- Moved from Wikipedia talk:Cite sources.
"Reputable" vs. "Appropriate" sources
The tricky thing about a requirement of "reputable" sources is that the appropriateness of a source really depends on what the source is being used for, and in some cases a very un-reputable source may be entirely appropriate. For example, you would never cite a transcript of a Hitler speech for any purpose other than to indicate what Hitler said, but for that purpose it would be the most appropriate source. To take a less extreme example, I would not quote an arts review in a minor newspaper to sum up the work of a major figure, but I would not hesitate to quote it to show critical opinion from a particular place and time. We went through an interesting round of this at Republican/Democrat In Name Only, having to come up with a standard for what would count as appropriate citation of someone being called a RINO or DINO. Maybe "appropriate" is a more operative word than "reputable"? -- Jmabel | Talk 06:48, Jan 24, 2005 (UTC)
- I agree with this. Might it be appropriate to explain the difference between using a source as a primary source, and as a secondary one? To quote Stormfront in an article about them is to use them as a primary source, and is entirely appropriate. To use them to discuss Martin Luther King in an article about him, is to use them as a secondary source, and is entirely inappropriate. Also, just thinking out loud here, could we adapt the term "good enough" that is used in psychoanalysis to describe the "good-enough" parent? Not perfect, but appropriate, adequate, and fulfulling basic requirements? We'd still have to come up with a description of "good enough," however. I have no problem with the word "appropriate." Whatever term we choose, we can define it using ostensive definition i.e. by giving examples, rather than trying to pin down a hard-and-fast definition. SlimVirgin 06:58, Jan 24, 2005 (UTC)
Types of sources
- I like Rbellin's suggestion: "Try to cite sources that are widely available, relevant, and as credible as possible." Maurreen 07:17, 24 Jan 2005 (UTC)
- Rbellin's phrasing is fine with me, too. -- Jmabel | Talk 19:13, Jan 24, 2005 (UTC)
- I have no problem with using "credibility". I'd like to create a subpage and start doing an edit of the current version, then show it to you, but I'm tied up today and won't be able to start it straightaway, just in case anyone's wondering why I mentioned it earlier but then didn't press on. There's an interesting discussion going on on the mailing list about the authority of sources, and the importance of distinguishing between primary and secondary, and when to use one or the other. Check out the archives here. . There were quite a few threads and the headers kept changing so you may have to scout around a bit, but most posts were in the last three to five days. SlimVirgin 19:58, Jan 24, 2005 (UTC)
- Rbellin's phrasing is fine with me, too. -- Jmabel | Talk 19:13, Jan 24, 2005 (UTC)
Uh, I am not sure these terms are being used here appropriately. I don't think there ever is such a thing as an ""appropriate" source as such, because it is appropriate only in relation to the context in which it is used. Indeed, ratehr than having a guideline that says "use appropriate sources" we should say "use sources appropriately" -- then the guideline is for the editor, not the source. In contrast, I do think there are sources that are reputable or disreputable. I think this is based in part on how transparent the author of the source (or, in the case of primary sources, the creator of the source) is about his or her methods and biases. It is also based on public opinion. Both of these can be hard to ascertain, but I think in fact most people in any given field have a pretty good idea of what is reputable in their field. In any case, the measure of a works reputation has something to do with the quality of the work itself; whether it is being used appropriately or not is entirely dependent on the specific situation in which it is being used. Slrubenstein 22:35, 26 Jan 2005 (UTC)
- Yes, fine, whatever, I think you are quibbling. Obviously "appropriate" means "appropriate to the siutation". My problem was with the word "reputable" which would be an inherent property of the sources. Appropriateness always involves a context. -- Jmabel | Talk 23:21, Jan 26, 2005 (UTC)
In choosing our vocabulary, it might be helpful to list the things we're trying to achieve. As I see it, the consensus on this page and on the mailing list is:
- (1) we don't want to discourage new or inexperienced editors from contributing;
- (2) we don't want to erect a scholarly standard so intimidating that editors without an academic background will feel unable to contribute;
- (3) we want to discourage (strongly discourage or prohibit?) editors using dodgy websites with strong POVs as secondary sources, but not as primary sources; that is, the Stormfront website may be used for an article about Stormfront, but it may not be used as a secondary source on Martin Luther King. Question: how to define "dodgy?" Answer: ostensive definition - give examples - we all know what we mean, more or less, and so will other editors;
- (4) we want to create a non-intimidating research culture, in which, as editors become more experienced, they will learn how to use reputable sources in an appropriate manner, how to cite accurately, when to spot that there's something wrong with a source, where to find good sources;
- (5) we want to create a great encyclopedia that will earn the trust of people all over the world, so we want editors to use sources compatible with that goal.
Probably the best way to achieve this is to describe what we're trying to achieve on the page, using all the words we've talked about: reputable, authoritative, credible, appropriate, qualifying each word as we use it, and giving examples of what we mean.
- Zero's stuff is good, with maybe one reservation: sometimes "just a random web page" is obviously well researched even if it lacks citation apparatus. I wouldn't consider something like that a good reference, but it's often better than nothing. I'm not sure what our citation policy should be on those. -- Jmabel | Talk 01:00, Jan 27, 2005 (UTC)
- I think any guidelines about evaluating sources, or which sources to use, should probably go on a different page. Maurreen 07:05, 27 Jan 2005 (UTC)
- As I posted on the mailing list earlier today, there are two different things being talked about here, I think: quality of citations, and quality of sources. Quality of citation is about accurately documenting a source; identifying it, identifying how to find the referenced information, and characterising the source's content accurately. Quality of sources is about how accurate and reliable the source is for that particular use. It might work better to talk about the first here (how to cite sources correctly) and leave the other one for another page.
- One sometimes has to use bad sources (sometimes there are no good sources, or one hasn't been found yet) but one should always do good cites, even of bad sources. —Morven 07:24, Jan 27, 2005 (UTC)
- Excellent point Morven. I mean seriously, people can spend years in graduate school and still struggle with evaluating the quality of source material, so I don't think it is realistic to expect a short Wikipedia article to even begin to address the problems involved. Of course we must still attempt to address it, but we need to be realistic with our expectations. I agree that the first step is encouaging and facilitating people to provide a citation for sources used in an article. Evaluating the credibility (or appropriateness) of sources can then get thrashed out as needed. older≠wiser 13:00, Jan 27, 2005 (UTC)
I agree to with Morven. I think it was in the same spirit that I made the point about appropriate verus reputable -- I don't think I am quibbling, Maureen -- I think we do not want to confuse newbies. I want to add one point to SlimVirgin's comments, all of which I agree with. I think the single biggest turn-off to newbies is not complex policy pages (and yes, I still agree that policy pages should not be too complex); the biggest turn off is when they soon find themselves in edit wars they do not understand. So my main criteria for developing policy pages is, we want to help them avoid getting into edit conflicts early on that would turn them off. Since a new editor can be criticized byothers for using non-reputable sources or for failing to use sources appropriately, we need to distinguish between the two (as Morven distinguishes between quality of sources and quality of citations) and explain them as simply and clearly as we can. But "reputable sources" is crucial, because many edit-wars are triggered over what constitutes a reputable sourc. Slrubenstein 17:28, 27 Jan 2005 (UTC)
- Just for clarification, Slrubenstein, I think you have me confused with someone else. I haven't said anything about quibbling. Maurreen 17:44, 27 Jan 2005 (UTC)
Damn -- I am very sorry, Maurreen. I meant Jmabel. Slrubenstein 19:05, 27 Jan 2005 (UTC)
- No problem, not to worry. Maurreen 05:04, 30 Jan 2005 (UTC)
Interpretation and original research in Wikipedia
Every text in Wikipedia that is not a direct quote is an interpretation of another text. Even a summary is an interpretation of the original text. Furthermore, deciding which arguments to present for and against a theory in an article is interpretation and original research. So all articles in Wikipedia contain some degree of interpretation and original research. Therefore the guidelines cannot just decide which information to allow or not to allow. It must also be guidelines for this discussion and interpretation.
A rough draft may be that all facts should be supported from outside sources, if demanded. The only theories in Wikipedia should be those held by "many" people or an authority. One guideline regarding the discussion of a theory and selection of arguments might be that one cannot argue against "stronger" arguments. A hierarchy might be
1. Peer-reviewed studies or government statistics 2. Academic press 3. Opinion held by an authority 4. Opinion held by many people who are not authorities (5. Opinion held by only one or a few people)
For example, it should not be allowed to argue against peer-reviewed studies by refering to common opinion. In the more controversial topics in Wikipedia this hierarchy for discussion is already in place by itself. Look for example at capitalism or race and intelligence. Ultramarine 12:59, 30 Jan 2005 (UTC)
- This makes a lot of sense to me. Maurreen 14:50, 30 Jan 2005 (UTC)
- One example giving rise to a problem with Ultramarine's heirarchy is David Irving. His early material was not peer-reviewed in the sense of being published in academic journals, but it was hailed by certain British academic historians who weren't specialists in Holocaust studies. Yet journalists who evidently knew more than the historians criticized it. We would want to be allowed to quote those journalists even in the absence of specialist criticism. SlimVirgin 02:40, Feb 1, 2005 (UTC)
- If the historians were not specialists then they were not authorities. Thus, if David Irving has not published in peer-reviewed journals or academic press then I think one can certainly quote journalists with good historic knowledge.
- Today an authority might be someone who has published academic material in this area previously. Authorities earlier in history might be those today recognized as such by current authorities.
- I hope that you agree that some kind of guidelines are needed? Otherwise anyone can quote an opinion from a newspaper as evidence against peer-reviewd studies. Ultramarine 14:23, 1 Feb 2005 (UTC)
There are three issues here. First, what do we mean by "interpretation." Fr the purposes of Wikipedia policy (as opposed to, say, a book on literary theory or hrmeneutics) I strongly disagree with Ultramarine's claim that everything is interpretion. I think what we mean here is an account of a text or event that makes claims about what it means that are not obvious -- meaning, that are not widely shared by others. Summaries of a text or event that claim to provide the "plain" meaning -- meaning, what most others would consider obvious -- are not for our purposes, interpretations.
Second, what is a high-quality source. I still believe that it is impossible to operationalize this in a way everyone here would find acceptable -- and I believe just as strongly that this is an ideal to which we should aspire. I think Ultramarine's list of sources could serve as examples but not as rules.
Third, what we can operationalize is the distinction between controversial and non-controversial sources. We need to make clear that the controversies surrounding controversial sources must be explained; that the NPOV rule applies and we should make sure other views are represented; that articles should try to use as many non-controversial sources as possible. Slrubenstein 16:24, 1 Feb 2005 (UTC)
- I agree that is is impossible to exactly operationalize what is a high-quality source. I also think my list or similar list might be useful as examples but probably not as exact rules. I do think that Wikipedia should allow sources with lower quality if there are no high quality sources available. For example, I think Wikipedia should briefly describe conspiracy theories popular among many people even if these are completely ignored in academic research.\
Great! Do you think the policy makes this clear? Slrubenstein 17:56, 1 Feb 2005 (UTC)
- I am more pessimistic about whether Wikipedia can be free from interpretations. If a particular summary of another text provides the "plain" meaning obvious to most people is always open to debate. Furthermore, the selection of sources and the decision which are of higher quality is also interpretation. Ultramarine 17:43, 1 Feb 2005 (UTC)
I am not questioning whether Wikipedia can be free of interpretations; I do not think anyone here believes this. I am, however, questioning your definition of the word "interpretation." Slrubenstein 17:56, 1 Feb 2005 (UTC)
Authoritativeness and Relevance of sources
Every text in Wikipedia that is not a direct quote is a interpretation of another text. Even a summary is an interpretation of the original text. Furthermore, deciding which arguments to present for and against a theory in an article is interpretation and original research. So all articles in Wikipedia contain some degree of interpretation and original research. Therefore the guidelines cannot just decide which information to allow or not to allow. It must also be guidelines for this discussion and interpretation. In the more controversial topics in Wikipedia informal rules for this discussion is already in place by itself. Look for example at capitalism or race and intelligence.
A rough draft might be that all facts should be supported from outside sources, if demanded. The only theories in Wikipedia should be those held by "many" people or an authority.
Regarding sources in the discussion of a theory, one important factor is the authoritativeness of the source. One guideline might be to avoid arguing against more authoritative arguments with less authoritative. For example, to avoid arguing against peer-reviewed studies by referring to common opinion. Although common opinion could certainly be mentioned, especially in areas where there are no academic research as in many conspiracy theories. An example of the hierarchy of authoritativeness might be:
1. Peer-reviewed studies or government statistics 2. Academic press 3. Opinion held by an authority. That might be a person who has previously done academic research in discussed area. Or dead persons considered authorities by authorities today. 4. Opinion held by many people who are not authorities (5. Opinion held by only one or a few people who are not authorities).
Another factor is the relevance of the source for the theory. A source with very little relevance for the theory should be avoided. An example of the hierarchy of relevance might be:
1. Discusses the theory directly. 2. Discusses general principle important for the theory. For example, a page about peak oil might reference an article that discusses solar power in general without mentioning peak oil. 3. Statistic that if generalized affects the theory. For example, a page about capitalism or Marxism might reference US government statistics about growing differences in income. (4. Anecdotal evidence for the theory. For example, a page about poverty in the third world might reference the income of a particular person.) Ultramarine
- I mostly agree; one demurral: the priority of peer-reviewed studies varies with subject matter. For example, a peer reviewed academic paper on punk rock would probably be authoritative on dates and names, but wouldn't necessarily carry any more weight than a review in Maximum Rock'n'Roll for its judgement of a band. There are simply fields where the real experts aren't in academia. -- Jmabel | Talk 21:53, Feb 1, 2005 (UTC)
I think we need to be careful about this heirarchy of relevance as it could be used by some to introduce their own synthesis of published material i.e. their own arguments. While others have made the point that we don't want to stifle creativity, and I agree with that, we also don't want people using Wikipedia as a platform to publish their own essays. Dipping into a variety of sources for a paragraph here and there that supports an argument is personal essay-writing. Ultramarine's example about a page about Marxism referencing U.S. government statistics about differences in income is an example of what editors ought not to do, in my view. Perhaps that's what you meant too, Ultramarine. I'm all for giving examples within a "heirarchy of relevance," but I'd like to see a cut-off point where we say: "And these examples would count as original research." SlimVirgin 22:39, Feb 4, 2005 (UTC)
Quoting of unknown pressure groups
Lately, I have come across a number of edits which included quotations (sometimes unattributed) from various groups which had a very specific point of view on a question. Of course, articles must represent all points of views, and I do not dispute that.
However, it appears to me that the value of a reference should be assessed on four aspects:
- Quality of the analysis: are there any glaring factual errors? (e.g. a study placing Italy in Africa should probably not be cited as reference)
- Expertise: is the author a recognizable institution working in the field? (e.g. a study by an obscure self-employed physicist, not published in a peer-reviewed journal, does not have the same value as a study by researcher in an internationally known institution publishing in a first-class scientific journal, even though the latter may contain mistakes)
- Identifiability: is the source traced to a clearly identifiable person, institution, publication, or group? Nowadays, anybody can start a "group" or "institute" and have Web pages, so an "institute" may actually be a handful of persons. That is why I much prefer it when the source has its own well-documented page on Wikipedia. Surely, it matters a lot, when reading the opinion of a source, to know whether the source is a reputed human rights organization, a personal project, or an institute affiliated with or funded by a political party or religious organization.
- Representativity: does the group represent a sizeable part of public opinion? To me, it seems that quotes from mostly unknown pressure groups are a subtle workaround of the policy on weasel words, the policy of citing sources, and the policy against citing one's personal opinion.
It seems to me that these aspects are too often ignored, and that some contributors push too much irrelevant opinion quotes in the guise of NPOV. To draw a parallel, the article on George W. Bush does not, and should not, contain opinion pieces from every left-wing group on Earth saying that Bush is stupid and that his policies are a disaster; I do not see why this should be different for other articles.
In addition, I think that, all too often, sources are improperly attributed. For instance, the opinion of a single judge in a lower court in a country is not the opinion of the government of that country. The opinion of some people invited at a conference organized by institution X is not the official point of view of association X. Etc.
I think that all this should be reflected in official Wikipedia policy. David.Monniaux 11:14, 11 Feb 2005 (UTC)
- You make very good points and I agree with substantially all of it. But I welcome this kind of debate as being on an entirely higher level of argument than arguing whether we should have sources and properly cited facts at all. I think your criteria above for deciding on the quality of a source is very good, and one that meets more of those criteria should be preferred over one that meets less. Maybe others will have some idea on improving the criteria, and maybe we can even create a separate page for references criteria, start with the above and build a consensus for ways to distinguish between more respected sources and less. It will of course nevr be perfect, but you have made a great start. - Taxman 13:55, Feb 11, 2005 (UTC)