Wikipedia talk:Reliable sources/Archive 48

Archive 45

Archive 46

Archive 47

Archive 48

Archive 49

Archive 50

→

Archive 55

Unexpectedly low degree of certainty in medical examiner findings.

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

We were discussing the reliability of a medical examiners cause of death finding over on the RS noticeboard, and we all or almost all, thought that an official medical examiner would be a reliable enough expert primary source to establish the fact of a cause of death, at least in the case of an uncontroversial finding. But then when I went looking at information about typical degrees of certainty in medical examiner findings, I found a document at the National Association of Medical Examiners website thename.org, A Guide For Manner of Death Classification (under General Info, Death Certification): https://netforum.avectra.com/Public/DocumentGenerate.aspx?wbn_key=38c0f1d2-11ec-45c7-80ca-ff872d0b22bc&SITE=NAME That states:

"In general, the certifier of death completes the cause-of-death section and attests that, to the best of the certifier’s knowledge, the person stated died of the cause(s) and circumstances reported on the death certificate. It is important to remember that these “facts” only represent the certifier’s opinion and are not written in stone or legally binding."

But it gets worse. It goes on:

Because the cause and manner of death are opinions, judgment is required to formulate both for reporting on the death certificate. The degree of certainty required to classify the manner of death depends sometimes on the circumstances of the death. Although such issues will be discussed in further detail below, a general scheme of incremental “degrees of certainty” is as follows:

• Undetermined (less than 50% certainty)
• Reasonable medical or investigative probability (Greater than a 50:50 chance; more likely than not)
• Preponderance of medical/investigative evidence (For practical purposes, let’s say about 70% or greater certainty)
• Clear and convincing medical/investigative evidence (For practical purposes, let’s say 90% or greater certainty)
• Beyond any reasonable doubt (essentially 100% certainty)
• Beyond any doubt (100% certainty)

Seldom, for the purpose of manner-of-death classification, is “beyond a reasonable doubt” required as the burden of proof. In many cases, “reasonable probability” will suffice, but in other instances such as suicide, case law or prudence may require a “preponderance” of evidence—or in homicide—“clear and convincing evidence” may be required or recommended.

Notice that a finding of homicide MAY require a heightened standard of 90%. But that means that the finding MAY NOT require even 90% certainty. A medical examiner may well be using the 70% "Preponderance of medical/investigative evidence" or possibly even the 51% "Reasonable medical or investigative probability" standard. A medical examiner may well have found a higher certainty in a particular case. But until we know, it seems clear to me that we can't state the examiner's opinion of the manner or cause of death as fact in Wikipedia without establishing what degree of certainty the medical examiner is claiming. And we should warn readers that medical examiner's findings may be determined to a surprisingly low degree of certainty, if we don't know the degree of certainty.

I think we need to modify the reliable sources guidelines to warn other Wikipedia editors that medical examiner findings may be only 51% certain, and that if greater weight is to be attributed to such findings, then the medical examiner's degree of certainty level must be verified on a case by case basis. Mindbuilder (talk) 10:33, 7 January 2015 (UTC)

? People decide on what is likely true all the time, especially when, as here, they have a choice to say they cannot state what is likely true (undetermined). Same with, it is likely true it was a homicide (that is, a volitional act by a person causing death to another), and at any rate, the notes make clear that the examiner in a homicide case would have a higher degree of certainty than likely true. The notes also state that after "manner of death" (eg. homicide, accident, etc.) it is then followed by a clarifying note on circumstances (eg gunshot, hanging, etc)Alanscottwalker (talk) 15:56, 7 January 2015 (UTC)

We're not just talking about homicide here. Wikipedia editors also need to realize in non-homicide cases that the findings may be only to the "more likely than not" degree of certainty. And any following clarifying note may also be established only to a 51% degree of certainty. But even if the degree of certainty is at least 90%, I don't think even that is nearly high enough to establish something as a flatly true fact to be stated in Wikipedia's voice. Something closer to "beyond a reasonable doubt" is a more appropriate standard for establishing a fact. If there is a reasonable doubt about a fact, then that is worth alerting the reader to, and taking a neutral point of view. Especially for a derogatory fact in a BLP scenario. And also remember, these are suggested guidelines. We just don't know what particular degree of certainty an examiner is using in a particular case even in homicide. Some examiners might use a 51% standard to establish cause and manner of death in all cases, because criminal cases often become civil cases, but simply include a separate comment to warn when their findings don't reach a higher degree of certainty. If we don't know then we have to find out first. We can't just assume reliability. Mindbuilder (talk) 17:24, 7 January 2015 (UTC)

We do not assume they are wrong - they state what they have concluded ("manner of death") and we relate it. They are obviously certain enough to state what they have concluded. Alanscottwalker (talk) 17:39, 7 January 2015 (UTC)

If you do not know if something is wrong or right, you should not assume it is right, just because you do not know that it is wrong. Nor should you assume that it is wrong, just because you do not know that it is right. The wise thing to do is not to assume it is right or wrong, but to just reserve judgment, and say you just don't know, unless you have sufficient evidence one way or the other. It was the assumption, thought to be obvious to almost all of us, that medical examiners would not declare an opinion unless they were certain enough to know the facts. But now it is clear that medical examiners will issue an opinion when they are not certain, and not even close to certain. If the degree of certainty does not exceed "more likely than not", then that degree of certainty means NOT certain, merely probable. According to Merriam Websters, certain means "not having any doubt about something" "Indisputable". A conclusion that may merely be "more likely than not" is nowhere close to indisputable. Mindbuilder (talk) 18:59, 7 January 2015 (UTC)

This is not a forum. If you have an issue with the way medical examiners word their findings, take it elsewhere. AndyTheGrump (talk) 19:10, 7 January 2015 (UTC)

I'm seeking consensus for changes to the guidelines for identifying reliable sources. I thought this was the place to do that. Is there somewhere else I should go to discuss proposed changes? Or should I just be bold and make the changes?

If I was suggesting medical examiners should word their findings differently, I would go make that suggestion where medical examiners congregate, not here. I'm just saying that I and many other Wikipedia editors were mistaken to assume that the expert findings of medical examiners were reliable. We now know that those "official" findings may be based on merely a "more likely than not" level of likelihood. Editors and readers need to be warned of this surprising fact. Mindbuilder (talk) 19:38, 7 January 2015 (UTC)

No, you're not "seeking consensus"; you're trying to avoid acknowledging it. You've already heard a consensus, that medical examiner findings are reliable sources for cause of death (for example, at RS/N). You don't agree with that consensus, and you're forum-shopping relentlessly to find a way around it. You're well into I-didn't-hear-that territory. While I can only guess at the motivations behind your crusade to discredit the M.E.'s findings in the Garner case, your behavior is disruptive and tendentious and will be grounds for being blocked from editing if it continues. You are free to raise questions or concerns about source reliability, but you are not free to disregard all input you disagree with, nor to keep asking the same question in different venues hoping for a different answer, particularly when you've omitted to link to other relevant ongoing discussions. Especially, you are not free to try to change a fundamental sourcing guideline to fit your personal agenda, when you've already been told by an overwhelming number of other editors that you're off-base. MastCell ^Talk 19:51, 7 January 2015 (UTC)

I was not the one initiating the previous forum changes related to this issue. That was Dyrnych. Previously I was following his forum changes. And I'm not suggesting his forum changes were unreasonable. I only changed forum once to this forum because I thought it was a more proper place to discuss, not just the reliability of a single source for a particular article, but a change to the reliable sources guideline page with respect to medical examiner findings in a wide variety of cases, both civil and criminal.

Furthermore I think my discovery that medical examiner opinions may be based on no more than a "more likely than not" level of probability, dramatically changes the vast majority of our previous discussion. Previously even I had agreed that they could be considered reliable when they were not disputed.

But lets get things perfectly clear. Is it your opinion that it is acceptable for a medical examiner's finding that is established to a level of "more likely than not", or for which we don't know the degree of certainty, to be flatly stated as a fact in Wikipedia's voice? Mindbuilder (talk) 20:35, 7 January 2015 (UTC)

? Apart from the medical examiners conclusion, there is nothing else that states the medical examiners conclusion. It goes into the mortality and statistics books that way. - Alanscottwalker (talk) 20:53, 7 January 2015 (UTC)

@Mindbuilder: It's my opinion that you're inappropriately trying to alter a fundamental guideline here in order to win a content dispute where consensus has gone against you. That's disruptive behavior.

The answer to your "more likely than not" question is mu. Seriously, spend a few minutes actually thinking it through. We report that the shortstop Omar Vizquel committed only 3 errors in the 2000 baseball season. But whether a particular play constitutes a hit or an error is a subjective judgement—an opinion—rendered by the official scorer. The threshold is for calling a play an error is "more likely than not", or 51%. But we report it as a matter of fact, in Wikipedia's voice, that Vizquel committed 3 errors. We don't include some sort of ridiculous, contrived nonsense about Vizquel's fielding percentage being a matter of opinion. Right? MastCell ^Talk 04:47, 8 January 2015 (UTC)

Changing a policy to try to win a dispute is the lowest of the low. Hipocrite (talk) 14:08, 8 January 2015 (UTC)

I intend to get the policy changed even if I do win the dispute. And if consensus is found that the policy needs to be changed, and that wins the dispute, then the dispute should be won, and there is nothing low at all about winning a dispute by improving the quality of Wikipedia. More on MastCell's baseball errors example later. By the way Hipocrite, do you think that a source that establishes its conclusions to a "more likely than not" standard should be considered a reliable source for statements of fact? Mindbuilder (talk) 21:11, 8 January 2015 (UTC)

I think you should be topic banned from this topic for failing to hear. I'll probably suggest that in a few more cycles of you not hearing. You lost, get over it. Hipocrite (talk) 21:40, 8 January 2015 (UTC)

That you dodged my last question is noted. It is understandable though, as your position is indefensible. And you act as if you have a strong consensus. By my last count it is only 6 to 4 against me since I made it known that medical examiner findings may be established to a mere "more likely than not" standard. And that changed the debate dramatically, so the previous opinions should not be relied upon unless they re-assert them.

In the baseball errors example, anyone that cares about the exact accuracy of the error number is likely to quickly find out that the true number of errors depends on some close calls by the official scorer. They're going to know that the precise value is not a highly reliable number. And even if for some reason some people don't figure it out, it doesn't really matter much. The precise true and fair number matters to the particular player as far as negotiating his salary, but anyone negotiating the salary of a baseball player most certainly knows the official number is not precisely reliable. And they know that Wikipedia will be reporting the official number, not Wikipedia's own finding. And they can do their own video confirmation of each error if they care to. Furthermore, for most purposes, a wrong close call or two probably doesn't mean much to the over-all reputation of the player. Few fans will probably care about a slight error in the exact number of errors. The cited Omar Vizquel article doesn't even mention Vizquel's career or yearly error numbers. Also, it's just a game. If Wikipedia's misses letting people know that the official baseball statistics aren't perfect, the omission is not that big a deal to the credibility of Wikipedia. And there is nothing wrong with Wikipedia flatly stating as fact what the official stats are. MLB probably has an official process that establishes the errors, and then probably that is the organization's final declaration. A medical examiner's finding is official, but it is nowhere near final. It can be easily changed upon new evidence, or overruled by a jury.

On the other hand, with medical examiner findings, I think few people will realize that they may be determined to such a low degree of probability. And people may be unlikely to find that out even when they look into a case deeper, as we have done. Finally, life and death findings can be very important to society and the parties involved.

@MastCell - Since you answered my question in the negative, that apparently means it is NOT your opinion that it is acceptable for a medical examiner's finding that is established to a level of "more likely than not" to be flatly stated as a fact in Wikipedia's voice. So what are we disagreeing about? Are we disagreeing about the situation when we don't know the level of probability the medical examiner found? Mindbuilder (talk) 21:55, 8 January 2015 (UTC)

• This is nothing less than forum shopping and disruptive. Maybe ANI should lead the way for this editor.TMCk (talk) 00:14, 9 January 2015 (UTC)

• As the initiator of the RS noticeboard discussion and one of the parties involved in the content dispute, I'm kind of shocked that I wasn't informed that this discussion was going on here. That smacks of forum shopping to me, especially when my initial change of forum is being brought up as a defense for moving the discussion to this page; I notified every editor involved in the discussion and posted a link in the previous discussion, neither of which has happened here. Also, can someone explain why we should accept the section of document that Mindbuilder has cherry-picked as authoritative and discount this statement from the "Preface and Caveats" of the document (page 2) as irrelevant? "This book is a Guide. The recommendations contained herein are not standards and should not be used to evaluate the performance of a given certifier in a given case. Death certification and manner-of-death classification require judgment, and room must be allowed for discretion on a case by case basis" (emphasis mine).

Mindbuilder's notion that there is not a robust consensus for accepting an ME's report as a reliable source for fact (unless disputed by equally qualified experts) is laughable. Not one of the editors opposing the acceptance of an ME's report as an RS has cited Mindbuilder's rationale for doing so; their objections mostly boil down to the fact that an ME is fallible, which is true but (in my opinion) irrelevant in the absence of expert dispute. We are not qualified to assess whether they have erred or not, and stating the mere possibility of error is not anywhere near enough to raise doubt about an ME's conclusion. On the other hand, most editors—both before and after the "bombshell"—have accepted the argument that in the absence of qualified dispute, an ME's report should be accepted as a reliable primary source for a cause of death. Dyrnych (talk) 01:04, 9 January 2015 (UTC)

• Comment (reposting from RSN): I can't believe it took me this long to notice this, but the fatal (if you'll pardon the pun) flaw in Mindbuilder's reasoning is that the excerpted segment applies to manner of death, not cause of death. Mindbuilder is conflating the two (whether from carelessness, ignorance, or whatever other reason), but they're entirely different; cause is a medical determination, while manner is not. So even accepting that this cherry-picked material has any application to this matter (which I don't), the thesis is that a medical examiner may require as low as 51% certainty before deeming a death to be one of the following: suicide, homicide, accident, or natural causes. That has no application whatsoever to the degree of certainty of the determination of the cause of death, i.e., the medical circumstances that ended the life of the decedent. That is the question: whether the medical examiner's report is a reliable source for the cause of death, not the manner of death. Dyrnych (talk) 03:01, 9 January 2015 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Using otherwise reliable sources that (without directly saying so) borrow details from Wikipedia

Obviously this page already rejects using Wikipedia and its derivatives as sources, but recently on Talk:Akira Toriyama someone tried to argue in favour of the status quo because an several external sources had copied minor stylistic details ("The subject is known for this and this") from Wikipedia. I've seen people do this in requested moves and the like as well ("X-source spells his name the same way we do on its title-page because that is the more familiar spelling of this very obscure person's name solely as a result of English Wikipedia. Therefore we shouldn't change the title.")

In my experience when it comes to minor stylistic details like this, otherwise reliable sources quite frequently copy Wikipedia just because the Wikipedia status quo is currently prominent online. When it comes to romanization and order of Japanese personal names, I know the Iwate Prefectural Government does this from time to time since I, on discussion with my other non-Japanese colleague, was the one who proposed it.

Anyone else think we should include a proviso warning against this kind of potential circular sourcing when defending the status quo?

Hijiri 88 (聖やや) 09:44, 16 January 2015 (UTC)

(By the way, while there have been a few such RMs, the one that jumps to mind is at Talk:Emperor Jimmu, where the majority of piecemeal mentions of the figure's name are in books written by people outside the field, and who therefore likely copied their spelling from either Wikipedia or the other source available for free online about the topic -- a 130-year-old public domain translation of the Kojiki. Such books are obviously of minimal value in defending the Wikipedia status quo. My recent comment on WT:MOS-JA cites another almost uncontrovertible example of an essay someone no doubt considers a reliable source on economics, written by someone who obviously consulted Wikipedia while writing. Hijiri 88 (聖やや) 13:28, 16 January 2015 (UTC) )