Wikipedia:Reliable sources/Noticeboard/Archive 313
From Wikipedia, the free encyclopedia
| This is an archive of past discussions on Wikipedia:Reliable sources/Noticeboard. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current main page. |
| Archive 310 | Archive 311 | Archive 312 | Archive 313 | Archive 314 | Archive 315 | → | Archive 320 |
Dogsbite.org, other dog attack-related advocacy websites
We have an AfD going on for list of fatal dog attacks in the United States and one of the questions people have about the page is about reliable source coverage of the subject as a group.
Among the websites offered for evidence of this coverage are dogsbite.org, dogbitelaw.com, daxtonsfriends.com, animals24-7.org, nationalpitbullvictimawareness.org, and fatalpitbullattacks.com
Not looking for opinions on notability here, of course, nor for people to weigh in at that AfD. Just for opinions as to whether advocacy websites, law firm websites, etc. like these should be considered reliable sources in this context.
It's very much not my topic area -- it just seems like I keep coming across thinly veiled anti-pitbull advocacy on Wikipedia lately, and I want to make sure I'm not missing something about these kinds of sources. I know that in some areas there are advocacy organizations routinely do high quality work, and perhaps these are among them. — Rhododendrites talk \\ 20:29, 12 September 2020 (UTC)
- All unreliable - self published sources by non-experts whose statistics and inference from said statistics are at odds with the expert veterinary literature. PearlSt82 (talk) 20:59, 12 September 2020 (UTC)
- Can you cite some examples? ImTheIP (talk) 05:02, 13 September 2020 (UTC)
- Sure - Animals 24-7 states that "Of the 210 fatal dog attacks occurring since January 1, 2010, 138––66%––have been inflicted by pit bulls.", and dogsbite.org makes similar claims that "From 2005 to 2019, pit bulls killed 346 Americans, a rate over 6.5 times higher than the next closest breed, rottweilers, with 51 deaths.". Dog bite statistics are not collected by the CDC by breed and dogsbite.org's statistics are disputed by some groups as misleading - this article cites the American Veterinary Society of Animal Behavior as an example. Dog bite statistics by breed are fraught with error, as noted in a 2006 Journal of Veterinary Behavior paper which states
Attack data are often seriously flawed with respect to collection, reporting, and analysis. In the United States, the term “pit bull” does not mean APBT: it is a generic term that includes all the bull and terrier breeds, and sometimes the other bull breeds such as boxers, bull mastiffs, American bulldogs (Rowan, 1987). Breed identification is seldom ver-ified or consistent (Beck et al., 1975), and even experts cannot always tell whether a dog is a pit bull (Rowan, 1987). More seriously, breed identification often is based upon newspaper accounts.
A 2013 paper in the JAVMA analysizes all US fatal dog attacks from 2000–2009 and notes that breed was only reliably identified for 17.6% of cases. Likewise, in a 2017 Journal of Applied Animal Welfare Science paper, the authors found that large portions of hospital intake statistics are incomplete or unknown with regard to the breed, an example isFor example, Dwyer, Douglas, and van As (2007) claim that “pit bull terriers” and German Shepherd dogs—presumed to be correctly identified—were the most common breeds in their study to “attack” children, but they note that only 1% of their sample reported the presumed breed of the dog, a percentage far too low to permit generalizations about the other 99% of the sample.
- Breed-specific legislation is opposed by the AVMA and the AVSAB, two of the major veterinary bodies in the United States on the grounds that it is not effective at reducing overall dog bite incidences, the AVSAB position statement in particular references several studies that shows overall dog bite incidences do not change when pit bull populations approach near zero. Dogsbite.org doesn't seem to dispute this, and their claims that BSL is effective specifically states laws reduce only the amount of pit bull attacks - not overall dog attacks or hospital intake.
- Dogsbite.org was also critized for their use of the term "science whore" to describe members of the scientific community. This mentioned by author Bronwen Dickey in both her book Pit Bull: The Battle over an American Icon and in an interview with Psychology Today. Radio Canada also criticized dogsbite.org for the use of the term - the site responded to this criticism saying that this term has only been used three times since the creation of the site in 2007, which is not a particularlly compelling defense. Perhaps more worrysome is that after a great deal of discussion in Feb 2019 on the Talk:Dogsbite.org page about the term "science whore", dogsbite.org scrubbed all the logos and branding off the maultalk "Science whore" entry as of the April 23 2019 snapsnot, indicating some kind of off site coordination to influence Wikpedia discussion. PearlSt82 (talk) 07:29, 13 September 2020 (UTC)
- But the specific claims you are objecting to are not contradicted by the literature. I did a literature search and found that it is not implausible that, currently, most fatal dog bites in the US are inflicted by pit bull type breeds: "Bite risk by breed from the literature review and bite severity by breed from our case series were combined to create a total bite risk plot. Injuries from Pitbull's and mixed breed dogs were both more frequent and more severe. " Essig Jr. et. al 2019 "During 1997 and 1998, at least 27 people died of dog bite attacks (18 in 1997 and 9 in 1998). At least 25 breeds of dogs have been involved in 238 human DBRF during the past 20 years. Pit bull-type dogs and Rottweilers were involved in more than half of these deaths." Sacks et. al 2000 Dog-Bite-Related Fatalities 1998 It is of course possible, as you suggest, that this research is based on faulty statistics, but that wouldn't make the conclusions false, only unproven.
- My point is that if you object to "Of the 210 fatal dog attacks occurring since January 1, 2010, 138––66%––have been inflicted by pit bulls" you should challenge the statement directly. Phillips publishes the data he supports his claims with so whether he is doctoring the data or not can be shown. For example, you could go through his list of dog bite fatalities for 2018 and show that he has incorrectly listed the breed as pit bull on one or more entries. To me, that would be conclusive evidence that the guy is a crackpot. ImTheIP (talk) 16:12, 13 September 2020 (UTC)
- The Radio Canada source does exactly this. Through Google Translate:
Here are a few cases that Animals 24-7 considers to be pit bull deaths: In 2009, in Wisconsin, a 55-year-old woman, Louanne Okapal, died after being punched in the face by her horse. The horse had been frightened by a pit bull. In 2009, a 48-year-old Connecticut woman, Teresa Foss, died of a head injury after being hit by a pit bull. The dog hadn't bitten her. In 2010, 64-year-old Texas man Richard Martratt stabbed a pit bull and shot a catahoula after the two dogs attacked a border collie on his property. The man was not attacked by the dogs, but when authorities arrived he collapsed and died of a heart attack. In 2010 in Georgia, 14-year-old Miracle Parham fled after being frightened by a dog that witnesses described as a pit bull. She was fatally hit by a car. In 2013, 63-year-old James Harding was hit by a car after trying to pull away from two pitbulls. A 6-year-old girl was strangled by a chain to which a pit bull was attached. The year and location are not specified.
andAnother case cited on DogsBite.org concerns a 57-year-old man from Tennessee, James Chapple, who suffered serious hand injuries from pit bulls in 2007. Four months later he died of atherosclerosis and alcoholism problems. Even so, DogsBite.org counts it as a death caused by pit bulls.
PearlSt82 (talk) 16:23, 13 September 2020 (UTC) - Another case, Julia Mazziotto is reported by dogsbite.org as a pit bull fatality, but page 204 of Pit Bull: The Battle Over an American Icon states
The Lee Institute’s forensic experts determined that Julia Mazziotto’s wounds had been inflicted postmortem and that the elderly woman had actually died of a cardiac arrhythmia. The dogs had not been involved in her death, but after the fact they had pawed, scratched, and bitten her body.
PearlSt82 (talk) 17:02, 13 September 2020 (UTC)
- The Radio Canada source does exactly this. Through Google Translate:
- Sure - Animals 24-7 states that "Of the 210 fatal dog attacks occurring since January 1, 2010, 138––66%––have been inflicted by pit bulls.", and dogsbite.org makes similar claims that "From 2005 to 2019, pit bulls killed 346 Americans, a rate over 6.5 times higher than the next closest breed, rottweilers, with 51 deaths.". Dog bite statistics are not collected by the CDC by breed and dogsbite.org's statistics are disputed by some groups as misleading - this article cites the American Veterinary Society of Animal Behavior as an example. Dog bite statistics by breed are fraught with error, as noted in a 2006 Journal of Veterinary Behavior paper which states
- Can you cite some examples? ImTheIP (talk) 05:02, 13 September 2020 (UTC)
- PearlSt82, your blind hatred for Clifton and Lynn comes out yet again. ImTheIP was referring to attorney Kenneth Phillips "doctoring the data or not" and you go on and on about Clifton and Lynn. By the way, according to the website DogsBite.org, Lynn doesn't "count" Chapple as a dog bite related fatality. See here where it says it isn't counted in the statistics and here where it discusses (in 2013) the dispute over the official cause of death. Julia Mazziotto was killed in 2002 and that pre-dates DogsBite.org's collection (which started in 2005, I believe), so that is not counted. From the New York Times article:
"An 80-year-old woman was fatally mauled in her home here on Monday by two pet pit bulls, the authorities said today. The woman, Julia Mazziotto, was bitten or clawed over 80 percent of her body, and the Bergen County medical examiner found that she had died of severe mutilating wounds inflicted by the dogs, said the county prosecutor ... the older of the two dogs ... was covered with blood..."
Allegedly, according to Dickey pages 176-178, the daughter of the victim wanted her dogs back and so hired a third party (Lee Institute) to reevaluate the findings. Despite Lee's opinion that the dogs didn't kill the victim, a judge in the matter was unconvinced. So why are we second-guessing a judicial opinion on Wikipedia? This isn't helping Dickey's reliability (the evidence keeps piling up). By the way, here is a video of Mr. Phillips testifying before the Tennessee Senate Judiciary Committee in 2007 on changing their dog bite law after Mr. Chapple's death. Chapple's testimony was video'd from his hospital bed to present to the committee. The committee changed the Tenneessee law. Normal Op (talk) 18:13, 13 September 2020 (UTC)- Kenneth Phillips heavily bases his statistics off of dogsbite.org and Animals 24-7, so they're all connected together. From the link you shared:
The results are reported below. More details about these attacks can be found on the authoritative and comprehensive web site of DogsBite.org, in the section on bite statistics.
. Dogsbite.org does indeed count Julia Mazziotto as a pit bull fatality - see the New Jersey Section here. PearlSt82 (talk) 18:18, 13 September 2020 (UTC) - And Chapelle is indeed listed as a fatality by dogsbite.org here. PearlSt82 (talk) 18:20, 13 September 2020 (UTC)
- Kenneth Phillips heavily bases his statistics off of dogsbite.org and Animals 24-7, so they're all connected together. From the link you shared:
- PearlSt82, your blind hatred for Clifton and Lynn comes out yet again. ImTheIP was referring to attorney Kenneth Phillips "doctoring the data or not" and you go on and on about Clifton and Lynn. By the way, according to the website DogsBite.org, Lynn doesn't "count" Chapple as a dog bite related fatality. See here where it says it isn't counted in the statistics and here where it discusses (in 2013) the dispute over the official cause of death. Julia Mazziotto was killed in 2002 and that pre-dates DogsBite.org's collection (which started in 2005, I believe), so that is not counted. From the New York Times article:
- That's not a valid argument ("heavily bases"). I could say the same thing about ACF/NCRF/NCRC/Delise/AFF and every single one of their "commissioned" studies which you trot out with regularity. Good grief! The first link you point to is talking about legislation in each state, compared to fatal attacks in that state. Doesn't mean that DogsBite.org "counts" these in a statistics report, which is what you are inferring. And the second link was created in 2009, long before the stink in 2013 over Chapple's 2007 cause of death. I can't even begin to imagine maintaining old information on a several-hundreds-pages and over ten-year-old website to cater to nitpickers. And if you actually READ the page about the Chapple conflict, you would understand why the average person WOULD count him as a dog bite related fatality. Need I quote from it? Or will you actually read it? Normal Op (talk) 18:41, 13 September 2020 (UTC)
- Dogsbite.org is counting the Chapple in its statistics though - in the page you linked here, it has an explanatory note about the controversy over cause of death, but does not say at all it is withdrawing it from the list of fatalities. On that page it says
35 U.S. dog bite-related fatalities occurred in 2007.
, and there are 35 incidents on the list, of which Chapple is one. PearlSt82 (talk) 18:58, 13 September 2020 (UTC)- [I wrote this commen before Normal Op and PearlSt82's latest exchanges so excuse me if some things are not relevant.] It's unfortunate that Rhododendrites listed many different sites under a common heading. They are not published by the same groups and it's very possible that some of them are reliable while the others are not. For example, your Radio Canada report (reproduced on the AntiBSL Reddit sub) mentions multiple inaccuracies in Merritt Clifton's data but only one in Colleen Lynn's. These series appear to be different; Clifton lists pit-bull-related deaths while Lynn lists dog-bite-related deaths.
- Regarding James Chapple. In my opinion, counting it as a dog bite fatality is correct. From the reporting: "Chapple's left arm was amputated to the elbow after the dogs tore his hand off in an attack ... Chapple's right arm was severely injured. He is recovering at the Regional Medical Center of Memphis."The photo of him in his hospital bed is gruesome. The dog attack clearly contributed to his death even if he was an alcoholic.
- Regarding Louanne Okapal. She is listed here under the heading Notes on unusual cases: "Louanne Okapal, 55, of Sauk County, Wisconsin, suffered severe facial injuries on February 14, 2009 when a pit bull attacked a horse she was saddling and the horse kicked her in the face." She didn't die, and contrary to Radio Canada's allegation, Clifton doesn't claim she did.
- Regarding Julia Mazziotto. Bronwen Dickey in Pit Bull: The Battle Over an American Icon doesn't cite any sources and she most definitely underplays Mazziotto's wounds: "Alaimo contacted the Henry C. Lee Institute of Forensic Science in West Haven, Connecticut, and asked it to reevaluate the medical examiner's findings. The Lee Institute's forensic experts determined that Julia Mazziotto's wounds had been inflicted postmortem ... The dogs had not been involved in her death, but after the fact they had pawed, scratched, and bitten her body." It's clear from context that Alaimo was the owner of the dogs and that s/he contacted the Institute in order to avoid having to put them down.
- If these examples are the best the "pro-pitbull" side can come up with (and they clearly have a vested interest in finding errors in Lynn's and Clifton's data), then that tells me that the data actually is fairly trustworthy. Good enough to source statements like "According to dogbites.org, blah" or "According to Colleen Lynn, bleh" imho. ImTheIP (talk) 19:12, 13 September 2020 (UTC)
- Dogsbite.org is counting the Chapple in its statistics though - in the page you linked here, it has an explanatory note about the controversy over cause of death, but does not say at all it is withdrawing it from the list of fatalities. On that page it says
- That's not a valid argument ("heavily bases"). I could say the same thing about ACF/NCRF/NCRC/Delise/AFF and every single one of their "commissioned" studies which you trot out with regularity. Good grief! The first link you point to is talking about legislation in each state, compared to fatal attacks in that state. Doesn't mean that DogsBite.org "counts" these in a statistics report, which is what you are inferring. And the second link was created in 2009, long before the stink in 2013 over Chapple's 2007 cause of death. I can't even begin to imagine maintaining old information on a several-hundreds-pages and over ten-year-old website to cater to nitpickers. And if you actually READ the page about the Chapple conflict, you would understand why the average person WOULD count him as a dog bite related fatality. Need I quote from it? Or will you actually read it? Normal Op (talk) 18:41, 13 September 2020 (UTC)
- Another difference between Clifton/Animals24-7 and Lynn/DogsBite.org is that Clifton covers "serious attacks", not just deaths. DogsBite.org covers fatalities. Their styles are completely different even if their coverage overlaps. Despite PearlSt82, Dickey and Radio Canada trying to paint them with the same paint brush in the same stroke, they are two separate individuals working different angles. PearlSt82, you're right about the 35 count; I was mistaken. You made me count the fatalities on that page (horrible deaths!) and I would count Chapple, too, as a dog bite related fatality. However, you still don't seem to have read the link I sent, so I will quote from it:
"Ms. Delise discusses the death of James Chapple Jr. She states that “Mr. Chapple received severe injuries but fully recovered and was discharged from the hospital.” Mr. Chapple’s left arm was amputated, his right arm was badly mauled. A full recovery is impossible in this circumstance. Mr. Chapple’s injuries were so severe that a bill changing Tennessee law regarding vicious dogs was introduced. Video equipment was set up in Mr. Chapple’s hospital room so he could testify to legislators. Mr. Chapple lived long enough to see the bill signed into law. As a hospital nurse, I recognize that there are several reasons for discharge from the hospital, one is recovery, and another is that there is no further treatment that can be offered to the patient, they are discharged home with family care and Home Health nursing care. The listing of cardiovascular complications on the death certificate would not be unexpected. As a Cardiac Rehab nurse, I would expect cardiovascular deterioration in a newly disabled person with underlying coronary artery disease." — Carol Miller, RN
. ImTheIP is on the right track on the subject of discrediting others' work. Normal Op (talk) 20:44, 13 September 2020 (UTC)
- Another difference between Clifton/Animals24-7 and Lynn/DogsBite.org is that Clifton covers "serious attacks", not just deaths. DogsBite.org covers fatalities. Their styles are completely different even if their coverage overlaps. Despite PearlSt82, Dickey and Radio Canada trying to paint them with the same paint brush in the same stroke, they are two separate individuals working different angles. PearlSt82, you're right about the 35 count; I was mistaken. You made me count the fatalities on that page (horrible deaths!) and I would count Chapple, too, as a dog bite related fatality. However, you still don't seem to have read the link I sent, so I will quote from it:
- Kenneth Phillips (dogbitelaw.com) is a well respected attorney and a subject-matter expert. Here is one write-up about him in the Washington Post. (free account signup required). Normal Op (talk) 21:08, 12 September 2020 (UTC)
- Even if he is excluded, the rest are sensationalist, and sometimes hypocritical too. Dogsbite.org, for example, crows about their opponents "killing dogs", and then turns around and says that certain breeds should be banned, which they admit kills dogs. And should someone with such a heavy financial stake in the matter, like Phillips, be considered reliable? - Sumanuil (talk) 06:31, 13 September 2020 (UTC)
- DBRF lists are reliable; breed data is unreliable The issue with these "sources" is less about their lists of DBRFs (dog bite-related fatalities) which are fairly accurate; but instead, more about their claims and "statistics" related to the breeds attributed to the attacks. There are numerous studies that have concluded that breed information (without DNA) is subject to high rates of inaccuracy -- on average between 40%-60%. This study determined that breed information in media reports differed from breed determination by animal control agencies up to 40% of the time; this study determined that visual identification techniques for some breeds differed from actual DNA results up to 60% of the time; this study determined that accurate breed identification varied between 10.4%-67.7% (depending on the number of breeds identified in a dog's DNA); and finally, this study determined that DNA results matched visual breed identification only around 25% of the time. Essentially, the problem with attempting to identify and list the breed(s) involved with the DBRFs is that there is a high probability that either the shelter and/or the owner have not correctly identified the specific breed of their dog -- which then leads to inaccurate information in media reports as almost all media reports on dog bites simply identify the breed as reported by the owner/shelter (which again is likely to differ from the dog's actual DNA around 50% of the time, per cited studies). Regardless, there is a lot of evidence to support that breed information (especially in media reports, which is the *primary* source of breed data for these sources) is prone to high rates of inaccuracy without DNA. So while their DRBF lists may be fairly accurate, their breed data/information is highly suspect. The fact (per multiple studies on canine DNA) is that accurate breed identification by DNA highly contradicts casual breed assumptions by visual inspection (breed assumptions by owners, shelters, etc.). Furthermore, there are far too many bulldog-type (aka "bully-type") breeds (and even more bully-type mixes) with very similar physical characteristics to the 3-5 breeds commonly classified as pit bulls and attempting to classify every single one of these different, unique breeds (and mixes) as "pit bulls" is entirely inaccurate; therefore, the fact that these sites attempt to claim that their breed information is accurate/factual is highly problematic and leads to well-warranted questions around the agenda and biases of the sources listed above.Michael2468b (talk) 20:54, 13 September 2020 (UTC)
- Here is what the study Co-occurrence of potentially preventable factors in 256 dog bite-related fatalities in the United States says
For single dog incidents (148 incidents), on the basis of the strict definition (exact match), breed descriptors in media reports were discordant for 32 of 148 (21.6%) dogs; animal control or local law enforcement assessment of breed differed from the media account for 45 of 129 (34.9%) dogs. On the basis of the expanded definition (any agreement between alleged breeds and mixes), breed descriptors among media reports were discordant for 19 of 148 (12.8%) dogs; animal control or local law enforcement assessment of breed differed from the media account for 18 of 129 (14.0%) dogs.
That is, in 21.6% of all DBRFs, media outlets reported different specific breeds. In 78.4% of all DBRFs, they were in agreement. But the most interesting part is the last figure; in only 14.0% of all DBRFs did animal control and media reports diverge on what breed of dog was involved. FurthermoreWith respect to pedigree or results of DNA analysis for single dog cases, pedigree documentation, parentage, or DNA information was available for 19 dogs. These data were discordant with media reports for 7 of 19 cases on the basis of the strict breed definition and 0 of 18 cases on the basis of the expanded breed definition.
So, given the "strict breed definition", the media reports were correct in 12 out of 19 cases and in 18 out of 18 cases given the "expanded breed definition". It is obvious that someone who spends a lot of time investigating DBRFs can reach an even higher accuracy than what time-pressured journalists can. I.e, all else being equal, dogsbite.org must be a more reliable source for breed identification than media reports. - The three other studies find poor correlation between visual assessement and DNA testing (but a high correlation between assessors!). That is unsurprising because pit bulls are a hetereogenous group of breeds with similar morphological features. Taking the argument to its extreme, one couldn't ever know if a dog is a pit bull or not. But courts have ruled that dog owners can:
In sum, we believe that the physical and behavioral traits of pit bulls together with the commonly available knowledge of dog breeds typically acquired by potential dog owners or otherwise possessed by veterinarians or breeders are sufficient to inform a dog owner as to whether he owns a dog commonly known as a pit bull dog.
Regardless, there is a lot of evidence to support that breed information (especially in media reports, which is the *primary* source of breed data for these sources) is prone to high rates of inaccuracy without DNA.
Lynn also collects breed identification photos for most dogs:Of the 48 dog bite fatalities recorded in 2019, a record 81% (39) had some form of an identification photograph, the highest percentage since we began our collection effort. Pit bulls and their mixes represent 74% of all images collected in 2019. Of the 39 cases with breed identification photographs, 59% (23) contained images captured or republished by news media; 59% (23) contained images located on social media pages of the dog's owner or family members; and 49% (19) contained images that were the result of DogsBite research and otherwise may have gone unpublished.
- If breed identification is as difficult as you say, then dogsbite's data must contain tons of errors. Then it must be possible for someone to point out at least a few of them.
- tl;dr: For List of fatal dog attacks in the United States, dogsbite.org is a good source. The page needs a disclaimer explaining why dog classification is difficult and somewhat subjective. Deleting the page and/or blacklisting dogsbite.org is throwing out the baby with the bathwater. ImTheIP (talk) 03:13, 14 September 2020 (UTC)
- As mentioned previously, that site can be a source for DBRFs -- but not for valid (true/factual) breed information. All 4 studies (and there are other studies as well with similar conclusions regarding visual breed identification vs. actual DNA) still validate the inherent challenges and inaccuracies of visual breed identification vs. actual/true breed determination by DNA. The Co-occurrence study still specifically concludes:
"For 401 dogs described in various media accounts, reported breed differed for 124 (30.9%); for 346 dogs with both media and animal control breed reports, breed differed for 139 (40.2%). Valid breed determination was possible for only 45 (17.6%) DBRFs; 20 breeds, including 2 known mixes, were identified.
-- Valid breed determination for only 17.6% of the incidents does not infer confidence in the breed data attributed to DBRFs by any measure. The other 3 studies validate the discrepancies between casual breed assumptions by visual inspection vs. actual DNA results -- ergo pictures are not a reliable method for determining breed. Therefore, even if there were pictures for 100% of the dogs involved, that would still not serve to validate breed and any site/individual claiming otherwise is going counter to the conclusions in the studies on this subject matter -- importantly, the studies found that even those familiar with dogs and breeds (veterinarians, shelter staff, etc.) are prone to being unable to accurately identify a dog's breed(s). Furthermore, there are far too many bully-type breeds (and even more bully-type mixes) that look very similar to the 3-5 breeds classified as pit bulls to be able to assert (as a fact) that a dog is a specific breed based only on a picture or shelter label etc. If a site is going to make breed-specific assertions about risk; then that site needs to be breed-specific -- and what these studies have concluded is that without DNA, any breed-specific claims are likely to be inaccurate and therefore cannot be used as a reliable source (for breed information). The main issue with these sites is that while they are entitled to guess/assume/suspect the breed(s) of a dog, they cannot with any reasonable amount of certainly or authority claim that their breed assumptions are "facts" without DNA (as per cited studies) -- and doing so is at best disingenuous. Canine DNA is far too complex for visual breed identification to be accurate which is why DNA is needed and is currently the most accurate method for valid breed determination -- without DNA, breed is just a guess.Michael2468b (talk) 04:04, 14 September 2020 (UTC)- The Co-occurrence study? You mean "Co-occurrence of potentially preventable factors in 256 dog bite–related fatalities in the United States (2000–2009)" by Gary J. Patronek, VMD, PhD; Jeffrey J. Sacks, MD, MPH; Karen M. Delise; Donald V. Cleary, BA; Amy R. Marder, VMD. Two other Wikipedia editors picked apart the contradictions/paradoxes in that exact study recently — one here in this thread, and one on the Pit bull talk page. Now it's my turn. That 2013 study's conflict of interest disclosure simply stated "The National Canine Research Council supported the efforts of Karen Delise from 2006 to 2011 for assembly of case reports" with NO MENTION of her actual position within NCRC, which was Founder and Director of Research. Note that even in late 2012 (pre-publish) and still in 2014 (no changes there) Delise is listed on the website as "Founder & Director of Research". So how far do we go with believing NCRC and their commissioned studies when they lie on the conflict of interest disclosure? Or did it take two years to "shop around" for a journal that would take the study (explaining the discrepancy of 2011 versus 2013); in which case, do we trust that this is "good science"? Oh, I forgot to mention the earlier history. National Canine Research Council (NCRC) is a pit bull advocacy organization, formerly known as National Canine Research Foundation, closely associated with (and co-staffed) with former LOBBYING organization known as Animal Canine Foundation and Washington Animal Foundation. The first two were founded by and run by Karen Delise, who even copyrighted the NCRF website, and she co-staffed with the other two organizations. In 2004, Karen Delise self-proclaimed on her website "Karen also is the most versed expert on fatal dog attacks and consults the American Canine Foundation and other organizations. Karen has spent over ten years researching fatal dog attacks." And in 2014 is billed as "More than twenty years of research and investigation have led to Ms. Delise being considered the nation's leading expert on dog bite-related fatalities." In 2013 (the year the study was published), NCRC and its several predecessor organizations had all been run by the same person for well nigh 20 years with a deliberate and focused pro-pit bull agenda. And we're supposed to believe that her only contribution was "assembly of case reports" with no post title. And you want to consider that not-properly-disclosed 2013 study as a reliable source? At least three of those four studies you mentioned were commissioned and/or authored by persons connected with NCRC. I gotta say, in the same spirit as WP:BOOMERANG (but for debates) if you're going to drag one side of a divide through a reliable sources gristmill, you should be encouraged to look at the other side, too. Normal Op (talk) 05:25, 14 September 2020 (UTC)
- Point taken with respect to the "Co-occurrence of potentially preventable factors in 256..." study; nevertheless, the study's conclusions are similar to the conclusions found in multiple other studies (provided above) regarding the accuracy (or lack thereof) of visual breed identification vs. accurate breed determination by DNA. If needed, we can delete the Co-occurrence study entirely from this conversation and the point still stands that making assertions about breed(s) without DNA is proven to be relatively inaccurate. It's pointless for me to clarify this point further because that's exactly what the studies (sure, not including the "Co-occurrence..." study) have already done.Michael2468b (talk) 05:56, 14 September 2020 (UTC)
- I spent a long time reading through Co-occurence, trying to understand it so I don't think that it should be discarded. :) Regarding the numbers, 14.0% was the missclassification rate for single-dog attacks using the expanded definition. The high missclassification rate of 40.2% you cite is for multi-dog attacks using the strict definition. They define the strict definition as requiring an exact match. The example they give is illuminating:
Thus, pit bull and American Staffordshire Terrier would be concordant, but pit bull and pit bull mix would be discordant, as would American Staffordshire Terrier and American Staffordshire Terrier mix.
In other words, if media reported pit bull and animal control reported pit bull mix, they consider that a missclassification. Such errors are completely unsurprising as journalists aren't experts. In their expanded definition, they lump different types of pit bulls together:For the expanded definition, concordances related to pit bull-type dogs were considered when reported as pit bull, pit bull terrier, pit bull mix, pit bull terrier mix, American Pit Bull Terrier, American Staffordshire Terrier, Staffordshire Bull Terrier, or any alleged mix thereof.
Thus, yielding a much lower missclassification rate. The classification scheme used on the sites in question is, as far as I can tell much closer to the expanded definition than the strict one. - Breed determination was only possible for 45 dogs. That is low but given the expanded definition all dogs matched!
These data were discordant with media reports for 7 of 19 cases on the basis of the strict breed definition and 0 of 18 cases on the basis of the expanded breed definition. ... For multiple dog cases, pedigree documentation, parentage, or DNA information was available for 28 dogs. These data were discordant with media reports for 7 of 28 (25.0%) cases on the basis of the strict breed definition and 0 of 28 (0%) cases on the basis of the expanded breed definition.
And this is for the media reports which obviously are less accurate than the animal control reports. Interestingly the study doesn't report the breed of the 45 dogs determined, I wonder why! - You also have Toledo v. Tellings (2007) and other rulings telling dog owners that they indeed can tell if their dog is a pit bull or not. And a metastudy from 2019:
Bite risk by breed from the literature review and bite severity by breed ... Injuries from Pitbull's and mixed breed dogs were both more frequent and more severe.
The claim implies that breed determination without DNA is possible (or else the whole metastudy is bogus). - There are only about 30 to 50 DBRFs per year and they all lead to police investigations. I find it hard to believe that they wouldn't be able to correctly determine the breed in at least the majority of cases. At least it flies in the face of common wisdom; telling a dog owner with a dog of breed X that they can't know it is an X unless they have had it DNA-tested will likely upset them. tl;dr The claim that breeds can't be determined without DNA-evidence is extraordinary and runs counter to conventional wisdom. ImTheIP (talk) 19:53, 14 September 2020 (UTC)
- Understood; however, respectfully I don’t think anyone is making an extraordinary claim of mistaking Chihuahuas for Bulldogs or Dachshunds for German Shepherds. The challenge is with specific unique breeds within breed types/groups - for example, accurate visual identification for specific breeds within the diverse bully-type (bulldog-type) breed group such as the various pit bull type breeds and other similar breeds and mixes such as Dogo Argentino, American Bulldog, Cane Corso, Boxer dog, Alapaha Blue Blood Bulldog, and many others. And unsurprisingly (per the DNA studies) there is a higher level of agreement/accuracy for purebred dogs than for mixed breed dogs. However, accuracy declines when attempting to identify the difference between, for example, an American Staffordshire Terrier mix and an Dogo Argentino mix or an American Bulldog mix and a Cane Corso mix. These are just a few examples, there are many breeds in the broad bully-type category that have similar physical characteristics that unless they are purebred, are harder to accurately identify their predominant breed and secondary breed(s) without DNA. Not an argument, just a clarifying point with respect to misidentifying dog breeds. Also, the "Co-occurrence" report is not a DNA study, it's admittedly a study on the discrepancies between media reports and animal control based only on assumed breed by visual breed identification (not DNA) -- which are discrepancies that are still important to study, but it's not a study on visual breed ID vs actual DNA. The studies that analyzed the accuracy of visual breed identification vs. actual breed determination by DNA did find larger percentages of discrepancies - especially for mixed breed dogs.Michael2468b (talk) 20:23, 14 September 2020 (UTC)
- I spent a long time reading through Co-occurence, trying to understand it so I don't think that it should be discarded. :) Regarding the numbers, 14.0% was the missclassification rate for single-dog attacks using the expanded definition. The high missclassification rate of 40.2% you cite is for multi-dog attacks using the strict definition. They define the strict definition as requiring an exact match. The example they give is illuminating:
- Point taken with respect to the "Co-occurrence of potentially preventable factors in 256..." study; nevertheless, the study's conclusions are similar to the conclusions found in multiple other studies (provided above) regarding the accuracy (or lack thereof) of visual breed identification vs. accurate breed determination by DNA. If needed, we can delete the Co-occurrence study entirely from this conversation and the point still stands that making assertions about breed(s) without DNA is proven to be relatively inaccurate. It's pointless for me to clarify this point further because that's exactly what the studies (sure, not including the "Co-occurrence..." study) have already done.Michael2468b (talk) 05:56, 14 September 2020 (UTC)
- The Co-occurrence study? You mean "Co-occurrence of potentially preventable factors in 256 dog bite–related fatalities in the United States (2000–2009)" by Gary J. Patronek, VMD, PhD; Jeffrey J. Sacks, MD, MPH; Karen M. Delise; Donald V. Cleary, BA; Amy R. Marder, VMD. Two other Wikipedia editors picked apart the contradictions/paradoxes in that exact study recently — one here in this thread, and one on the Pit bull talk page. Now it's my turn. That 2013 study's conflict of interest disclosure simply stated "The National Canine Research Council supported the efforts of Karen Delise from 2006 to 2011 for assembly of case reports" with NO MENTION of her actual position within NCRC, which was Founder and Director of Research. Note that even in late 2012 (pre-publish) and still in 2014 (no changes there) Delise is listed on the website as "Founder & Director of Research". So how far do we go with believing NCRC and their commissioned studies when they lie on the conflict of interest disclosure? Or did it take two years to "shop around" for a journal that would take the study (explaining the discrepancy of 2011 versus 2013); in which case, do we trust that this is "good science"? Oh, I forgot to mention the earlier history. National Canine Research Council (NCRC) is a pit bull advocacy organization, formerly known as National Canine Research Foundation, closely associated with (and co-staffed) with former LOBBYING organization known as Animal Canine Foundation and Washington Animal Foundation. The first two were founded by and run by Karen Delise, who even copyrighted the NCRF website, and she co-staffed with the other two organizations. In 2004, Karen Delise self-proclaimed on her website "Karen also is the most versed expert on fatal dog attacks and consults the American Canine Foundation and other organizations. Karen has spent over ten years researching fatal dog attacks." And in 2014 is billed as "More than twenty years of research and investigation have led to Ms. Delise being considered the nation's leading expert on dog bite-related fatalities." In 2013 (the year the study was published), NCRC and its several predecessor organizations had all been run by the same person for well nigh 20 years with a deliberate and focused pro-pit bull agenda. And we're supposed to believe that her only contribution was "assembly of case reports" with no post title. And you want to consider that not-properly-disclosed 2013 study as a reliable source? At least three of those four studies you mentioned were commissioned and/or authored by persons connected with NCRC. I gotta say, in the same spirit as WP:BOOMERANG (but for debates) if you're going to drag one side of a divide through a reliable sources gristmill, you should be encouraged to look at the other side, too. Normal Op (talk) 05:25, 14 September 2020 (UTC)
- As mentioned previously, that site can be a source for DBRFs -- but not for valid (true/factual) breed information. All 4 studies (and there are other studies as well with similar conclusions regarding visual breed identification vs. actual DNA) still validate the inherent challenges and inaccuracies of visual breed identification vs. actual/true breed determination by DNA. The Co-occurrence study still specifically concludes:
- Here is what the study Co-occurrence of potentially preventable factors in 256 dog bite-related fatalities in the United States says
Anecdotal/personal |
|---|
| ** FWIW, as an anecdotal example, my neighbor has a smaller Dogo Argentino mix (it's 65 pounds whereas most Dogo Argentinos are around 100 pounds) and when she adopted him from the shelter, he was labeled as an "American Pit Bull Terrier mix". However, she bought a DNA test at the vet and the results were (if I remember correctly) something like 60% Dogo Argentino, 20% Labrador Retriever, and the balance of his DNA was a mix of several other breeds including Boston Terrier and Swiss Shepherd -- no mention of American Pit Bull Terrier in his DNA. So this is just an example of how dogs with bully-type ancestry and features commonly get misidentified as "pit bulls" just because of their physical characteristics. Also, I thought the dog was more of an American Bulldog mix - but I was wrong as well... and I'm pretty familiar with most of the bully breeds.Michael2468b (talk) 21:46, 13 September 2020 (UTC) |
- @Michael2468b: Per Wikipedia:Talk page guidelines,
"Do not use the talk page as a forum or soapbox for discussing the topic. The talk page is for discussing how to improve the article, not vent your feelings about it."
Normal Op (talk) 22:11, 13 September 2020 (UTC)- Fair enough; but to be clear, there was no "venting" - just an anecdotal account which I 100% made clear that it was only that -- an anecdotal account. Nevertheless, it is a real-world example of how breeds can very easily be misidentified.Michael2468b (talk) 22:22, 13 September 2020 (UTC)
- @Michael2468b: Per Wikipedia:Talk page guidelines,
- Not reliable - to begin, it is already proven that attempts to identify dogs by their appearances only are not reliable. Breed identification requires a positive ID based on corroborating official breed registry records and/or DNA testing. WikiProject Dogs has started a list of RS that may prove helpful. Atsme Talk 📧 22:08, 13 September 2020 (UTC)
- Thanks for the link, Atsme. I note that these sources aren't included there (as reliable or unreliable). Are they out of scope for that page? Maybe WPDogs wants to take on writing up some guidance about dog attacks? Of course, part of the reason I posted here was in the hope of attracting fresh eyes to this. There seem to be a few editors on Wikipedia primarily to argue one way or the other on this issue. — Rhododendrites talk \\ 22:21, 13 September 2020 (UTC)
- You're quite welcome, Rhododendrites. I posted notice of this discussion on the project page, and remain optimistic that we will be looking further into these types of sources in the days to come. We are a bit snowed-under right now trying to clean-up as much of the promotional material, misinformation and unreliable sourcing that our time allows. Cavalryman has been most diligent, and we welcome more volunteers to the project! Atsme Talk 📧 22:36, 13 September 2020 (UTC)
- Sorry if I'm missing this in the growing threads above, but is there really no source of these statistics that isn't an advocacy organization first and foremost? I have a hard time trusting the statistics of a website set up to make a certain conclusion which collects statistics that happen to support that conclusion (likewise a lawyer who profits from a particular view advocating for that view). It cuts both ways -- if someone were using statistics on, say, "savethepitbullz.com" (arbitrary -- no idea if it exists) I would likewise default to not considering it reliable.
- I also think it's important to separate the reliability of the statistics from the reliability of, well, everything else. We don't typically treat a database/statistics the same way we do coverage of a topic that uses those statistics, and there's a reason we try not to use opinion pieces, PR, propaganda, etc. as reliable sources, even if its data collection seems ok enough (there are exceptions, but they're hard-earned). — Rhododendrites talk \\ 22:15, 13 September 2020 (UTC)
- Statistics by breed aren't captured by veterinary orgs or the CDC due to their inherent unreliability - as far as I know these (DBO/A247) are the only organizations that publish these sort of statistics. PearlSt82 (talk) 22:20, 13 September 2020 (UTC)
- Wrong! Even NCRC and its predecessor NCRF published incident reports and lists of fatalities. And you know that! . And your excuse about the CDC has no basis in fact. Normal Op (talk) 00:45, 14 September 2020 (UTC)
- I was not aware of these sites, and would appreciate less WP:ASPERSIONS. PearlSt82 (talk) 01:10, 14 September 2020 (UTC)
- Wrong! Even NCRC and its predecessor NCRF published incident reports and lists of fatalities. And you know that! . And your excuse about the CDC has no basis in fact. Normal Op (talk) 00:45, 14 September 2020 (UTC)
- Statistics by breed aren't captured by veterinary orgs or the CDC due to their inherent unreliability - as far as I know these (DBO/A247) are the only organizations that publish these sort of statistics. PearlSt82 (talk) 22:20, 13 September 2020 (UTC)
- @Rhododendrites: To answer your earlier question, there's probably no neutral side publishing any related data because you can boil it down simply enough to two rabidly-opposed sides: (A) people who like or want to keep pit bulls as pets, and no-kill shelters full of pit bulls that need to offload them somewhere, and (B) people who view pit bulls as a public safety menace and want better oversight, accountability, and laws. Group A labels Group B as "anti-pit bull", most likely because pit bulls are over-represented proportionally in the collection of attacks so any efforts against dangerous dog hazards seems like an anti-pitbull agenda. Group A is well-funded (millions) and has created organisations such as NCRC, AFF & BFAS which can "fund studies" to prove their viewpoints (see AFF & BFAS Form 990s for proof that they are funding the studies that later show up on NCRC's website), then take their paid lobbyists to change legislation using their collection of studies as proof. Unlike Group A, Group B has no huge well-funded organizations, philantropists funding it, or lobbyists; this group includes DogsBite.org, Clifton, & Phillips. PearlSt82 posts only Group-A-favored studies (yes, I've checked) and vehemently opposes all medical studies that point to a higher incidence of pit bulls causing disfigurements or fatalities, and opposes any article that mentions favorably anyone from Group B. So sorry, no neutral middle ground to pull from, I'm afraid. Normal Op (talk) 01:34, 14 September 2020 (UTC)
- All unreliable, as stated above they are a collection of data selected to support the preferred conclusions of website creators. Cavalryman (talk) 22:32, 13 September 2020 (UTC).
- Sounds unreliable. On the other issue, readers may be interested in this British parliamentary committee report, Controlling Dangerous Dogs, wherein appear the words The British Veterinary Association and British Small Animal Veterinary Association, the RSPCA, Dogs Trust, Blue Cross, Battersea Dogs & Cats Home, the Kennel Club, and David Ryan, former Chair of the Association of Pet Behaviour Counsellors, were unanimous in their condemnation of the breed specific provisions in Section 1. of the Dangerous Dogs Act 1994 which banned dogs of "pit bull terrier type". (As noted, "pit bull terrier" is not per se a recognized breed in Britain.) The National Police Chiefs Council also said that "we would like to move away from a specific list [of banned types]" and told the committee
"traditional fighting dog lines had been diluted” to the degree that such dogs were often now found to be ‘near types’ that fell outside the Section 1 classifications."
One not-necessarily reliable source cites a weak methodology to claim Labrador retrievers attack the most or rather, that the most frequently owned breed is responsible for the greatest proportion of attacks for which insurance is claimed. (And maybe that Labrador owners are more likely have insurance than are pit bull owners.) GPinkerton (talk) 23:24, 13 September 2020 (UTC)- This is not an international discussion, but a discussion related to US fatal dog attacks, and is a spinoff of the discussion on Wikipedia:Articles for deletion/List of fatal dog attacks in the United States. Normal Op (talk) 00:53, 14 September 2020 (UTC)
- Yes I know, just wanted to point out that the issues are international and there is a List of fatal dog attacks in the United Kingdom which if its American equivalent is deleted will also require attention. GPinkerton (talk) 12:32, 14 September 2020 (UTC)
- This is not an international discussion, but a discussion related to US fatal dog attacks, and is a spinoff of the discussion on Wikipedia:Articles for deletion/List of fatal dog attacks in the United States. Normal Op (talk) 00:53, 14 September 2020 (UTC)
- Sounds WP:NOTRELIABLE due to WP:SELFPUB. An advocacy source might be a WP:BIASED RS, but I think that self-published material would only be included where some other source that is RS used it. Cheers Markbassett (talk) 00:56, 14 September 2020 (UTC)
- Comment: DogExpert.com/fatal-dog-attacks is another website discussing and listing fatal dog attacks. (I just now found it.) What do you guys think about this one? Normal Op (talk) 07:22, 14 September 2020 (UTC)
- Another personal for-profit website by someone involved in dog bite litigation. This time a professional witness rather than a lawyer. Doesn't mean he doesn't know what he's talking about, of course, but like some of the others there's (a) no editorial oversight, (b) the purpose is to make money, (c) advocacy. — Rhododendrites talk \\ 14:45, 14 September 2020 (UTC)
- Comment: Why are we even discussing this here. No one is using those websites as citations anyway and this conversation is a FORK of the conversation at Wikipedia:Articles for deletion/List of fatal dog attacks in the United States. The discussion belongs on that page. Any special angle of this conversation has already run its course and you've now circled back around to the AfD arguments. Normal Op (talk) 17:42, 14 September 2020 (UTC)
- A linksearch when I opened this thread showed that some of these sources do appear in articles (though, granted, not many). I opened this because whether they are reliable transcends that particular AfD, and if they are unreliable they're not relevant to that AfD. Don't know if it was the most efficient approach, but it does seem to be attracting a couple people who were not previously involved. — Rhododendrites talk \\ 18:11, 14 September 2020 (UTC)
- Comment/request: In previous comments, Normal Op made several statements to the effect that the National Canine Resarch Council should be treated as inherently unreliable, including when people connected with them are published in journals. However, this source is listed at Wikipedia:WikiProject_Dogs/Reliable_sources. Since its somewhat related to the discussion at hand, would this be an appropriate venue to establish reliability of that organization as well? PearlSt82 (talk) 18:22, 14 September 2020 (UTC)
- @PearlSt82: Was there any discussion that lead to it being listed? --Guerillero | Parlez Moi 21:52, 14 September 2020 (UTC)
- I don't think there was discussion on this, but based on their about page, I would think it would be at least reliable enough to quote with attribution. PearlSt82 (talk) 22:01, 14 September 2020 (UTC)
- I don't see any as reliable --Guerillero | Parlez Moi 21:50, 14 September 2020 (UTC)
Is Mark Durie as a source for Criticism of Joe Biden okay?
I came across following write up about Mark Durie. Would that be considered reliable source for Criticism of Joe Biden article (I am contemplating for).
https://www.jns.org/opinion/joe-biden-courts-islam-not-muslims/
Bookku (talk) 12:41, 18 September 2020 (UTC)
- I doubt it, why is his opinion worthy of note, who is he?Slatersteven (talk) 12:45, 18 September 2020 (UTC)
- No way, most "criticism of living person X" have been deleted for NPOV and BLP reasons. (t · c) buidhe 07:54, 19 September 2020 (UTC)
Can Ibn Warraq be used as a reliable source on Islam?
Ibn Warraq is a notable critic of Islam, and his use at criticism of Islam is probably appropriate. Can he also be used as a WP:RS at other Islam related articles? There is a discussion at Talk:Superstitions_in_Muslim_societies#RfC_Whether_to_allow_Ibn_Warraq_and_Swami_Vivekananda_opinions_in_the_article? on whether he is a reliable source for the claim that the "rituals of the Pilgrimage to Mecca" constitute superstition. The source is Warraq's Why I Am Not a Muslim published by Prometheus Books. Warraq is controversial (Ibn_Warraq#Reception) and not an academic.VR talk 11:30, 16 September 2020 (UTC)
- He is reliable for his own views. Including his views in "Criticism of Islam" would only be appropriate if covered by a third party RS. Emir of Wikipedia (talk) 16:13, 16 September 2020 (UTC)
- He is reliable for his views, where they are noteworthy/due (e.g. his criticism of Islam), but he is not reliable as a source on Islam, including on superstition in Islam, and his views should be considered FRINGE. BobFromBrockley (talk) 09:20, 18 September 2020 (UTC)
- My view is that he should not be used as a neutral source on Islamic belief, and I can't really see a circumstance where his attributed views would even fit within the scope of Islamic articles. Possible exceptions might be criticism sections related to specific Islamic practices such as circumcision, halal meat or the niqab. But even then, I would proceed with a great deal of caution. --Boynamedsue (talk) 16:45, 19 September 2020 (UTC)