Honestly, all of this has almost nothing to do with records committees deciding whether a given record has sufficient evidence to be accepted or not. For birders identifying birds, and for committees evaluating records, any quantum uncertainty is infinitesimal and therefore irrelevant. For a birder or a committee it matters little whether p=1 or p~1. For practical purposes close to certainty is certainty. Of course, many records or even most records are not certain, but as Buford Myers said earlier: "what bird records committee has ever stated or even implied that its decisions are expressions of certainty." This is right to the point. Thanks, Richard S. Heil S. Peabody, MA rsheil AT comcast.net Massachusetts Avian Records Committee -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports (RESEND) From: laurent raty <l_raty AT HOTMAIL.COM> Date: Thu, 3 Jul 2008 02:28:47 +0200 Hi Ted, > Our Type I error rates would go up slightly, but our Type II error rates would go down substantially. I'm not sure to see why the opposite could not also happen... I.e., your Type I error rates could increase rather dramatically, while your Type II error rates might in practice go down very little. If you consider the entire process of bird detection/identification, there is little doubt that Type II error rates are often huge, particularly for drab and hard-to-identify species, like your Alder Flycatcher. Not because of RCs' policies... But simply because, each time an Alder Flycatcher enters Colorado but does not choose to perch right in the front of a knowledgeable birder, you get a Type II error - a bird was there, but it's not going to be accepted. Changing the standards of acceptance would have no effect on this; even if we lowered them substantially, the Type II error rates would still remain very high. On the other hand, (as I know you know,) for a rare bird confusable with common species, the actual probability of an ID being correct is conditional to its relative rarity. If this species is really rare, even very suggestive (but non-diagnostic) evidence can produce very high rates of false positives. The extreme situation being that the "rare" bird is in fact not there at all, in which case using "preponderance of evidence" instead of "proof beyond a reasonable doubt", you could end up building a whole body of data in which the Type I error rate would be, unavoidably, 100%... Cheers, Laurent - -- Laurent Raty Brussels, Belgium l_raty AT hotmail.com _________________________________________________________________ Discover the new Windows Vista http://search.msn.com/results.aspx?q=windows+vista&mkt=en-US&form=QBRE -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports (RESEND) From: Ted Floyd <tedfloyd57 AT HOTMAIL.COM> Date: Wed, 2 Jul 2008 13:50:16 -0700 SORRY! THESE LESS-THAN AND GREATER-THAN SYMBOLS ARE TOO VEXING FOR HOTMAIL. I'M RESENDING WITHOUT THE SYMBOLS! --TED Hello, Everybody. Thanks for the recent responses from Mac Myers, Paul Guris, David Christie, and Laurent Raty. Here's a real-life anecdote that I think gets at many of the points we've been dicussing here. Earlier this year, I observed two interesting birds within about 100 feet of one another. I reported one as a Prairie Warbler, which is a review species in Colorado (27 accepted records for the state). I reported the other as a possible Alder Flycatcher, also a review species (18 accepted records for the state). I right away submitted extensive written documentation for the Prairie Warbler, which I saw and heard at close range. Although there were a few problems (it was raining, the plumage of the bird was ambiguous), the documentation was, I suspect, pretty straightforward. I saw an easy-to-identify bird, I described its plumage and structure, I noted various aspects of its behavior, and I heard its vocalizations. In qualitative terms, I'd say I was nearly certain that the bird was a Prairie Warbler, or p~1.00 in quantitative terms. As to the possible Alder Flycatcher, I wasn't so sure. Here, we're dealing with three probabilities: the probability that the bird was an Empidonax flycatcher, the probability that it was a "Traill's" flycatcher, and the probability that it was an Alder Flycatcher. Let's call those 3 probabilities p(1), p(2), and p(3). I'd assign them the following values: p(1)~1.00, p(2)=0.99, p(3)[greater than]0.50. (Actually, we could probably set the second probability at 0.999, but let's keep it to 2 sig figs, just for consistency's sake.) If you pressed me on p(3), and I'm specifically thinking of Paul Guris's question now, I'd haveta say: 0.50[less than]p(3)[less than]0.99. That's about the best I can say. The most likely "outcome," as statisticians say, is that the bird was an Alder Flycatcher. There is a "preponderance of evidence" that it was "guilty" (now I'm thinking of what Laurent Raty wrote) of being an Alder Flycatcher; but there was not "proof beyond a reasonable doubt," to keep this tortured legal metaphor going. Now here's where it gets tricky, and here's where BRCs come into play. I did not submit documentation of this possible Alder Flycatcher to the Colorado Bird Records Committee (CBRC). A few weeks later, the CBRC chairman politely and solicitously asked me if I would be submitting documentation, and I declined. My reasoning: I wasn't sure. My own personal standard--and, truly, this is merely a personal standard--comes fairly close to "proof beyond a reasonable doubt." I strongly suspect--indeed, I know--that many of you out there hold up similar standards. Unless you're nearly certain (but never completely certain, because, as we all know, there is no such thing as certainty), you don't bother with a submission to a BRC. (Why waste all the time and paper, as Alan Wormington says.) I assume we all see the problem here? It probably *was* an Alder Flycatcher. But the CBRC won't be evaluating the report. It won't even be "rejecting" the report. Simply, there is no report at all. And that's a pity. It's a pity, in particular, because, this past spring, there appears to have been an unusually strong flight of Alder Flycatchers across the western Great Plains. (Getting a little off subject, there was a strong westerly push of boreal breeders in general: Alder and Yellow-bellied Flycatchers, Philadelphia Vireo, Gray-cheeked Thrush, and especially Blackpoll Warbler.) Any sort of quantitative assessment of the Alder Flycatcher flight this past spring is basically invalidated by the fact that merely "possible" records are being excluded from the data. Like it or not, if an observer suspects that a report will be "rejected" by a BRC, then that observer is considerably less likely to report than if he or she is reasonably confident of "acceptance." Which brings me to an especially thorny angle here. Given what I know of the literature on Alder and Willow Flycatchers, and given, to a lesser extent, what I know of the composition of the CBRC, I could have produced documentation that would likely result in "acceptance," despite my own misgivings about the report. (No need to go into the details, but I could have over-emphasized the call-note, which I heard repeatedly, and which was perfect for Alder; I could have over-emphasized the field marks that were good for Alder; and I could have downplayed the field marks that were suggestive of Willow.) No matter how neutral or impartial we strive to be, we humans do an excellent (?) job of telling stories, providing witness testimony, and, yes, describing birds in ways that are consistent with the outcomes that we desire. My conundrum in a nutshell: 1. There was, in my opinion, a preponderance of evidence (0.50[less than]p[less than]0.99) that the bird was an Alder Flycatcher. 2. My guess was that a mere preponderance of evidence would not be good enough for the CBRC. 3. I didn't want to submit a somehow dishonest report that implied p~.99. 4. So there is no report at all. 5. And a potentially interesting phenomenon (unusually strong flight of Alder Flycatcher) goes undocumented. One final thought. I'm a realist, and I accept that the accept/reject dichotomy isn't going to go away anytime soon. But what if we moved away from what I sense is a "proof beyond a reasonable doubt" mentality for accepting records, and instead settled on a "preponderance of evidence" standard? It would certainly be more accurate. Our Type I error rates would go up slightly, but our Type II error rates would go down substantially. Plus, reports such as my Alder Flycatcher would actually get evaluated by BRCs, as opposed to languishing in my field notebook. Ted Floyd tedfloyd57 AT hotmail.com Lafayette, Boulder County, Colorado P.s. I hope I haven't come across, in any way at all, as critical of the CBRC, which I think does magnificent work. In the areas of timeliness, thoroughness, accuracy, professionalism, fairness, competency, and even wisdom (have I missed anything?), the CBRC is exemplary. _________________________________________________________________ Windows Live Hotmail is giving away Zunes. Enter for your chance to win. http://www.windowslive-hotmail.com/ZuneADay/?locale=en-US&ocid=TXT_TAGLM_Mobile_Zune_V3 -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Ted Floyd <tedfloyd57 AT HOTMAIL.COM> Date: Wed, 2 Jul 2008 13:35:46 -0700 Hello, Everybody. Thanks for the recent responses from Mac Myers, Paul Guris, David Christie, and Laurent Raty. Here's a real-life anecdote that I think gets at many of the points we've been dicussing here. Earlier this year, I observed two interesting birds within about 100 feet of one another. I reported one as a Prairie Warbler, which is a review species in Colorado (27 accepted records for the state). I reported the other as a possible Alder Flycatcher, also a review species (18 accepted records for the state). I right away submitted extensive written documentation for the Prairie Warbler, which I saw and heard at close range. Although there were a few problems (it was raining, the plumage of the bird was ambiguous), the documentation was, I suspect, pretty straightforward. I saw an easy-to-identify bird, I described its plumage and structure, I noted various aspects of its behavior, and I heard its vocalizations. In qualitative terms, I'd say I was nearly certain that the bird was a Prairie Warbler, or p~1.00 in quantitative terms. As to the possible Alder Flycatcher, I wasn't so sure. Here, we're dealing with three probabilities: the probability that the bird was an Empidonax flycatcher, the probability that it was a "Traill's" flycatcher, and the probability that it was an Alder Flycatcher. Let's call those 3 probabilities p(1), p(2), and p(3). I'd assign them the following values: p(1)~1.00, p(2)=0.99, p(3)>0.50. (Actually, we could probably set the second probability at 0.999, but let's keep it to 2 sig figs, just for consistency's sake.) If you pressed me on p(3), and I'm specifically thinking of Paul Guris's question now, I'd haveta say: 0.50 Subject: Black-bellied Whistling-Duck sighting request From: Phil Davis <pdavis AT IX.NETCOM.COM> Date: Wed, 2 Jul 2008 02:55:31 -0400 Hi Chatters and BRCF-L: Ian McLaren and I have been discussing some issues related to Black-bellied Whistling Duck vagrancy. Ian is not currently subscribed to Birdchat, so I am forwarding a message from him, below. If anyone has any information related to this flock, including one bird with a distinctive deformed bill, please advise and please copy Ian on any replies. Thanks. Phil >Date: Tue, 1 Jul 2008 11:47:38 -0300 >From: "Ian A. McLaren" >To: Phil Davis >Subject: BB Whistling-Duck posting > >I wonder if there have been any further sightings of (or searches >for) the flock >of ten Black-bellied Whistling Duck that occurred 25 May-4 June in >Nova Scotia? >Nine of them then appeared briefly 6 June at Ipswich, MA, as confirmed by a >distinctive abnormal bill growth on one bird. Everything about this flock >suggests that it was of wild origin - no bands, unclipped hind toes, and its >apparent return trajectory (towards FL?). If the bird with the bill growth is >found in a flock further southeast, this might add to the picture. > >See photos of the birds in NS and MA, including the one with the abnormal bill >growth at: > > http://www.nebirdsplus.org/BBWDuck.htm > >Cheers, Ian M. > >Ian A. McLaren >Biology Department >Dalhousie University >Halifax, NS Canada B3H 4J1 =================================================== Phil Davis, Secretary MD/DC Records Committee 2549 Vale Court Davidsonville, Maryland 21035 USA 301-261-0184 mailto:PDavis AT ix.netcom.com MD/DCRC Web site: http://www.MDBirds.org/mddcrc/rcindex.html =================================================== -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Buford Myers <bmyers990 AT EARTHLINK.NET> Date: Tue, 1 Jul 2008 14:37:42 -0500 Ted, Others have pointed out some problems with your arguments. It seems to me that there is a fundamental error in your assumptions and subsequent analysis. What bird records committee has ever stated or even implied that its decisions are expressions of certainty? When a committee votes to accept or not accept a record, it is simply evaluating the evidence as best it can. Nobody believes this is error-free. Nobody believes it is a truth for the ages. And it is most certainly not a statement about the sighting itself. Rather, it is a statement about the evidence produced in support of the purported sighting. Bird record committees are not list police. But when it comes time to write a book about the birds of Colorado or do a checklist of the birds of Louisiana, somebody is going to make decisions about what to include and what to exclude. I have yet to see one that lists a whole bunch of species and then assigns a probability (say, 0.01 through 0.99) for each, and I hope I never do. Mac Louisiana Bird Records Committee alternate and former voting member -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Questions for Ted on Records Probability From: "Paul A. Guris" <paul AT PAULAGICS.COM> Date: Tue, 1 Jul 2008 12:23:27 -0400 Ted, this may be a bit of a bush league response, but I come from the IT field, not biology, so I may simply not have the background to understand your view. Please bear with me and feel free to knock down anything I say/ask. You said the following: "If pressed, they would say that there was a 6% chance that the bird was something other than a Golden-winged Warbler." Now in the computer world, we have an ancient (for computers) term called GIGO; Garbage In, Garbage Out. Looking at your argument, to me it seems based on an immensely faulty premise; that committee members can assign probability with anything remotely approaching this level of accuracy. Without first defining a detailed data model (i.e. defining the criteria that determine HOW probabilities are assigned), I think the process itself is useless. Think of it in significant figures. Can an individual reasonably assign a probability of 35%? Is their granularity really closer to +/-10%, i.e. rating a record from 1 to 10? Maybe it's more like +/-25%, i.e. rating a record from 1 to 4? Maybe it's different for every committee member? Now let's look at what it would take to define the data model. Because of the unquantifiable backgrounds of the committee members, as well as the unquantifiable differences between committee members, plus the enormous number of actual report variables, it appears to me that your system would be completely unworkable at the accuracy levels you're attempting to use. It would simply give the appearance of greater accuracy rather than an actual increase in accuracy. I also see distribution problems with your system. For example, think of a record where 11 members of the committee have assigned probabilities ranging from 20-90%? What do you do? Do you do "Olympic scoring", throwing out the high and low numbers? What if one of those outlying numbers is from the most knowledgeable person on that particular species? What happens when somebody assigns a value of 20%, then listens to the argument of the 90% person, and changes their "probability" to 75%? Is their "probability" rating really worth all that much? Doesn't their 75% differ vastly from the 75% assigned by a person coming from a totally different path and, if so, are both paths equally valid? Isn't a 20-90 split that averages out to, say 60, very different than a 50-70 split that also averages out to 60? Now think about the current system in use. In NJ, anyway, a record is not simply tagged accept/not accept. The vote tally is published with the record. A "not accept" vote of 0-11 is VERY different than one that is 6-5. That information is available to everybody. I think it does show "probability", and probably (no pun intended) at a significant figure level that is much more valid than the one you propose. The next question is, would it be a better system if votes were 1-4 rather than yes/no? Maybe more levels? Maybe fewer? I'd have to see the description of what each level meant before accepting any such system, but I imagine I could be talked into it, but I think I would absolutely reject the ability of committee members to reasonably assign probability at even 10% increments. FYI, Pennsylvania does use levels for both accept and not accept, and it seems like a reasonable approach to me. They can be found at: http://www.pabirds.org/PORC/PORC_Purpose.htm There is one other very important point where I think we disagree. You seem to feel the vote should be on the probability that the bird occurred. I have sat on Pennsylvania's committee and am currently on New Jersey's, and my view has always been that the I'm voting first on whether or not the documentation supports the ID, and second if the documentation is credible. Using your system, I would see the following examples (coming from a discussion literally days ago on the NJ committee): - A report of Ash-throated Flycatcher from Cape May, NJ in November would get a high probability, even with no written description. At that point it seems like I just believe every occurrence where time/place looks pretty good, with no actual documentation required. - A report of virtually anything by any number of world class birders would be assigned a high probability, even without a description. There are people in NJ that I would personally believe if they told me they saw just about anything. At that point, I would simply accept records by reputation. If I do that, how does somebody assess my vote in 100 years? Do I simply write "I give this record a 90% probability because Ace Q. Birder said he saw it, and he's really, really good?" I may be misinterpreting your stance, but it sure sounds to me like you're arguing for voting on the likelihood that the bird occurred, not necessarily on the quality of the documentation filtered through the likelihood that the bird occurred. As I said, feel free to school me where you feel I'm completely off base. Sorry for going on so long. I hope I didn't enter into the realm of Department of Redundancy Department. -PAG Paul A. Guris Green Lane, PA www.paulagics.com paul AT paulagics.com -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Acceptance/non-acceptance of records, Type I & II errors From: David Christie <maryspt AT NBNET.NB.CA> Date: Tue, 1 Jul 2008 09:39:49 -0300 [This thread began as "Question on protocol in records committee journal reports" but recently discussion has turned to committees' review of records.] At 1:58 PM -0700 6/30/08, Ted Floyd wrote: >Many rejected reports are less clear-cut. Many rejected reports are, >in fact, probably correct. > >Committees can make two types of errors. They can falsely accept >incorrect data (Type I error), and they can falsely reject correcet >data (Type II error). > >Mathematically, logically, and philosophically speaking, Type I and >Type II errors are equally, well, erroneous. They're both mistakes, >plain and simple. In the New Brunswick Bird Records Committee, (in addition to considering the likelihood of natural occurrence) what we vote on is whether or not the documentation adequately supports the reported identification. In effect, this amounts to accepting adequately documented correct identifications as being correct, but not necessarily that inadequately documented ones are incorrect. Identification of the latter is considered "questionable," but "does not necessarily mean that... a mistake was made." (The wording for our voting categories is on the Web at http://www3.nbnet.nb.ca/maryspt/BRC/Voting.html .) Ours is a small committee handling a small number of records. Out of curiosity I looked back at about a hundred records voted on during the late 1990s and early 2000s. 78% were accepted, 18% were not accepted because the identification was questionable, and 4% not accepted because natural occurrence was unlikely. Of those in the questionable identification category, the documentation of more than 12% (of total records) actually indicated a different identification than the one submitted (4 of them were dull Baltimore Orioles that had been reported as Bullock's), and a bit more than 5% might have been either the reported species or another one equally rare or more common. Some of these 5% could include the "egregious Type II errors" that Ted is concerned about. But honestly, most of the ones in this category were seen briefly or under marginal conditions or were reports where the observers apparently did not see the birds well enough or examine them critically enough to notice important features distinguishing them from similar species. I acknowledge that 2% might possibly be Type II errors but the quality of these reports is not what you would want for incorporation in a scientific study. The likelihood of Type 1 errors would be less than this because the documentation of accepted records is generally much better, imparting a higher certainty to the decisions. Therefore I believe that the total "error" rate in these hundred records is likely less than 3%. That seems quite acceptable. As others have stated, the non-accepted records are not rejected; they are mentioned in committee reports and archived so others can study them if desired. Indeed, occasionally one might be reconsidered and voted on again in light of additional information. We also receive some records each year of "review species" that have been documented by a single observer without any other supporting evidence (photo, video, audio, a feather, or a description from a confirming observer). We decline voting on these, and Ted might consider that equivalent to a Type II error, because most are possibly or probably correct identifications. Someone once wrote me that we are suppressing information by not voting on these records, but that's not true. We enumerate them in our published reports and archive the documentation. Most are also mentioned in seasonal reviews published in the N.B. Naturalist, and some in North American Birds. They simply do not meet our minimum requirement for voting. That requirement helps encourage most observers to make the effort to get photos and additional people to confirm their sighting and it doesn't put the committee in the position of having to make a judgment on the word of just one person. In addition to records voted on, our committee also receives documentation of twice as many records of less unusual occurrences, which we acknowledge and preserve without voting. >Would you want for, say, a climate change researcher to exclude from >analysis "rejected" reports of extralimital birds that really were >there? If he thinks it is useful, that researcher is free to examine the data and make his own judgment on what is valid. If using other criteria, one might come to somewhat different decisions than a records committee. Generally, however, I suspect there wouldn't be much difference. > >Let's go back to that Golden-winged Warbler.... > >Anyhow, the record is rejected 10-2, with the comment that the 2 >members felt that the report wasn't entirely conclusive. In many records committees I suspect that this vote would result in the record being accepted. If not, then it would be recirculated with each member seeing the reasons for the others' choice. >To make this less fanciful, I really do know a climate change >researcher who will not use BRC data because of their unacceptably >high Type II error rates. This person's conclusion is that BRC >deliberations are nice for policing recreational bird listing (and >there's nothing wrong with that) Which committees police recreational birding? I look at the role of a records committee as being to encourage observers to document unusual occurrences, to review the quality of the most unusual ones, and to preserve the data for study. This provides some order and a standard against which to compare the sometimes astonishing array of reports that surface on the Internet and elsewhere. I believe this does have some use to science. >, but that BRC deliberations are not relevant to science. Don't like >that? Then, at the very least, BRCs should strive for a balance >between Type I and Type II errors. And, honestly, where there >remains uncertainty, I'd say err on the side of including bad data, >rather than excluding good data. If correct identifications are poorly supported by evidence, we can't include them as "good data", except probably by also including as much "bad data." I don't think that's an improvement. >At 2:28 PM -0400 6/30/08, Phil Davis forwarded Ted Floyd's message >of Mon, 30 Jun 2008 10:06:22 -0700: > > >>More to the point, though, how can a BRC possibly know? We live in >>a post-Heisenbergian world. Sheesh, we live in a post-Darwinian >>world. There is no certainty. We can assign various levels of >>probability ("likely," "very likely," "exceedingly likely," "not so >>likely," whatever), but we just cannot say "accept" or "reject." In >>most of my submissions to BRCs, I make a point of saying that I'm >>not certain of the identification. (I like how the California form >>specifically asks the question of certainty, and I typically say, >>No, I'm not certain.) I can see assigning probabilities but wonder whether the results would be much different. I suspect that votes would group at 80-100% (i.e. accepted) and at 30% or less (not accepted), with very few in between, except in cases where there is some question about natural occurrence. -- David Christie Mary's Point, Harvey, Albert Co., New Brunswick, Canada http://www3.nbnet.nb.ca/maryspt/BRC -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Laurent Raty <l_raty AT HOTMAIL.COM> Date: Tue, 1 Jul 2008 14:21:07 +0200 Ted, > Practically speaking, Type II errors are often more egregious than > Type I errors. Would you want for your doctor to miss ("reject") a > cancer diagnosis? Would you want for a judge to imprison ("reject") > an innocent person? Would you want for, say, a climate change > researcher to exclude from analysis "rejected" reports of > extralimital birds that really were there? Any error could always conceivably have bad consequences, be them Type 1 or Type 2. Your doctor could also remove you a sane lung, the judge/jury could fail to imprison a serial killer and let him free in your immediate neighborhood, and incorrectly accepted records of a very rare bird could divert protection funds in favor of an area of little actual biological value, leaving more important places unprotected... Besides, although I've already seen the same comparison several times elsewhere, I believe the judge/jury case is not comparable to the other two in this way. Missing a cancer diagnosis or excluding an out-of-range bird record from an analysis are failures to reject the null hypothesis (the patient should normally not have a cancer; the bird should normally be whithin its range), and are indeed Type 2 errors. But this is /not/ true of a jury imprisoning an innocent person: here the null hypothesis is that this person is innocent and, if this gets wrongly rejected, this is a Type 1 error. If you want to compare a committee to a jury, it makes more sense to do it in the following way, in my opinion: - The committee is there to assess the possibility that an unusual/ unexpected behavior occurred in a bird, resulting in an out-of-range record. The jury is there to assess the possibilty that a deviant behavior occurred in a person, resulting in a crime. - The reporting of the presence of the bird at a given location and time by some observers is comparable to a witness testimony during a process. This should of course be presumed honnest/reliable unless there are good reasons to assume the opposite. But this honnesty/reliability is simply not the object of the verdict given by tha jury/committee. - The object of the verdict of the jury is the culpability of the accused person; in the case of the committee, the "accused" is the bird, and the "charge" is "vagrancy". - If declared guilty, the person goes to jail. If declared "guilty", the bird goes in the report. Cheers, Laurent - -- Laurent Raty l_raty AT hotmail.com Brussels, Belgium -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Ted Floyd <tedfloyd57 AT HOTMAIL.COM> Date: Tue, 1 Jul 2008 05:18:56 -0700 Hi, all. Thanks for the great discussion. Richard Heil writes: > In Ted's later posts regarding Type I and Type II errors, he is of > course correct that both are equally erroneous, and that records > committees should be wary of making either type of error. However, > Ted then creates unrelated and irrelevant examples (juries sending > innocents to prison, patients dying) that are extremely amicable and > beneficial to his tenuous point that records committees rejecting > records with insufficient, poor, or even contrary (!) evidence is > somehow evil or unscientific, when actually the opposite is true. How so? That is, how is it that the opposite is true? Let's say I've determined with 35% certainty that a large meteor will strike the earth in the next 10 years, i.e., I am 65% certain that it will not. Shouldn't I report my discovery, even with a large (65%) certainty that I am wrong? And let's say that I'm only 35% certain of my identification of an out-of-range Golden-winged Warbler. Shouldn't I report my discovery? Mightn't the result be of interest to a climate change scientist, or a bird navigation expert, or just somebody with interest in the natural history of Golden-winged Warbler? Richard cited examples of near-certainty (straight-out photo of Spotted Redshank, juvenile Bobolink misidentified as Grasshopper Sparrow), but I am increasingly persuaded that there are many examples that fall into the 65%/35% range. Certainly, Golden-winged Warbler is one such example. In last week's ABA workshops on documenting birds, we actually, believe it or not, got into a fine discussion of ontology-vs.-epistemology vis-a-vis Golden-winged Warbler. The ontological problem with Golden-winged Warbler, of course, is that the very definition of Golden-winged Warbler is problematic: What do you call a bird that looks like a Golden-winged Warbler, but that has the genes of a Blue-winged Warbler? What IS it? The question goes beyond mere epistemology (what do we know?) and gets is into the realm of ontology (what is it?). Of course, there are examples of near-certainty: Richard's juvenile Bobolink, Alan Wormington's White-throated Sparrow, etc. But there are also many examples of decided uncertainty: flight calls of Ammodramus sparrows and Passerina buntings, adult Hermit Warblers and Golden-winged Warblers with yellow-tinged breasts, overexposed photos of Willow Flycatchers and Thayer's Gulls, Azure Gallinules and Red-footed Falcons that may have been ship-assisted, Pink-footed Geese and Northern Cardinals that may have escaped from captivity, etc., etc. In my experience with bird records committees, such examples are legion. And in my experience, such examples are best treated as uncertain. We just don't know. And as to straight-out certainty, I'm just not buying it. Not since 1859. We now know that we live in a universe that is evolutionary (1859), relativistic (1905), uncertain (1927), impermanent (1929), incomplete (1931), chaotic (1961), entangled (1982), and otherwise fuzzy, blurry, and indeterminate. That's not political correctness or some "cult of uncertainty." That's 149 years of scientific progress. Old ideas die hard. A lot of folks still cling to Platonistic ideals of certainty, truth, and beauty. But not scientists. Not since 1859. We now know that it is better to say p=0.35, or p=0.99, or even p~1.00. But not p=1.00. Indeed, it is in the realm of those very small (p~0.00) and very large (p~1.00) probabilities where some of the most surprising of mathematical and philosophical issues play out. A fine introduction, in my opinion, is Brian Greene's Fabric of the Cosmos (Knopf, 2004). An important sub-theme--and it is one that runs throughout Fabric of the Cosmos--is that it is a mistake to equate near-certainty with certainty. The best we can say about the universe around us is that it is a continuum of uncertainties; and the best we can do is describe reality in probabilistic terms. For sure, that is a complete repudiation of Plato. But it does seem to be where modern science has taken us. Ted Floyd tedfloyd57 AT hotmail.com Lafayette, Boulder County, Colorado _________________________________________________________________ Need to know now? Get instant answers with Windows Live Messenger. http://www.windowslive.com/messenger/connect_your_way.html?ocid=TXT_TAGLM_WL_messenger_072008 -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: MD/DC Records Committee web page updates and Review List changes From: Phil Davis <pdavis AT IX.NETCOM.COM> Date: Tue, 1 Jul 2008 00:11:40 -0400 MD Osprey, MD Birds, BRCF-L: We have just updated a number of our MD/DC Records Committee (MD/DCRC) web site data products. Information highlights include: NEW SPECIES. No new species were added to the Maryland or DC Official Lists during this update cycle; however, potential new species under active consideration include: Maryland: Southern Lapwing, Cape Verde Shearwater, Western Meadowlark, Scopoli's (Cory's) Shearwater (subspecies) District of Columbia: Swainson's Warbler MARYLAND AND DC REVIEW LISTS. The following changes were made to the MD and DC Review Lists: Anhinga. There is a fairly even distribution of reports and records since the early 1990s; only one DC record. Removed from Maryland review list; retained on the DC review list. Sooty Tern. The pattern of displacement following hurricanes is well-established; however, only one accepted record for DC. Moved to Category 4B on MD list (reviewable if seen west of the coastal plain), retained in Category 2 on DC list (reviewable anywhere). Cave Swallow. Fall appearances since 2002 are well documented, no DC records. Removed from MD list. Painted Bunting. Averaging 3-4 reports and records per year for MD over the past 10 years; no records for DC. Removed from MD list. Trumpeter Swan. Although this taxon is on the Official List of Maryland Birds as an extirpated species, it has occurred in the region in recent years apparently as a result of reintroduction programs in the Northeast and upper Midwest. Other complications include possible escapes of hybrid "Trumpling Swans" from regional lakes and the Smithsonian research facility at Arlie, VA. The committee decided that it does want to continue to accept documentation of putative Trumpeter Swans, so the action is to add this species to MD review list. Since the species is not considered to be extirpated from DC and since there are no DC records, it would not be added to the DC Review List, but rather would be reviewed as a new species for DC, if observed there. Gambel's White-crowned Sparrow (subspecies). Reports have been averaging about 2 per year since 2000, with more reports in the 1990s; committee reviews of this taxon are backlogged, but most reports include images; no records for DC. Removed from the MD "Subspecies/Form" Review List. As with all unusual subspecies or forms, observers should continue to report sightings to MD Osprey and/or the seasonal editors of Maryland Birdlife and North American Birds. MD Review List: http://www.mdbirds.org/mddcrc/pdf/mdreview.pdf DC Review List: http://www.mdbirds.org/mddcrc/pdf/dcreview.pdf WEB DATABASES: Our historical canvass is nearing completion and many additional historical reports have been added to the MD and DC databases and further corrections made. For those keeping track, there are still a few taxa that need some work to verify earlier reports and dates, notable Le Conte's Sparrow. MD database: http://www.mdbirds.org/mddcrc/pdf/mddatabase.pdf (a very large PDF document ... search, but do not print!!!) DC database: http://www.mdbirds.org/mddcrc/pdf/dcdatabase.pdf (ditto, but not as large) 2008 ANNUAL MEETING MINUTES. This detailed report is posted here … http://www.mdbirds.org/mddcrc/pdf/rcannual2008.pdf Hope this helps ... Phil =================================================== Phil Davis, Secretary MD/DC Records Committee 2549 Vale Court Davidsonville, Maryland 21035 USA 301-261-0184 mailto:PDavis AT ix.netcom.com MD/DCRC Web site: http://www.MDBirds.org/mddcrc/rcindex.html =================================================== -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Rick Fridell <rfridell AT BURGOYNE.COM> Date: Mon, 30 Jun 2008 21:26:45 -0600 Hello, Despite my best intentions, I usually can't find time to get involved in these discussions, however, this one is getting too interesting not to comment. The Utah Bird Records Committee has gradually (very gradually) been shifting away from the traditional role, and have been moving more towards several of the concepts expressed in Ted's post. The basic philosophy shift has been to put more emphasize on the submitted records rather than the committee's review and individual committee member opinions of the records. Primarily thanks to an active and innovative secretary, Milton Moody, we've increased our efforts on archiving and maintaining electronic records and copies of associated photos, recordings, etc. including scanning many early records and photos. We still vote on the records and document the votes in an annual write-up. Since 2003 we've published the committee report and summary of the records in Utah Birds (the Journal of the Utah Ornithological Society). In the write-up we give equal weight to all the records, list the records in taxonomic order, publish all observer names with the records, and include the committee's vote at the end. Here's an example of how we handled a traditional "non-accepted" / "rejected" record: "Northern Bobwhite (Colinus virginianus). 2003-13. One adult male was observed near a backyard bird feeder in Emigration Canyon, Salt Lake Co., on 8 May 2003 by Susan and Jeffrey Saffle. Documentation provided by S&JS. Although there was no question that this was indeed a Northern Bobwhite, this species is not established in Utah, and thus, this individual was most likely an introduced bird. First Rd. 0 Y - 7 N." (Fridell et al. 2003, Utah Birds 17:75-81). Obviously this example is a bit more straightforward than others, but we use the same approach on obvious mis-identifications (0-7 first round votes) and difficult reviews that go three rounds and result in a split committee vote of 3-4 or 4-3. Personally, I do agree with Ted that all submittals are important, and the long-term scientific / ornithological value lies with the record itself. While the committee's review or individual committee member opinions may be important factors for inclusion on a state/ county/ region/ or observer's list, these factors are only somewhat relevant in the hypothetical example, of say, a researcher looking at patterns of vagrancy in Golden-winged Warblers. Having access to the records and all submitted documentation allows an interested researcher to come to their own conclusions based on whatever parameters or current information they are using at the time. Thanks for your time, respectfully, Rick Fridell Hurricane, Utah -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Subject: Re: Question on protocol in records committee journal reports From: Geoff Malosh <pomarine AT EARTHLINK.NET> Date: Mon, 30 Jun 2008 23:17:26 -0400 As per usual when iconoclastic ideas are voiced, both the iconoclast and his detractors argue by way of inappropriate analogy. A contentious Golden-winged Warbler record in Nevada is not the same as juror deliberations over life sentences; nor is the suggestion that uncertainty exists in the evaluation of bird records a nod to the philosophy of wishy-washy political correctness and the awarding of gold stars to D- students so they don't feel bad about themselves. Truth rarely exists at such extremes. I personally don't take such a dim view of Ted's analysis. He *is* correct, on a very fundamental level. The great scientific discovery of the 20th century was precisely that uncertainty exists in everything and not just at the quantum level, the corollary to which is that the act of observation alters the state of that which is observed; in other words observation itself introduces change to the system and uncertainty in the data observed. There really are very few absolutes in science and associated knowledge, and certainly the identification of every individual bird to the species level is not one of them, much less is the concept of species itself (see related discussion currently underway on BirdChat). Where the rubber meets the road, so to speak, is where the tolerances for uncertainties are set, and how they are applied to the final analysis. To bring the argument into the realm of records committees, consider not just the problems of genetic expression and species concepts but also the problem of questionable origin. Maybe, for instance, we have in hand an inarguable photo of a Redwing in Pennsylvania that was later confirmed by dozens of observers (which really happened -- Feb 2005 in Bucks County north of Philadelphia). How can it be shown to the committee that the bird actually arrived at the location on its own? Is it *really* not possible that it was released or escaped at the site from a local collection? Is it totally out of the question that it might have arrived on North American soil by stowing away on an airplane for a few hours? Besides these questions, without DNA analysis, can we truly know that it is a pure, genetically unsullied Redwing? Maybe it just looks like one but really is a second-generation descendant of some unusual pairing. Or maybe we don't know as much as we think about Redwing taxonomy. The fact of the matter is that any of these scenarios are possibilities, no matter how small, and if confirmed, any of these scenarios would compel the committee to "reject" the record. There *is* inherent uncertainty in the work of the committee that reviews this Redwing record, indeed in the examination of every record, as Ted argues. And as Ted suggests, the records committee could express their decision on the Redwing in terms of alphas and betas, and they would be entirely correct, and perhaps even more accurate, in doing so. But this is where Ted's ideal breaks down -- during the "application phase" of this knowledge. Even though there really is some undeniable uncertainty in our knowledge of Redwing occurrence in Pennsylvania, the alphas and betas involved are, statistically and practically, just a little ways south of totally meaningless. Yes, the bird could have been released, but the possibility is so remote that it is not worth calculating the percentages, much less authoring the associated qualifiers into the committee decision. This is true so much of the time that, when considering the practical application of knowledge, absolutely nothing additional is gained from a committee ruling that a Redwing in Pennsylvania is only 99.999% certain due to uncertainties, as opposed to doing what records committees do now, which is to "accept" (declare 100% certain) the record of the Redwing. The practical applications of every kind of science, from engineering to policy decisions to BRC deliberations, regularly (and correctly) disregard uncertainties this small. Ted's ideas are correct, but they must be applied at the macro (real-world) level only where appropriate. Some of the deliberations of a records committee, I have long thought, are best expressed in probabilities, especially in cases of questionable origin. For instance consider the occurrence of Black-billed Magpies in Ohio in late summer, which is another actual real-world example (September 2005 near Youngstown). "Accepting or Rejecting 100%" such magpies does the official ornithological record of Ohio a disservice, because neither approach is correct. The possibilities of natural vagrancy and assisted origin of magpies in Ohio are both non-negligible, and both unknowable. By either accepting or rejecting, the committee only guarantees that it will get it wrong half the time. Ted's ideas, I think, are correctly applied to a problem of this sort. They may also be correctly applied to some of the more thorny problems of species and subspecies definitions, for example, when evaluating out of place races of Dark-eyed Junco. But to extend this ideal to an expectation that the fourth and fifth decimals after the "99." be computed for every single record, and to outrightly reject the established concepts of "acceptance and rejection" simply for the sake of being rigorous, is missing the forest for the trees. Geoff Malosh Pennsylvania Ornithological Records Committee Geoff Malosh | Editor, Pennsylvania Birds 450 Amherst Avenue | Moon Township, PA 15108-2654 | 412.735.3128 pomarine AT earthlink.net | http://home.earthlink.net/~pomarine/index.html =========================================================================== Pennsylvania Birds is published by the Pennsylvania Society for Ornithology Preview the latest issue: http://www.pabirds.org/pabirds/pb_sample.html Subscription information: http://www.pabirds.org/PSOJoin.htm -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Richard Heil <rsheil AT COMCAST.NET> Date: Mon, 30 Jun 2008 20:40:16 -0400 Ted Floyd says there are no certainties, apparently oblivious to the fact that he is uttering one. Ted is sputtering absolute nonsense, the first part contradicts the second. Literally it is gibberish. Ted, are you certain there are no certainties? This is the 'fallacy of the stolen concept', which is displayed by the fact that one must use and therefore accept a concept, in this case "certainty", in any attempt to deny it. Quantum uncertainty, if correctly interpreted, is relevant only on the particle level, and is not very applicable to everyday life, or to everyday judgements required by, and necessary to maintain, life. If someone submits a straight-out photo of a Spotted Redshank, that is later confirmed by other observers at the reported site, how then is that not a certain record? Likewise, if someone submits a description and photo of an out of range Grasshopper Sparrow, the only problem being that the photo actually and unambiguously depicts a juvenile Bobolink, is not this record also equally certain - in this case false? Where exactly is the uncertainty in these examples? The former warrants an "accept", the latter warrants a "reject". Given a photo of a juvenile Bobolink, and asked the question: This is a Grasshopper Sparrow, true or false? How do you answer? Do you "accept" the proposition that it is a Grasshopper Sparrow, or do you "reject" the proposition that it is a Grasshopper Sparrow? I forgot, Mr. Floyd doesn't believe in the accept/reject, true/false dichotomy. Sorry, but this is just silly. In Ted's world apparently everyone is a winner (or a loser), no one is right, no one is wrong, everybody gets a trophy and a star on the forehead. There is no certainty. There are no truths, no falsehoods. This type of 'thinking', this gray cult of uncertainty, is an epistemological dead end, and is actually antithetical to science, which is knowledge based on principles and truths (laws). However, Infallibility is not a precondition to knowledge, and no thinking person will claim infallibility. Knowledge is contextual, if the evidence supports X and there is no evidence to support any alternative, then X constitutes knowledge in that context. The photo of the Bobolink submitted as a Grasshopper Sparrow and thus appropriately rejected has no value to the historical record whatsoever, other than to enlighten one that Bobolinks are sometimes mistaken as Grasshopper Sparrows. In Ted's later posts regarding Type I and Type II errors, he is of course correct that both are equally erroneous, and that records committees should be wary of making either type of error. However, Ted then creates unrelated and irrelevant examples (juries sending innocents to prison, patients dying) that are extremely amicable and beneficial to his tenuous point that records committees rejecting records with insufficient, poor, or even contrary (!) evidence is somehow evil or unscientific, when actually the opposite is true. Scrap the whole accept/reject dichotomy? I think not. Richard S. Heil S. Peabody, MA rsheil AT comcast.net Massachusetts Avian Records Committee -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Ted Floyd <tedfloyd57 AT HOTMAIL.COM> Date: Mon, 30 Jun 2008 13:58:13 -0700 Hello, all. My previous posting got cut off. (Not the fault of Phil, who forwarded it; not the fault of BRCF-L, either. If you're curious, and this is geeky and arcane, it had to do with my use of greater-than and less-than symbols that outgoing HoTMaiL can't handle.) Anyhow, as I was saying, my previous posting got cut off, and the cut-off portion is relevant to something that Alan Wormington said: > For most rejected records I fail to see how they are still > a part of the ornithological record. If someone submits a > report labelled "Golden-winged Warbler" and by the description > (or even submitted photo) it is determined that the bird was > actually a White-throated Sparrow, I fail to see how this > submission has any value (other than demonstrating that > misidentifications occur). Many rejected reports are less clear-cut. Many rejected reports are, in fact, probably correct. Committees can make two types of errors. They can falsely accept incorrect data (Type I error), and they can falsely reject correcet data (Type II error). Mathematically, logically, and philosophically speaking, Type I and Type II errors are equally, well, erroneous. They're both mistakes, plain and simple. Practically speaking, Type II errors are often more egregious than Type I errors. Would you want for your doctor to miss ("reject") a cancer diagnosis? Would you want for a judge to imprison ("reject") an innocent person? Would you want for, say, a climate change researcher to exclude from analysis "rejected" reports of extralimital birds that really were there? There is relatively little harm in committing a Type I error. You can simply go out and see for yourself. (Indeed, I note that there are good Bayesian methods for dealing with exactly such scenarios.) There is potentially great harm in committing Type II errors: People die of cancer, innocent people go to jail, early detection of climate change is thwarted, etc. Let's go back to that Golden-winged Warbler. Let's say that 2 committee members felt that the report could well have been correct, but that the report wasn't air-tight. If pressed, they would say that there was a 6% chance that the bird was something other than a Golden-winged Warbler. Conversely, and simplistically, there was therefore a 94% chance that it was a Golden-winged Warbler. Sampling distributions aren't actually that simple, and it is not always the case that p(it wasn't) + p(it was) = 1.0, but let's move on. Anyhow, the record is rejected 10-2, with the comment that the 2 members felt that the report wasn't entirely conclusive. Or: 10 jurors say innocent, 2 say probably innocent but aren't totally sure, and someone goes to jail. To make this less fanciful, I really do know a climate change researcher who will not use BRC data because of their unacceptably high Type II error rates. This person's conclusion is that BRC deliberations are nice for policing recreational bird listing (and there's nothing wrong with that), but that BRC deliberations are not relevant to science. Don't like that? Then, at the very least, BRCs should strive for a balance between Type I and Type II errors. And, honestly, where there remains uncertainty, I'd say err on the side of including bad data, rather than excluding good data. Or, better yet, just scrap the whole accept/reject dichotomy, report p-values (alpha and beta), and, well, I'm just repeating myself now, so I shall relent. Ted Floyd tedfloyd57 AT hotmail.com Lafayette, Boulder County, Colorado _________________________________________________________________ Need to know now? Get instant answers with Windows Live Messenger. http://www.windowslive.com/messenger/connect_your_way.html?ocid=TXT_TAGLM_WL_Refresh_messenger_062008 -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Fwd: Re: Question on protocol in records committee journal reports From: Alan Wormington <wormington AT JUNO.COM> Date: Mon, 30 Jun 2008 15:20:15 -0400 Everyone, A bit of a follow-up. First of all I disagree with some of what Ted says below. For most rejected records I fail to see how they are still a part of the ornithological record. If someone submits a report labelled "Golden-winged Warbler" and by the description (or even submitted photo) it is determined that the bird was actually a White-throated Sparrow, I fail to see how this submission has any value (other than demonstrating that misidentifications occur). To me its a junk record. (And in reality a lot of time and paper has been wasted.) And since the majority of rejected records pertain to misidentifications, I do not hesitate in making the statement that the majority of rejected records are thus junk, pure and simple. What future ornithological study would ever incorporate these rejected records? Back to the Ontario Bird Records Committee, where I stated that we do not list observer names for Rejected Records. I should also have stated that each of these observers receives a letter in the mail notifying them that their record was not accepted, and this mailing also includes the comments that were made by voting members during the review process. So we do acknowledge (and value) their submissions, we just don't list their name in the Annual Report. Alan Wormington Assistant to the OBRC Secretary On Mon, 30 Jun 2008 14:28:12 -0400 Phil Davis writes: > BRCF-L subscribers: > > FYI ... forwarded from Ted Floyd: > > Phil > > >From: Ted Floyd > >To: > >Subject: Re: Question on protocol in records committee journal > reports > >Date: Mon, 30 Jun 2008 10:06:22 -0700 > > > > > >Hi, all. > > > >As Donna says: > > > > > Whether any report is accepted or not, it is important to > > > acknowledge those that take their time to contribute to > > > the process and archive.... > > > >True, and, more to the point, those folks simply *are*, bottom > line, > >a part of the record. They have contributed to the ornithological > >record, period, and their data are very much a part of that record. > > >If we say that, in Nevada (just as a hypothetical example...), 3 > >reports of Golden-winged Warbler have been accepted and that 1 has > >been rejected, well, that 1 rejected report is clearly a part of > the > >ornithological record. Isn't it? Isn't it just as valuable as the > >"accepted" (see below) records? > > > >In a (somewhat) similar vein, I led 3 workshops on Bird Records > >Committees last week at the ABA Convention, and, if there was 1 > >thing and 1 thing only that everyone agreed on, it was the > >problematic, at best, distinction between "accepted" and "rejected" > > >(or "non-accepted," if you prefer) records. I feel more strongly > >than ever--especially now that it turns out that all of humanity > >agrees with me--that records ought not to be accepted or rejected. > > > >For one thing, that sort of distinction goes against 50+ years > (and, > >really, 400+ years) of scientific thinking. Assign a p-value, if > you > >must, and, of course, assign p for critical-alpha and > critical-beta. > >Many of the workshop participants rightly criticized BRCs for > >rejecting good records with unacceptably high critical-betas. > > > >More to the point, though, how can a BRC possibly know? We live in > a > >post-Heisenbergian world. Sheesh, we live in a post-Darwinian > world. > >There is no certainty. We can assign various levels of probability > >("likely," "very likely," "exceedingly likely," "not so likely," > >whatever), but we just cannot say "accept" or "reject." In most of > >my submissions to BRCs, I make a point of saying that I'm not > >certain of the identification. (I like how the California form > >specifically asks the question of certainty, and I typically say, > >No, I'm not certain.) > > > >I do acknowledge that "accept" vs "reject" has great value if we > are > >playing list-police. But not if our effort has scientific merit. > Not > >if our efforts might be useful to wildlife managers and > conservation > >biologists seeking to understand actual patterns of avian status > and > >distribution. If our work has scientific, professional, or any > >practical merit, then our job is to lay the data out there, to > >archive the data the best we can, to make qualitative ("believed by > > >two committee members to be inconclusive") or even quantitative > (p-alpha > > ================================== > Phil Davis Davidsonville, Maryland USA > mailto:PDavis AT ix.netcom.com > ================================== > > -------------------------------------------------- > Bird Records Committee Forum archives: > http://listserv.indiana.edu/archives/brcf-l.html > "The Early Worm Gets The Bird!" --- Alan Wormington -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Kimball Garrett <kgarrett AT NHM.ORG> Date: Mon, 30 Jun 2008 11:45:43 -0700 Just a brief comment from the California Bird Record Committee perspective. We certainly agree that observer names are a critical part of every record (and, in fact, a "report" doesn't even meet our minimal criteria for a "record" if there is no observer name attached to it). In our annual published reports we have always listed observer initials for all "Accepted" (sorry Ted) records, but have not done so for "Not Accepted, Identification Not Established" records. We do list all submitting observers in each annual report, so, indeed, one could go through the whole list of observers and deduce which ones submitted only "non-accepted" records. I wonder if anyone has ever bothered. Of course, all observer names are in the complete CBRC database and in the hard-copy files. Nevada's practices are therefore the same as California's (not surprising, since the current Nevada committee sought considerable advice on procedures from their neighbors to the west). The one thing I would add is that we do use initials rather than full names for observers in our annual reports (with a complete list of all submitting observers at the end of the report). Unlike Alan's situation in Ontario, we often have observers submitting 10-20 (or more) records per year. To always write out the full observer's name for each record would add considerable length to our published reports. The down side of using initials is the headache involved in avoiding redundant initials and achieving consistency in the initials used for each individual observer (the same headaches that NAB regional editors experience). Kimball Kimball L. Garrett Ornithology Collections Manager Natural History Museum of Los Angeles County 900 Exposition Blvd. Los Angeles CA 90007 (213) 763-3368 (213) 746-2999 FAX kgarrett AT nhm.org > -----Original Message----- > From: Bird Records Committee Forum [mailto:BRCF-L AT LISTSERV.INDIANA.EDU] On > Behalf Of Alan Wormington > Sent: Saturday, June 28, 2008 9:01 PM > To: BRCF-L AT LISTSERV.INDIANA.EDU > Subject: Re: Question on protocol in records committee journal reports > > Martin and all, > > In Ontario we have a system that is easy to follow and never has any > problems. > > FOR ACCEPTED RECORDS --- We always include the FULL name of contributors > for each and every record. Why would you want to list just initials of a > contributor? Is the record not important enough to provide as much > information (detail) as possible? I doubt if you need to list just > initials to save space. How many records is one observer likely to > submit in a single year? For the majority of observers I suspect it is > one only. -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Fwd: Re: Question on protocol in records committee journal reports From: Phil Davis <pdavis AT IX.NETCOM.COM> Date: Mon, 30 Jun 2008 14:28:12 -0400 BRCF-L subscribers: FYI ... forwarded from Ted Floyd: Phil >From: Ted Floyd >To: >Subject: Re: Question on protocol in records committee journal reports >Date: Mon, 30 Jun 2008 10:06:22 -0700 > > >Hi, all. > >As Donna says: > > > Whether any report is accepted or not, it is important to > > acknowledge those that take their time to contribute to > > the process and archive.... > >True, and, more to the point, those folks simply *are*, bottom line, >a part of the record. They have contributed to the ornithological >record, period, and their data are very much a part of that record. >If we say that, in Nevada (just as a hypothetical example...), 3 >reports of Golden-winged Warbler have been accepted and that 1 has >been rejected, well, that 1 rejected report is clearly a part of the >ornithological record. Isn't it? Isn't it just as valuable as the >"accepted" (see below) records? > >In a (somewhat) similar vein, I led 3 workshops on Bird Records >Committees last week at the ABA Convention, and, if there was 1 >thing and 1 thing only that everyone agreed on, it was the >problematic, at best, distinction between "accepted" and "rejected" >(or "non-accepted," if you prefer) records. I feel more strongly >than ever--especially now that it turns out that all of humanity >agrees with me--that records ought not to be accepted or rejected. > >For one thing, that sort of distinction goes against 50+ years (and, >really, 400+ years) of scientific thinking. Assign a p-value, if you >must, and, of course, assign p for critical-alpha and critical-beta. >Many of the workshop participants rightly criticized BRCs for >rejecting good records with unacceptably high critical-betas. > >More to the point, though, how can a BRC possibly know? We live in a >post-Heisenbergian world. Sheesh, we live in a post-Darwinian world. >There is no certainty. We can assign various levels of probability >("likely," "very likely," "exceedingly likely," "not so likely," >whatever), but we just cannot say "accept" or "reject." In most of >my submissions to BRCs, I make a point of saying that I'm not >certain of the identification. (I like how the California form >specifically asks the question of certainty, and I typically say, >No, I'm not certain.) > >I do acknowledge that "accept" vs "reject" has great value if we are >playing list-police. But not if our effort has scientific merit. Not >if our efforts might be useful to wildlife managers and conservation >biologists seeking to understand actual patterns of avian status and >distribution. If our work has scientific, professional, or any >practical merit, then our job is to lay the data out there, to >archive the data the best we can, to make qualitative ("believed by >two committee members to be inconclusive") or even quantitative (p-alpha ================================== Phil Davis Davidsonville, Maryland USA mailto:PDavis AT ix.netcom.com ================================== -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: "Donna L. Dittmann, ddittma" <ddittma AT LSU.EDU> Date: Mon, 30 Jun 2008 10:22:58 -0500 The Louisiana Bird Records Committee also does indicate in the published report the name of submitting observer(s) whose records are not accepted. However we do acknowledge and thank each and every one of them in the acknowledgement section along with anyone else who may have contributed additional information/discussion to published records contained in the report - here these observers are not obviously highlighted as authors of rejected records and can be thus thanked for their contribution to the report. Whether any report is accepted or not, it is important to acknowledge those that take their time to contribute to the process and archive.... Donna L. Dittmann LBRC Secretary Museum of Natural Science 119 Foster Hall Louisiana State University Baton Rouge, LA 70803 ph: 225-578-2009, fax: 225-578-2855 Ddittma AT lsu.edu On 6/28/08 9:13 PM, "Martin Meyers" wrote: > I'm curious what the prevailing policy/practice is regarding identifying > those who report sightings which are not accepted. > > The Nevada committee does not list any identification information when it > publishes a record which was not accepted. This policy is even in the > bylaws, and I've followed it both in the report published in Great Basin > Birds (the journal of the Great Basin Bird Observatory) and on the NBRC > website. (Whether I agree with that policy or not is a different issue...) > > But I've come up with a situation I'm not quite sure how to handle. > In the journal write-up for each individual (accepted) record, I include the > initials of the submitter(s). (I also put the full name into at least one > record, just because it seems like a nice idea, but that's irrelevant to > this discussion.) At the end of the annual report, I am including a list of > "contributors", at least in part to allow readers to look up the full name > for any of the initials. > > The question I (and others on the committee) are wrestling with is whether > or not to include the names of "contributors" who have no accepted records > in the report (although they did "contribute" records.) I feel like leaving > them out is a slap in the face -- it makes it appear that we don't consider > those people to have contributed. On the other hand, it does create the > possibility that a reader could see someone's name, then go look to see what > records they submitted, not find any, and conclude (correctly) that all of > that person's submissions were rejected ("non-accepted", in our more gentle > wording). And this would (I suppose) go against at least the spirit of the > bylaws. (I could, of course, change the heading "contributors" to something > like "Index to submitter initials" or something like that, I suppose, if the > final decision is to only include the names of those with accepted records.) > > Anyway, I'd be very interested in knowing how various committees deal with > the whole question about identifying submitters of non-accepted records. > > Thanks. > > Martin > (from the smoke-filled skies of Truckee, CA) > > ---------------------------------------------- > Martin Meyers > Secretary, Nevada Bird Records Committee > website: www.gbbo.org/nbrc > email: nbrc AT gbbo.org > > -------------------------------------------------- > Bird Records Committee Forum archives: > http://listserv.indiana.edu/archives/brcf-l.html -------------------------------------------------- Bird Records Committee Forum archives: http://listserv.indiana.edu/archives/brcf-l.html Subject: Re: Question on protocol in records committee journal reports From: Phil Davis <pdavis AT IX.NETCOM.COM> Date: Sun, 29 Jun 2008 01:58:30 -0400 Hi Martin, et al. We basically use the same system as Ontario - names for accepted records; no published or posted names for not accepted records. You do raise an interesting issue about "tha