Wikipedia talk:No original research/Primary v. secondary sources discussion/Archive 1

Discrepant Imports of "Primary" and "Secondary"
I did some editing of the two paragraphs following the description of primary and secondary sources. I do not believe I made any substantive change. My intention was to clarify the policy, largely by rearranging a couple of sentences. Slrubenstein  |  Talk 16:52, 21 August 2006 (UTC)

JA: No, the changes are non-trivial and very disagreeable, but it's probably not your fault. Back in sec. Jon Awbrey 19:40, 21 August 2006 (UTC)

JA: One of the first things that I noticed in WP was this odd definition of "primary source" that I read on one of its policy pages, with a consequential oddness in the value that it imputed to primary versus secondary sources. I think that it probably has to to with the different strokes that different folks in journalism versus scholarship attach to those two terms. Though it has caused a strange undercurrent in many discussuions of sources, it hasn't come up in a really flagrant way till now, so I let sleeping dogs lie, for once. But Cerberus has waked, as it were, and it looks like it can't be avoided any further. Jon Awbrey 19:52, 21 August 2006 (UTC)


 * On Wikipedia, a "primary" source is typically used in the sense of "a source which creates new 'facts' never before seen". An eye-witness report of a crime is primary, the documentation of the sale of your house is primary, a transcript of a court trial is primary. Wjhonson 19:55, 21 August 2006 (UTC)


 * However part of your proposed re-wording would state that primary sources must not be used, and that's incorrect. You state "Original research that draws on primary sources is also not allowed." You can simplify this to "Original research is not allowed."  But that has nothing to do with *this* page.  I think you are confusing research with original research.  Wikipedia makes a distinction between "collecting sources together to form a coherent article" and "creating new facts." Wjhonson 19:55, 21 August 2006 (UTC)


 * The version of the policy as of 12:54, 21 August 2006 first states "Original research that draws on primary sources is also not allowed.". It then contradicts itself by saying "However, in some cases, where an article (1) makes descriptive claims the accuracy of which is easily verifiable by any reasonable adult without specialist knowledge, and (2) makes no analytic, synthetic, interpretive, or evaluative claims, a Wikipedia article may be based entirely or primarily on primary sources. . . ." I object to the first statement not only because it makes the policy contradict itself, but because the phrase "draws on primary sources" does not specify to what extent an article may draw on primary sources. It could be interpreted to mean articles should not use primary sources at all.


 * One could argue the alteration is not as bad as I make it out to be, but it is clearly worse than what was there before, and this degradation was introduced without first being discussed on the talk page. --Gerry Ashton 20:07, 21 August 2006 (UTC)

JA: This is a serious problem, so let's all take our time and fix it right. One of the things that WP:NOR means is that WP cannot make up it's own private language for the meanings of words already in common use. The terms primary source and secondary source are terms in common use, even if they are commonly used with many different imports, and so Job 1 for us is finding out and sorting out what those meanings and imports already are. Jon Awbrey 20:12, 21 August 2006 (UTC)

JA: The edit by Slrubenstein was ambiguous, as it could be read as extending the definition of OR rather than drawing on a prior definition, so that will not work. Jon Awbrey 20:21, 21 August 2006 (UTC)

This editing worries me a bit, as you and I were still discussing basic wiki guidelines as it is. Kim Bruning 20:34, 21 August 2006 (UTC)

Original research is the creation of new "facts" not citable to any other source. Citing a primary source is not original research (nor is citing a secondary source). Frequently editors will shout "original research" when what they mean is "the person did *research*". Research is not forbidden, it is the creation of new facts that cannot be cited that is forbidden. Wjhonson 20:41, 21 August 2006 (UTC)

JA: Check. Jon Awbrey 20:44, 21 August 2006 (UTC)

Looking at the Chicago and APA style manuals, and thinking of my own experience, I can think of two definitions for primary source:
 * 1) A source which reports original research
 * 2) A source which contains quotes or summaries from other sources, but these quotes or summaries have never been used in a tertiary source.

Similarly, I can think of two definitions for secondary source:
 * 1) A source, such as a review article or book review, which exists mainly to summarize, synthesize, or review other sources
 * 2) A source which quotes or summarizes a primary source, and which in turn is quoted or summarized by a tertiary source.

These definitions overlap, but there are areas in which they differ. For example, an electronics engineer who reads Donath's article in IEEE Transactions on Circuits and Systems, V. CAS-26 No. 4 will see that it predominantly reports original research, but it does briefly summarize the work of Rent. If I quote Donath's summary of Rent, instead of reading and citing Rent's work directly, I am using Donath as a secondary source. I think the NOR policy uses the meanings that I've labeled 1, but I think all the meanings are in general use and scholars routinely have to distinguish the meaning from context. --Gerry Ashton 21:20, 21 August 2006 (UTC)


 * Additionally I would point out that while Donath's article on his own research is a primary source for that research, his *same* article on his summary of Rent is a *secondary* source for that portion of the article. Wjhonson 21:38, 21 August 2006 (UTC)


 * While that article may be a *secondary* source for the work of Rent, it is a *primary* source on how Donath would summarize Rent. - O^O 22:15, 21 August 2006 (UTC)


 * For the *general method* he uses perhaps, but that would go on *his* page not on Rent's page. "Donath frequently uses the pi-beta-epsilon method of summarizing..." While on Rent's page you would simply quote it and cite it.  Now, in the case where there are conflicting secondary sources, I always go for inclusion versus silence.  "Fred Flintstone summarizes Dino's work as annoying, while Wilma summarizes it as playful and cute."  We have room for both secondary sources. Wjhonson 03:09, 22 August 2006 (UTC)

JA: I see some ideas above that make sense to me, or at least sound familiar to me, but I think overall that it's making things too complicated to drag in mention of tertiary sources. Normally, I think of Classical examples, but that gets us into all the vagaries of trans*lations and whether Socrates said what Plato said he said and so on, so let's FF to semi-modern times. What William James wrote is the definitive primary source for William James' philosophy — the buck stops there as far as that goes. What he wrote about the philosophy of C.S. Peirce is a secondary source for that. And what Bertrand Russell wrote about the philosophy of William James is a (not always very sympathetic or understanding) secondary source for that. I think that is plain and simple enough, and gives the gist of the basic idea. This means that primary sources have a slightly different value in that brand of scholarship than they might have in some brands of journalism, say. They are to be valued as the rock-bottom grounds of grounded research when it comes to questions about what a given writer's writings actually said. Jon Awbrey 02:40, 22 August 2006 (UTC)

Slrubenstein' edit is a good one, and as for the attempts above to redefine the meaning of "primary source", please spare us.... ≈ jossi ≈ t &bull; @ 14:26, 22 August 2006 (UTC)


 * There is about 205kB of text above. Could you clarify who is attempting to redefine "primary source"? - O^O 17:29, 22 August 2006 (UTC)

JA: No, the intention was good, but it took me 3 readings to guess what it was — we can't afford that kind of ambiguous construction. Jon Awbrey 14:42, 22 August 2006 (UTC)


 * Wikipedia defines primary sources very clearly: "present information or data, such as archeological artifacts; film, video or photographs (but see below); historical documents such as a diary, census, transcript of a public hearing, trial, or interview; tabulated results of surveys or questionnaires, records of laboratory assays or observations; records of field observations." This has been a stable definition for a very long time and we should not try to rewrite it.  My edit in no way altered or modificed this definition.  It is unfair to suggest that my edit in any way changed what we mean by primary source. Moreover, my edit did not say that use of primary sources is never allowed.  ll I did was clarify the two paragraphs so as to eliminate any possible inconsistent reading of the policy.  The policy allows the use of primary sources under very very spoecific circumstances.  My edit in no way changed the circumstances under which primary sources can be used.  I did not extend the definition of OR at all.  Slrubenstein   |  Talk 18:03, 22 August 2006 (UTC)
 * That was my reading as well. ≈ jossi ≈ t &bull; @ 18:11, 22 August 2006 (UTC)


 * I'd like to see this proposed sentence deleted or reworded: "Original research that draws on primary sources is also not allowed." I'd prefer to see it struck altogether, as it could imply to some that there is some class of original research that is allowed. - O^O 18:15, 22 August 2006 (UTC)

I have reverted a change by User:jayjg mostly because there is no consensus for the change. A secondary reason is it claims that "Original research that draws on primary sources is generally discouraged" (my emphasis). Draws on is too vague. A perfectly acceptable article that relies mostly on secondary sources might cite some primary sources to good effect. Valid reasons for using primary sources could be to discredit secondary sources that misquote the primary, or to fill in details that are not mentioned in a secondary source but which are of interest in the Wikipedia article. If the phrase had been "Original research that predominantly draws on primary sources is generally discouraged", and if the phrase had been agreed to on the talk page, I would have gone along with it.--Gerry Ashton


 * "Valid reasons for using primary sources could be to discredit secondary sources that misquote the primary, or to fill in details that are not mentioned in a secondary source but which are of interest in the Wikipedia article." This use of primary sources serves to forward an argument and thus is in violation of NOR.  This is precisely why I made the edit I did - the policy needs to be crystal clear.  But I have not changed the policy.  Using primary sources to promote an editor´s interpretation or argument is just a violation of polcy.  Stop trying to destoy the policy. Slrubenstein   |  Talk 04:02, 23 August 2006 (UTC)


 * If we change the word "discredit" to "properly quote" then the entire quoted phrase is correct. We can quote primary sources, and actually SHOULD do so in cases where secondary sources misquote them.  And we can certainly, absolutely, fill-in details from primary sources.  That is one of the main reasons to use primary sources in the first place.  We can quote what Baha U'llah said about his religion, in the same paragraph where a critic is saying something else.  That is perfectly acceptable. Wjhonson 04:31, 23 August 2006 (UTC)


 * I used the word "discredit" only to mean correcting narrow factual errors that crept in between the primary and secondary source. Overturning a position put forward by a secondary source would require another secondary source, possibly supported by primary sources.


 * As for filling in details, I was thinking of a case where a secondary source has already established the importance and general meaning of a primary source, but omitted some detail that is useful for the article. There is no reason it couldn't be included.


 * Such primary references could be handy when an editor inserts a garbled, unsourced "fact" into an article. One such case occurred in the Metrication article. The article correctly stated that Jefferson proposed a decimal measurement system; an editor changed it to say Jefferson had proposed the metric system. Unfortunately neither version had a citation, so I corrected it and cited Jefferson's report to Congress. My change did not change the general significance of the fact in the article, namely, the United States considered decimal measurement early in its history; I just corrected a fact.


 * Of course, any such details should not undermine the position(s) put forth by the secondary sources that the article relies on; using details to advance a position different from the positions of the secondary sources is indeed original research. --Gerry Ashton 15:55, 23 August 2006 (UTC)

literary works
I deleted a paragraph about literary works and interpretations of literary works - not becuase it was a bad idea, just unnecessarily wordy. I added literary works to our list of primary sources. The effect is the same, just mor economical Slrubenstein  |  Talk 18:12, 22 August 2006 (UTC)

I deleted this:
 * The relative value of primary and secondary sources with respect to WP:Verifiability depends on the primary topic of the article. For example, when discussing an author's work, what the author actually wrote is a primary source that is definitive on that subject, and it is treated as the most reliable source, over and above what any secondary source may say about it.  In contrast, when discussing primary source reports of real events, common sense tells us that a comparatively disinterested secondary source, especially one that collates and verifies many different primary accounts, may be able to be more neutral in its reporting.

because it is unnecessary, argumentative, and violates NPOV. It is better that this policy not address the issue at all, it is irrelevant to the policy because it does not change the definition of primary or secondary source and has no bearing on original research. This paragraph is about weighing the value of a particular secondary source, which is a separate matter. Moreover, the paragraph makes a claim about the weight of a particular secondary source and this is simply the creation of a new policy which is utterly unwaranted. We shouldn´t make new policy without discussing it and I think the proposed new policy here is a bad one. In fact, the question as to whether an author´s interpretation of his or her own work is authoritative is something that has been hotly debated and there are scholars on both sides of the argument. How should one consider an essay by Phillip Roth on the work of Phillip Roth, or an interview with Julien Barnes on work by Julien Barnes (or, what to do with Pale Fire)? This should be handled on a case by case basis, and this policy should not argue for a bias one way or the other. Some scholars surely will claim that Roth´s own commentary or interpretation has precedence. NPOV requires presenting their views in the article. But other authors will argue the opposite and their views too must be represented, to comply with NPOV. NPOV is inviolate. The POV in the paragraph I deleted cannot be made into policy. Slrubenstein  |  Talk 18:20, 22 August 2006 (UTC)

JA: This policy page is intended to state the WP:Policy of WP:NOR, to help editors understand its proper application, and to explain the reasons behind the policy. I always like it better when Bossy People give me a Reason Why I ought to do a particular thing. All I did was state some of the senses of the terms in question that are commonly used in scholarship, amd I see nothing argumentative or controversial about that. I simply explained standards of practice already in use, from which WP derives its authority and justification for its own policies. I did not introduce the additional issues that you raise. The point is that the publication of a work is a historical event in its own right, and definitive in itself. Obviously, no one claims that what an author says about his or her own work is anything more than yet another secondary opinion about it. Jon Awbrey 18:40, 22 August 2006 (UTC)

JA: Again let me emphasize: The paragraph does not assert anything like "an author´s interpretation of his or her own work is authoritative". That would be just plain silly. Jon Awbrey 18:44, 22 August 2006 (UTC)


 * Jon, you wrote: " ... when discussing an author's work, what the author actually wrote is a primary source that is definitive on that subject, and it is treated as the most reliable source, over and above what any secondary source may say about it ..."


 * The implication of this is, for example, that material written by a particular neo-Nazi author (whose article I edit a fair bit) must be treated as authoritative, regardless of any criticism or contradiction of it by the London Times.
 * Rather than changing the policy, could you say which part of that section you feel is problematic and needs to be changed? SlimVirgin (talk) 18:55, 22 August 2006 (UTC)

Proposed Explanation of Primary & Secondary
JA: Here is the primary source for my last revision of the paragraphs in question, which got mugged by an edit conflict: In discussing the works of an individual writer, for instance, in compiling a bibliography of writings by and about an author, primary sources are works actually written by that author, while secondary sources are works written about that author's life and work by others. The relative value of primary and secondary sources with respect to verfiability depends on the primary topic of the article. For example, when discussing an author's work, what the author actually wrote in that work is a primary source that is definitive as to its content, and it is treated as the most reliable source, over and above what any secondary source may say about it. In contrast, when discussing primary source reports of real events, common sense tells us that a comparatively disinterested secondary source, especially one that collates and verifies many different primary accounts, may be capable of more neutrality in its reporting. JA: I think that a careful reading will show that I have clearly not said any of those silly things that some people are saying I said. So the only question is how to prevent most any reader from thinking that I said that. Jon Awbrey 19:02, 22 August 2006 (UTC)
 * Primary sources present information or data, such as archeological artifacts; film, video or photographs (but see below); historical documents such as a diary, census, transcript of a public hearing, trial, or interview; tabulated results of surveys or questionnaires, records of laboratory assays or observations; records of field observations.
 * Secondary sources present a generalization, analysis, synthesis, interpretation, or evaluation of information or data from other sources.


 * You are losing me here Jon, and I am sure you are losing others... ≈ jossi ≈ t &bull; @ 19:03, 22 August 2006 (UTC)


 * Yes, I'm lost too. Jon, this sentence of yours is clearly saying the primary source is authoritative: " ... when discussing an author's work, what the author actually wrote in that work is a primary source that is definitive as to its content, and it is treated as the most reliable source, over and above what any secondary source may say about it ..." And that is problematic. SlimVirgin (talk) 19:17, 22 August 2006 (UTC)

JA: Look, it's really a lot simpler than some folks are making it. Why do we cite published sources? So the reader can go and check what is actually written in that text. That's the basic principle here. The text is the definitive source for answering questions about what's in the text. That's all. Nothing else. The writer can preface the work, and many do, by writing, "This is the greatest work that human civilization has ever produced". The text is the last word for the fact that the author wrote that. It is not the last word on whether that statement is true, and nobody would even think of saying that. Jon Awbrey 19:20, 22 August 2006 (UTC)


 * Can you say what is wrong with the section as it stands? SlimVirgin (talk) 19:21, 22 August 2006 (UTC)

JA: Trust me, I don't come to any of these Policy pages until somebody drags me here by the whiskers of some tag or template or other. There is a real misunderstanding in the usership about the value of primary sources, and it seems to be this issue about how the value shifts from context to context. I have many times had folks object to the fact that I use a quotation from X to support a statement about the philosophy of X. Nobody in real world scholarship would make that objection, so there must be an extra source of confusion in WP. Some of it seems to be in the current definition of primary source, that makes it seem like all primary sources are suspect almost all of the time. There's even a template somwhere that tags articles for citing too many primary sources, and I have seen that get misapplied, apparently due to this misunderstanding. Jon Awbrey 19:38, 22 August 2006 (UTC)

A series of edits and reverts just occurred which ended up with the phrase "Original research that draws on primary sources is generally discouraged". This is wrong. The use of primary sources is encouraged. It is only when the article predominantly uses primary sources, or synthesizes them in ways that are not obvious to a reasonable adult, that it becomes original research and thus forbidden. Maybe that's what "draws from" means to whoever wrote this, but to me "draws from" means uses one or more words from a primary source. --Gerry Ashton 19:50, 22 August 2006 (UTC)


 * I picked up on that also. I think that addition has been reverted away, but it appears to have briefly gotten mixed up with the "original author" topic being discussed above - O^O 19:15, 22 August 2006 (UTC)
 * Which version exactly did you revert to? SlimVirgin (talk) 20:08, 22 August 2006 (UTC)
 * "18:13, 22 August 2006 Slrubenstein" - O^O 20:12, 22 August 2006 (UTC)
 * P.S. I have no opinion on the recent Duncharris change (regarding VAIN). But the  Jayjg change would have reinserted the 18:25 Slrubenstein text.  Reverting back to Duncharris, which is based on the 18:13 Slruberstain text . - O^O 21:59, 22 August 2006 (UTC)
 * 18:25 was a very clear time in the history of this policy; clearer than at any time in the past. Jayjg (talk) 22:40, 22 August 2006 (UTC)
 * Indeed, that phrase is wrong: often secondary sources misquote and distort primary sources. Thus, alhough for synthesis one should use secondary sources where primary sources are lacking of that, for accuracy it's essential to use primary sources as much as possible. This has been discussed before.
 * PS: I propose to entirely remove mention of "primary" and "secondary" sources, as it's rather artificial and only complicates the description of the NOR policy. It's all rather obvious without such hair splitting. Harald88 22:13, 22 August 2006 (UTC)


 * Harald, this has indeed been discussed many times, but you're getting it wrong. It is not essential to use primary sources as much as possible. Quite the reverse. Using primary source material correctly requires training for the most part. Wikipedia editors should stick to secondary sources as much as possible, and use primary sources with great care. SlimVirgin (talk) 00:50, 23 August 2006 (UTC)


 * Removing the distinction between primary and secondary sources will create total chaos. Not an acceptable proposition, IMO. ≈ jossi ≈ t &bull; @ 22:19, 22 August 2006 (UTC)


 * Indeed. A "primary" characteristic of Original Research is that it is almost always built on novel interpretations of primary sources. Jayjg (talk) 22:40, 22 August 2006 (UTC)


 * Original research may be a novel interpretation of primary sources, but not always. It may be a first report of the result of an experiment, or a record of an event. It need not include much interpretation; it might consist of rather raw data that is presented in the hopes that someone else can interpret it. --Gerry Ashton 23:13, 22 August 2006 (UTC)
 * That is why we discourage its use. SlimVirgin (talk) 00:51, 23 August 2006 (UTC)

Exactly. Lacking other forms of quality control, we DO discoursage the use of primary sources. We do not prohibit the use, and one editor called my attention to this and I corrected my edit. We do allow for their use under specific conditions. Under other conditions we discourage their use. Whenever any editor turns to primary sources to "corect" a secondary source, we are almost certainly witnessing a violation of NOR. Apparently there are some editors here who want to weaken the policy. My edits are solely to clairfy the policy as it exists and defend it against those who want to weaken it. Slrubenstein  |  Talk 04:10, 23 August 2006 (UTC)


 * That depends on exactly what you mean. If a secondary source states that a Rush Limbaugh in his blog said "George Bush is fat" and you can cite the actual blog saying "George Bush is flat" then that is definitely *not* a violation of OR.  Rather that is "source-based research".  OR involves *you* creating *new* facts, not citing those facts in primary sources. Wjhonson 04:22, 23 August 2006 (UTC)


 * User:Slrubenstein writes that "whenever any editor turns to primary sources to 'corect' a secondary source, we are almost certainly witnessing a violation of NOR." There is no original research if the correction consists of correcting an obvious misquote. This may not be a big deal in the humanities, but in math and technical topics, a single character can be critical. I do not take it for granted that the vast majority of corrections to secondary sources are something more subjective than clear factual corrections. Ideally any use of a primary source to correct a secondary source would be used in conjunction with one or more secondary sources that concur with the correction. --Gerry Ashton 04:45, 23 August 2006 (UTC)


 * I agree with everything Gerry said, except where he states that "Ideally any use of a primary source... would be used in conjuction with...secondary sources...." I would amend that to add "but that is not necessary" to make clear that the primary/secondary see-saw is evenly balanced.Wjhonson 04:50, 23 August 2006 (UTC)

JA: I don't know how to say this strongly enough. Any policy that can be interpreted as discouraging editors from citing X in support of statements about X's beliefs or philosophy is just plain out of touch with the realities of scholarship in the real world, and the definitions of primary source and secondary source as they are commonly understood there. Jon Awbrey 04:28, 23 August 2006 (UTC)


 * I strongly object to this "Original research that draws on primary sources is generally discouraged, in favor of research based on secondary sources" proposal that has popped up today, and at least three other editors appear to have caught this and are object to it as well. IMHO, this is a significant deprecation of the use of primary sources from the accepted standard here.  I don't see this as a "clarification", but as a wholesale attempt to discourage the use of primary sources. - O^O 05:04, 23 August 2006 (UTC)


 * We discourage primary sources, this has ben our NOR policy for ages. This does not mean we do not allow them.  It means we encourage primary sources.  This too has been our NOR policy for ages.Slrubenstein   |  Talk 14:29, 23 August 2006 (UTC)


 * Wjhonson writes, "OR involves *you* creating *new* facts, not citing those facts in primary sources." This is not te.  OR ALSO consists o using primary sources to forward new interpretations, arguments, explanations, analyses, and syntheses.  Again, this is long-established policy.Slrubenstein   |  Talk 14:29, 23 August 2006 (UTC)


 * AS to Jon Awbrey, he wrote about the importance "citing X in support of statements about X's beliefs or philosophy." There is no need for any clarification of the policy here.  X´s statements about X´s own beliefs or philosophy may be a primary source.  In such cases, as the policy states (both in the earlier verion and in my revision) tht such a primary source can be used if it is not used to promote any argument, analysis, interpretation.  X´s statements may also be a secondary source (e.g. Phil Roth writes about Philip Roth) in which case we draw on it like any other secondry source.  All of this is IN the policy - it was in the policy last week, and after my edit.  But Awbrey also suggsted that such statements count for more than statements by others (e.g. other secondary sources).  This is a flot out violation of NPOV.Slrubenstein   |  Talk 14:29, 23 August 2006 (UTC)


 * I interpreted Jon Awbrey's statement to only mean that if there were a question about whether a secondary source had correctly quoted or paraphrased an author, the primary source, that is the author's published work, would be more reliable than the secondary source. I didn't take it to mean that an author's commentary on his own work has a privileged position compared to other critics. --Gerry Ashton 16:34, 23 August 2006 (UTC)

New synthesis/is this OR?
I have been telling another editor that a contribution they made is probably OR. However, I am questioning whether I am being too strict, so I am here to try to get some clarity on the policy. Here is the scenerio. This one source says that 6 people have died using mifepristone. This other source says that over 460,000 doses have been shipped for Mifeprex (brand name of mifepristone), and most likely the vast majority of the doses equals one administration of the drug. Now, in the wikipedia article, we give the figure "6 in 460,000" as the mortality rate for the drug. However, I cannot find this figure listed in any source online. Most sources online simply say "less than 1 in 100,000". I have said that wikipedia cannot be the first place to publish the figure "6 in 460,000", but the counter argument is that its simply math. So can we say the mortality rate of a drug is X, if there isn't a single source backing X up, but instead a synthesis of multiple sources? The conflict is being discussed at Talk:Mifepristone. Thanks for your time.--Andrew c 00:24, 23 August 2006 (UTC)


 * What kind of encyclopedia is it that questions its right to put two related facts from reputable sources together, and do the relevant math? One that is still finding its way, one is inclined to conclude. --londheart 00:36, 23 August 2006 (UTC)


 * No wikipedia editor can prove that he or she is an expert on drug distribution. Even if the editor is an expert, there is no mechanism on Wikipedia to establish the editor's identity. So, can a non-expert put the number 6 people together with 460,000 doses shipped? I don't think so. Maybe the number does not include free samples distributed by physicians, or used during clinical trials. Maybe it does include doses that expired sitting on pharmacy shells and had to be thrown away. I would want to see a reliable source that puts the two numbers together. --Gerry Ashton 00:43, 23 August 2006 (UTC)


 * Furthermore, it ignores the fact that the drug has been used in Europe for over 15 years and probably used by over 600,000 people.--Andrew c 00:46, 23 August 2006 (UTC)


 * I agree that a source would need to be found who makes that analysis. SlimVirgin (talk) 00:48, 23 August 2006 (UTC)


 * So long as the parameters for the discussion keep being changed here, we have to. --londheart 00:55, 23 August 2006 (UTC)

he has found the numbers online, per the reliable sources i cited. 1) the FDA=6 fatalities. 2) the manufacturer of the drug says 460,000 units. (the New England Journal of Medicine notes that units shipped may not equal procedures, and therefore the number may be lower, but does not quote a number--this would mean a higher fatality rate) the number is listed in the article as "6 fatalaties per 460,000, may be higher/about thirteen times the rate of surgical abortion" in the absence of reliable data about how many it does mean, the number with a source is cited.

his quote of "600,000 in europe" fails to note any fatality statistics, as stats are not available.

the 1 in 100,000 figure he gets predates wider use of the drug/is quite old. the new statistic would be about 1.34. this number is probably actually closer to 2. in the absence of any current, more accurate numbers, 6 in 460,000 is the best data we have.

it may also help you to know that he originally posted (or was happy with/completely uncritical of) the estimate 1 in 100,000 (medical abortion) compared with .5 in 100,000 (surgical abortion). when i corrected the number for surgical abortion to .1 in 100,000/one in a million, and the numbers for medical abortion to 6 in 460,000, he at first questioned 6/460,000 and thought it should be expressed as a measure of something-in-100,000. but he objected to the conversion of 4/4600,000 to 1.34 in 100,000 as "original research." we are both pro-choice (but my edits tend towards pointing out the risks of the drug; his to prochoice vs. prolife arguments.) he would like to revert to 1 in 100,000 even though it is outdated; i want to use the most current numbers available (and have not made an argument for upping the fatality percentage by statistical analysis of the likely lower number of procedures than the drug company claims "units shipped.") Cindery 01:48, 23 August 2006 (UTC)


 * I'm not trying to push a POV, and I feel you miscaracterized my editing habits in general, and my concern over this issue in specific (but we can try and work that out on user talk if necessary). The main point: Find a source that says 6 in 460,000, or it fails WP:V and WP:OR. The only content I desire to be in the article, is content that meets wikipedia policy. I mentioned that the only sources I could find say "less that 1 in 100,000". Whether you feel this information is accurate or not doesn't matter, if you don't have another source that corrects it. It's not about the numbers. I'd be happy with 1.35 in 100,000 or whatever, if it is backed up with a citation.--Andrew c 01:59, 23 August 2006 (UTC)

was attempt at impartial summary re we both have some pov--i don't exactly feel my edits "emphasize the risks" of the drug--i would say they strive for accuracy (and that that is not appreciated by people who feel that it weakens the prochoice argument against the prolifers...) nevertheless, some context seemed helpful. i feel you may not be entirely self aware re your pov (plase read back over the "straw man" sections i pointed out that you particularly resisted any editing adjustments to, the ones where issues were characterized as being only "prolife" and then "refuted" by prochoicers.) in any case, what matters is: are these numbers accurately cited from reliable sources (per wiki def of reliable sources)? does converting them via fractions make them innaccurate? note also, i had/have no objection to leaving them at 6/460,000--i'm still not sure why you think they have to be converted (and therefore magically transmogrified into "original research" to which you think you can reasonably object?) Cindery 02:31, 23 August 2006 (UTC)


 * Converting the number isn't the OR research part. Please see Gerry Ashton's reply above. Coming up with our own mortality rate that isn't listed in any source is the OR part. I don't care about pro-choice vs. pro-life. This is about policy, please try to stay on topic.--Andrew c 02:42, 23 August 2006 (UTC)
 * Somedays I am so glad we have this policy in place. Today is one of these days, after reading the discussions above. Provide a source for that 6 in 460,000 stat, or mercilessly delete it. ≈ jossi ≈ t &bull; @ 04:05, 23 August 2006 (UTC)

Sorry it took me so long to jump in, I was busy doing some long division that I'm going to submit to the New England Journal of Medicine as original research. But I just checked out the article and that 6 in 460,000 is well sourced. Mumblio 04:23, 23 August 2006 (UTC)

Biographical talk pages
This policy currently states "Like most Wikipedia policies, No original research applies to articles, not to talk pages or project pages, although it is regarded as poor taste to discuss personal theories on talk pages."(as of 01:52, 24 August 2005) I believe this particular point in policy should be amended to state that original research shouldn't be in biographical talk pages at all per WP:BLP: "''Editors should remove any negative material that is either unsourced or relies upon sources that do not meet standards specified in Wikipedia:Reliable sources from any page, including those concerning living persons and related talk pages, without discussion; this is also listed as an exception to the three-revert rule. This principle also applies to biographical material about living persons found anywhere in Wikipedia. Administrators may enforce the removal of unsourced material with page protection and blocks, even if they have been editing the article themselves. Editors who re-insert the material may be warned and blocked.''"(as of 21:15, 5 June 2006) (Emphasis added) --HResearcher 11:45, 23 August 2006 (UTC)
 * I'd prefer if this rule was only applied strictly to biographical articles. On other pages, we can be a lot more eventualist about it, imho. Kim Bruning 12:49, 23 August 2006 (UTC)
 * Strongly object to "biographical". If this rule is applied it should be strictly limited to  . Wjhonson 16:47, 23 August 2006 (UTC)
 * I agree with that. It should be made clear that this is in regards to living people. --HResearcher 02:55, 24 August 2006 (UTC)

No Safety In Secondary Sources
JA: The idea that secondary sources somehow save you from the wages of original sin is just plain absurd, and we should not be encouraging editors in that delusion. The rule against adding unverifiable claims or interpretations to cited sources applies across the board, to primary, secondary, tertiary, and so on down the line. There is nothing about sane editorical policy that makes secondary sources preferable. Jon Awbrey 16:32, 23 August 2006 (UTC)


 * The problem is simply that certain things cannot be obtained from primary sources except through a level of analysis that exceeds our OR standard. Primary sources are very valuable for direct facts, and for verification of the accuracy of secondary sources, but there are many things they cannot do. Matthew Brown (Morven) (T:C) 17:55, 23 August 2006 (UTC)

JA: Please refer to my previous remarks about the different uses of the terms primary source and secondary source. Both uses are perfectly sensible and standard, but they apply in different contexts. One is the "journalistic or judicial" (JoJ) sense, "eyewiteness" vs. "secondary reports". The other is the "bibliographic or scholarly" (BoS) sense, "Works by X" vs. "Works about X", and you can't write a Bib without it. Most folks here are talking about the JoJ sense, and there is makes sense to rely on multiple independent sources. I am worried about the BoS sense, and all those article writers in WP who copy right out of a single secondary source because it's too much trouble to do actual research, and who will actually scream bludey muder about it if you dare to quote the locus classicus. That should not be encouraged. That is poor and substandard scholarship. That is usually very POV. That is many other bad things. Jon Awbrey 18:10, 23 August 2006 (UTC)


 * I don't think anyone (sensible, at any rate) disagrees that there are times and places for primary sources. Thus, I'm not sure why the issue. Matthew Brown (Morven) (T:C) 20:20, 23 August 2006 (UTC)

JA: I for one, and several others for more, have been saying exactly why this is an issue. We have also described in general terms the sorts of concrete cases that actually arise in WP practice where people have read this policy as discouraging the use of quotations from primary sources in the bibliographic sense. We would not have even dreamed this possible if it did not happen and keep on happening. Jon Awbrey 20:48, 23 August 2006 (UTC)

JA: The is the very thing that first brought me WP. My routine web searches on particular topics and writers keep turning up howlers that no reputable scholars or reliable sources in those areas would ever dream of putting in print. I eventually discovered that most of the more flagrant ones came from a few popular works with no credibility on those subjects. But try and correct a pet popularization of the subject by actually quoting the works that were supposedly being summarized, or even try and suggest the reading of that all imporant second book? — well, forget about it! WP discourages that sort of thing! Jon Awbrey 21:54, 23 August 2006 (UTC)

Anautocitation (Non (Self Reference)) Is Not Justly Enforceable in Wikipedia
JA: Some journals allow their authors to self-cite and self-reference, some journals do not. Either way, they can allow or ban this practice with justice, because their editors know the real names of their authors. Obviously, no such policies can be enforced with any pretense of equal application in Wikipedia. Consequently, all policies whose validity depends on knowing the true identities of WP contributors are null and void and must be purged from the policy and guideline pages. I'll do what I can, but life is short and the art is long. Jon Awbrey 16:48, 23 August 2006 (UTC)


 * We can, if we wish, state our desire that people not self-cite, even if this is unenforceable - I don't see why this necessitates not stating it. Matthew Brown (Morven) (T:C) 17:47, 23 August 2006 (UTC)


 * Because,e.g., many "expert editors" would be put off by such a statement without it having any good effect at all. Unless unnecessarily irritating those "skilled in the art" is your purpose, naturally.  There is a principle taught military commanders-- "Don't give an order you know will not be obeyed".Pproctor 22:30, 24 August 2006 (UTC)


 * I see general ways to state the policy, acknowledging that most readers and editors have no way to know when a self-citation occurs. One way is to just say that all editors are encouraged to remove material that is not supported by adequate citations, and editors may also delete material that overemphasizes one point of view, period. Leave it to the expert to deduce that if he uses his expert unpublished knowledge, it will be deleted if anyone takes notice of it. Similarly, the expert will deduce that if he edits an article giving an overview of electric generators to include three paragraphs detailing the contributions he made to better generator bearings, the material will probably be deleted if anyone notices.


 * The alternative is call attention to the expert the likely consequences of adding unsourced or unbalanced material, just in case the expert doesn't deduce the consequences without prompting. If the advise is ignored, the edit summaries should contain words like "deleted unsourced material", not words like "deleted material from expert."


 * I don't support words like "We wish that expert editors would not draw on their personal and direct knowledge…" becase that strikes me as a weak statement. I'd rather see a statement along the lines of "We remind expert editors that information based only on their personal and direct knowledge will appear to readers and other editors as unsupported information, and will be subject to deletion."


 * Finally, I don't care for the phrase "third-party publications." Just who exactly are the first, second and third parties? The phrase could be interpreted to mean that works written by the expert shouldn't be used, even if published by a ruputable publisher; that isn't what it means, but it could be interpreted that way. --Gerry Ashton 19:07, 23 August 2006 (UTC)

I agree that there should be appropriate discussion here, but WP:NOR and WP:VAIN should not be contradictory, as they were before. It is unfortunate that there are trolls and cranks, and such trolls do have their own sources that they like to cite. In some cases, even if they are good peer-reviewed work they may choose to overstate their importance and violate WP:NPOV. I also disliked the original formulation of "expert editor" as this is elitism - surely there is a gradation of expertise rather than overt categorisation into expert v non-expert. &mdash; Dunc|&#9786; 19:24, 23 August 2006 (UTC)


 * First, "overstating" is a completely subjective matter.  Even someone "skilled in the art" (a non-elitist patent-office definition) would have difficulty discerning it in most technical fields.    How can someone who is not "skilled in the art" do so?  It is not,  like Justice Jackson's definition of "Pornography",  "I know it when I see it".


 * Likewise, this is properly a matter of NPOV and not of "vanity",  which is even more subjective,  like (e.g.) "heresy".  The proper way to correct an NPOV cite is to (e.g.) contradict it with another cite,  if you can.  The best way to do this is to lure someone else "skilled in the art" to give their input.  Your proposed rule change is not the way to accomplish this. Pproctor 20:18, 24 August 2006 (UTC)

Major policy change re: primary sources
The proposed change represents a major change in how primary and secondary sources would be treated. Without going through the text line-by-line, consider just this:

Before: However, research that consists of collecting and organizing information from existing primary and/or secondary sources is, of course, strongly encouraged.

After: Research that consists of collecting and organizing information from secondary sources is, of course, strongly encouraged.

Use of primary sources has always been strongly encouraged, and in no way have secondary sources been preferred to primary sources except that an article was rarely supposed to rely solely on primary sources. This new policy makes primary sources distinctly less preferred than secondary sources, is a major policy change, and I do not support it. - O^O 18:00, 23 August 2006 (UTC)

JA: " — Jon Awbrey 18:14, 23 August 2006 (UTC)


 * Many of the policies on WP are based on the fact that many of the editors are not trained scholars, and even the scholars have no mechanism to prove their status. So the policies should be worded not only to provide justification for deleting passages in articles, or entire articles; the policy should also be worded to educate editors who are not trained scholars. Some of the proposed changes create a tone that would imply that editors shouldn't bother reading primary sources at all (unless they feel like it for some reason other than contributing to WP). I actually wouldn't have an issue with the precise meaning of "Research that draws predominantly on primary sources is generally discouraged…" but I don't think that phrasing sets the right tone. If someone were creating an article on Shakespeare's The Merchant of Venice, I really don't think we want to imply that the editor would be better off not reading the play, even though an article that cites nothing but the play would be OR and subject to deletion. --Gerry Ashton 18:38, 23 August 2006 (UTC)


 * Actually, I'm sure it would be possible to cite nothing but the play and avoid OR, but then all the article would be would be a synopsis of the play and could not have any analysis, criticism, or much else. Matthew Brown (Morven) (T:C) 20:23, 23 August 2006 (UTC)

The current version, again cannot stand, as it denigrates primary sources too far. I wish you all would stop verting and reverting and reverting all over the wikisphere. Wjhonson 20:22, 23 August 2006 (UTC)


 * Seconded. The existence of a revert war right now over this policy's content suggests that we do not have consensus.  How about TALKING ABOUT IT? Matthew Brown (Morven) (T:C) 20:28, 23 August 2006 (UTC)


 * The status quo has been that use of primary and secondary sources are on completely equal footing, with the sole exception that a wikipedia article will not be "based entirely" on primary sources except in limited cases. Other than that narrow exception, the two types of sources are equally welcome to be used.  It has never been policy that secondary sources are "preferred" over primary sources in any manner whatsoever.  The proposal being discussed is a major and dramatic change of existing policy, not a "clarification" of what already is.  - O^O 20:43, 23 August 2006 (UTC)


 * OO, secondary sources have indeed always been preferred to primary sources, simply because use of the latter often requires analysis that we don't allow. SlimVirgin (talk) 21:53, 23 August 2006 (UTC)


 * SV, I respect your right to hold that opinion, but it does not appear to match the text of the policy. Please point out where in the policy it has ever said secondary sources are preferred to primary sources. - O^O 22:15, 23 August 2006 (UTC)

Proposal to include (and in one case use as a replacement) the 'trinity' quote.
Following discussion here, I'm going to propose that we include the following to the lead of each policy. In the case of the NPOV policy, this also includes removing a previous quote on the matter which has been superceeded.


 * In the words of Wikipedia founder Jimmy Wales, "I consider all three of these to be different aspects of the same thing, ultimately. And at the moment, when I think about any examples of apparent tensions between the three, I think the right answer is to follow all three of them or else just leave it out of Wikipedia."

For the sake of clarity, please discuss this on Wikipedia talk:Neutral point of view. --Barberio 18:18, 23 August 2006 (UTC)

My edit summary
In case anyone thinks I left a misleading edit summary here, saying "tweak" when in fact there are lots of changes, it was because I had made a minor edit to Jayjg's version without realizing he had reverted to another in the meantime. I shouldn't have gotten an edit conflict, but didn't, which I've noticed happening a lot recently. Anyway, I do also think the version I tweaked is the better one. SlimVirgin (talk) 20:27, 23 August 2006 (UTC)

Protected the page
It's obvious that an edit war is occuring and people are not discussing, but simply reverting. Time to cool off.

(I consider myself uninvolved since I've not edited the page, although I have made comments on talk. I'm not sure it's possible to find an admin truly unconnected to a core Wikipedia policy in any case) Matthew Brown (Morven) (T:C) 20:34, 23 August 2006 (UTC)