This is an old revision of this page, as edited by EEng (talk | contribs) at 23:28, 7 August 2012 (→Ancestry.com: //very// limited used of Ancestry as source on WP). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.
Revision as of 23:28, 7 August 2012 by EEng (talk | contribs) (→Ancestry.com: //very// limited used of Ancestry as source on WP)(diff) ← Previous revision | Latest revision (diff) | Newer revision → (diff) | Biography: Arts and Entertainment C‑class | |||||||||
| ||||||||||
 | Nevada (inactive) | |||
| ||||
Other TV show
I saw Richard Harrison appear on a tv show on Destination America called Top TEn Ways to Win in Vegas, made in 2002. — Preceding unsigned comment added by 72.213.183.214 (talk) 03:00, 19 June 2012 (UTC)
Ancestry.com
 |
|
1. Are birth and death records on Ancestry.com user-generated? 2. Would they be usable for that info in BLP articles like this one? 3. Even if the public records there are not added by users, would WP:BLPPRIMARY preclude using them?
Ancestry.com is not a reliable source, because the material there is user-generated, and sources whose content is user-generated are not reliable, under WP:USERG. This is not my declaration by fiat, mind you, but is reflective of the community, as illustrated by the following four past discussions at the Reliable Sources Noticeboard:
- March 2010 Three editors say it's unreliable because its content is user-generated: David Underdown, Blueboar and radek.
- June 2010 Three additional editors agree that it's unreliable: Jayjg, A Quest For Knowledge, Bielle, and MarmadukePercy. One editor, JakeInJoisey, disagrees. Another, Dlabtot, questions JakeInJoisey, but doesn't state outright an opinion of his own on Ancestry.com.
- July 2010 Four additional editors concur with previous findings: Auntof6, Gadget850 (Ed), TheFeds, and Slp1.
- March 2011 Another editor, Jezebel'sPonyo, defers to the previous discussions, and two other editors, Andrew Lancaster and CalvinTy, also argue that census records should either not be used or be used in a highly limited manner because they are primary sources instead of secondary ones.
The bottom line is, if the information there comes from users, as with imdb, wikis, blogs, open forums, etc., they it's a user-generated site, which violates WP:USERG. One editor from the second discussion, JakeInJoisey, says, "There is, I believe, considerable 'editorial oversight'", but when another editor responds to that statement to ask him what the basis is of that assertion, JakeInJoisey never responds. And another editor in that same discussion, Bielle, even says that he is a contributor to Ancestry.com, and adds information frequently, saying, "There is no editorial oversight." Another editor, A Quest For Knowledge, acknowledges that user-submitted content is not reliable, but asks if one of the articles written by Ancestry.com staffers would be, and MarmadukePercy says, "I am unfamiliar with the magazine, but from what I've seen, I'd be reluctant to cite Ancestry.com as a source for anything – except...where Ancestry obviously struck a deal to carry proprietary content from a reputable source, the Dictionary of American Family Names published by Oxford University Press." Thus, it seems fairly clear that most in the community feel that most of the material on that site is unreliable. Nightscream (talk) 00:19, 7 August 2012 (UTC)
- The certificates are certainly not user-generated, I'm not sure what you are saying (family trees are, but their records aren't) - if we can verify 'for sure that the registry entry is for him, that it should be okay. – Connormah (talk) 01:36, 7 August 2012 (UTC)
- As a recommendation, if you're going to respond here, then it's a bit of overkill to express the same sentiment on my talk page. The most you should do if you think I won't see your response here is to simply inform me that you've made a response here, rather than repeat it. And I can assure you that this talk page is indeed on my watchlist, just as the article itself is.
- As to your point, who do you think adds those certificates and records? Simple. Users. Not reporters, journalists or researchers, but people with registered accounts on Ancestry.com over whom editorial oversight is not exercised, as established above.
- In addition, public records should not be used for biographical material, as indicated BLP: Misuse of primary sources. Nightscream (talk) 02:05, 7 August 2012 (UTC)
- No - the records are not added by ordinary users - I cannot add records nor can you - the offices obtain the records and index them accordingly. I do discourage using these records, but I can see some use for them in some limited circumstances, but in any case, as I stated on your talk, there has to be a certainty that the correct record is being used. – Connormah (talk) 02:36, 7 August 2012 (UTC)
- Re: the first thread above - yes, rootsweb is user generated, but it is completely different from Ancestry. Rootsweb IIRC is a site where you can string together a family tree using the original records from Ancestry. The Ancestry records themselves are not user generated. – Connormah (talk) 02:38, 7 August 2012 (UTC)
- Thank you for your considered response. Unfortunately it is based on a false premise that ancestry.com is a user-generated site. You say: "The bottom line is, if the information there comes from users, as with imdb, wikis, blogs, open forums, etc., they it's a user-generated site, which violates WP:USERG". Ancestry.com is not a user-generated site. While it contains some user-generated information (family trees), almost eveything else is from publically available official records eg county or state halls of record eg births, marriages, and deaths, as well as census, military (including draft cards), newspapers, etc. The modus operandi is that ancestry.com pays for their employees to scan official documents and they upload these records, not subscribers or "users". It is similar to what Google does with Google Books or Google News, or www.newspaperarchive.com does. The sources are RS, and should be evaluated accordingly. You acknowledge that with reference to Dictionary of American Family Names that there is the possibility of RS sources within the site. Did you notice the specifics of each reference provided? In addition to the ancestry.com attribution, there is specific reference to such sources as the North Carolina Birth Index, and the means to verify those sources. As regards editorial oversight, none is needed as the records are photostatic scans - there is no analysis or interpretation provided or needed. It presents the documents which speak for themselves. This is relevant as the prohibition (or rather reluctance to use Primary sources), is where these form the basis of OR. Again, none is provided or needed. For example, the assertion that Richard Kevin Harrison was born on March 22, 1965 in Lexington, Norh Carolina is appropriately sourced. It requires no mental cogitation on behalf of the WP editor. The record lists his parents and there are other details provided that ensure that the right person's records are accessed and used. Ancestry.com is a helpful tool for biographical articles. It is more accurate generally that newspaper obits (which are generally sourced from info provided by a family member).smjwalsh (talk) 03:21, 7 August 2012 (UTC)
Connormah: "The Ancestry records themselves are not user generated."
Smjwalsh: "As regards editorial oversight, none is needed as the records are photostatic scans..."
And who adds those records and photostats to the site? What documentation or evidence do you have that they're added by paid employees, and not users? If 11 different people say that the material on that site is added by users, including one editor who says that he himself is an ancestry.com user who has added material himself, that pretty much clinches it, I think.
Moreover, even if what you said were true, did you not see the final portion of my message above regarding WP:BLPPRIMARY? Did you read that portion of the BLP policy? It says:
"Do not use trial transcripts and other court records, or other public documents, to support assertions about a living person. Do not use public records that include personal details, such as date of birth, home value, traffic citations, vehicle registrations, and home or business addresses."
So even if the material on ancestry.com came from public records, they wouldn't be usable.
As for Rick's birthdate, that isn't sourced to Ancestry, it's sourced to a Dispatch article.
I have requested that editors from those previous discussions join this discussion to address the issue of both the assertion about paid employees, and BLPPRIMARY. I have not restricted my invitations to those who opined that the site is user-generated, but also extended it to JakeInJoisey, who asserted that there was considerable editorial oversight there. I will also open an RfC to attract more editors.
And incidentally, this has nothing to do with OR. OR refers to material based on personal knowledge, which has nothing to do with the approach to using primary sources, a completely separate and unrelated policy. While PSTS is indeed located on the OR policy page, OR itself refers to material that is based on no published sources, not primary ones, as it says in the opening sentence of the OR page. Nightscream (talk) 09:22, 7 August 2012 (UTC)
- Hi. Commenting as per request.
- First general point: ancestry.com is quite likely to be an RS for some purposes. We always need to remember that reliability is context relative. Very few sources can not be used for anything. Very few sources would be considered reliable for everything. I am not in favour of creating over simplified rules about sourcing which make a single source always good or always bad, and I believe my point here is a widely shared one.
- Second general point. Ancestry.com is a big website with many types of source. We should not mix those things up.
- For example they have some scanned books, something like google books. These are often neither user-generated nor primary. Of course in such cases, it should be possible to name those and avoid a debate?
- User generated sources are rarely reliable. I would really have to be convinced about any particular example, and I am having trouble thinking of any.
- It is correct to point out that wherever a living person is concerned, and the information might be sensitive, we should be more strict than usual about sourcing.
- Primary sources in ancestry.com are probably the most common tricky area. But some people over-simplfy. There is no blanket ban on using primary sources, only a caution that they are not always appropriate. However many previous discussions concerning ancestry.com have been about things like using censuses or parish registers for dates. These are classic cases of where primary sources are normally not able to be used because as any good genealogist will tell you, interpretation of such documents is "non obvious", and often trick up even very good researchers. There MIGHT be cases where such a document might be a good source for some very particular type of information, but they would be special cases. If we are just talking about sifting through raw data to find a date or age or place of birth, then this is rarely going to be "non obvious".--Andrew Lancaster (talk) 12:19, 7 August 2012 (UTC)
- Andrew, actually there is a blanket ban on using primary sources such as public records in BLP articles. See WP:BLPPRIMARY--Slp1 (talk) 18:06, 7 August 2012 (UTC)
- RFC comment
- This looks like a WP:RSN request. Every RSN discussion needs two things:
- 1) What is the article and content under discussion?
- 2) What is the source being used to support it?
- This RFC provides neither 1) nor 2), so no answer is possible. It's about as meaningless asking saying "Is cnn.com a reliable source?" You'd have to answer, Yes (or at least parts of it) for some things, No for others. I'd say cnn.com news articles are generally a reliable source for factual information about world events, but cnn.com-hosted user-generated "iReport" videos or the forum-type discussions of articles that cnn.com hosts are not reliable sources for BLP information.
- As has been pointed out, ancestry.com has a lot of different kinds of information available and it is absolutely incorrect to make a blanket statement like ancestry.com "is" or "is not" a reliable source. The user-generated family trees fail WP:USERG and cannot be used to support claims of relationships between people. The many scans of primary source documents can be used as primary sources with care.
- Regarding WP:BLPPRIMARY, what it actually says is:
- Exercise caution in using primary sources. Do not use trial transcripts and other court records, or other public documents, to support assertions about a living person. Do not use public records that include personal details, such as date of birth, home value, traffic citations, vehicle registrations, and home or business addresses. Where primary-source material has been discussed by a reliable secondary source, it may be acceptable to rely on it to augment the secondary source, subject to the restrictions of this policy, no original research, and the other sourcing policies.
- "Exercise caution" does not mean "it is forbidden."
- So the answer is, "It depends." We need to have the article content and which exact ancestry.com-provided source to come up with an answer.
Zad6813:24, 7 August 2012 (UTC)- For the record, I am a professional historian who has used ancestry.com for at least 6 years. As indicated above, subscribers to the site CANNOT add official records. All they can do is identify transcription errors or quality of scan issues. What other WP editors indicate is still consistent with my position if their primary involvement is in the area of creating family trees. However, even in relation to that there is no capacity to upload records or documents relating to birth, marriage, death, military, newspapers, etc. I am aware of the FULL statement regarding using primary sources and am an experienced enough editor to be able to use those sources with extreme diligence. However, the check is there for other users to dispute contentious material and the subjects of BLP to request removal if a line is crossed. While this is a good general discussion. perhaps it would be best to focus on specific sources used in THIS article and associated articles.smjwalsh (talk) 14:02, 7 August 2012 (UTC)
- SMJ, I am actually concerned about the edit you just made using ancestry.com. What reliable source do we have that indicates the primary-source record you found matches this individual?
Zad6814:05, 7 August 2012 (UTC)- Thanks for your interest. Give me an example of what you are concerned about and I can discuss better. The birth certs indicate both father and mother, place and date of birth, address of parents, etc. These correlate with other sources eg Rick Harrison's book, various newspaper articles referenced already in the article. However, would you prefer ancestry.com reference as intermediate source be restored?smjwalsh (talk) 14:49, 7 August 2012 (UTC)
- SMJ, I am actually concerned about the edit you just made using ancestry.com. What reliable source do we have that indicates the primary-source record you found matches this individual?
- For the record, I am a professional historian who has used ancestry.com for at least 6 years. As indicated above, subscribers to the site CANNOT add official records. All they can do is identify transcription errors or quality of scan issues. What other WP editors indicate is still consistent with my position if their primary involvement is in the area of creating family trees. However, even in relation to that there is no capacity to upload records or documents relating to birth, marriage, death, military, newspapers, etc. I am aware of the FULL statement regarding using primary sources and am an experienced enough editor to be able to use those sources with extreme diligence. However, the check is there for other users to dispute contentious material and the subjects of BLP to request removal if a line is crossed. While this is a good general discussion. perhaps it would be best to focus on specific sources used in THIS article and associated articles.smjwalsh (talk) 14:02, 7 August 2012 (UTC)
- This looks like a WP:RSN request. Every RSN discussion needs two things:
My comment following is in response to a request on my talk page from Nightscream: Others are making this key point: to the best of my knowledge, the original "institutional" records are added by the site owners. That makes them as reliable (or not) as any original historic record of, for example, passenger manifests, or death registers or parishes records. Personal records, like copies of birth certificates or marriage certificates, are added by users, and there is nothing to say they have not been altered or correct connections have not been made. The connections between names on the various institutional sources, and specific individuals, are made by site users. I found, for example, the names of an aunt and a cousin on a ship's manifest on a date I knew they had sailed from England for Canada. The names are sufficiently distinctive that it is unlikely to have been any two but them. However, to connect them to me (that is, that they are my aunt and my cousin) you would have to know what I "know" and accept all I know as historical fact. That is where Ancestry.com is not reliable. Bielle (talk) 14:44, 7 August 2012 (UTC)
- I'm also respoindig to a request made on my talk page. My comment in a previous discussion on WP:RSN relating to the specific usage of some information has eben slightly misrepresented. In that instance the information was from a user-generated portion of the Ancestry website (the website includes discussion boards, user-published family trees and the like), but as I mentioned in my comment, large parts of Ancestry are the scans of source documents and associated transcriptions/indexes. Those are under the editorial control of Ancestry staff, and it certainly appears that the information in this case falls into that category (on a matter of sourcing, I'd prefer to see a direct link to the specific recod, with a note that a subscription is required). However, as others have noted, while WP:PRIMARY does allow for the cautious use of reliably published primary sources, the issue is the interpretation, without other information we cannot be sure that the correct record has been identified (the subject's names are not that unusual), so really we need secondary sources too to confirm that it's the correct birthday, and/or both the parents' names are both correct and the location of the birth is correct and so on. There is some tension between WP:PRIMARY and WP:BLP in these sorts of cases - we shouldn't be "outing" things that aren't generally already in the public sphere (that's less of an issue for historical figures). Also of concern is the level of reversion that's been going on in this article, WP:3RR hasn't been breached I don't think, but once a source is challenged, it's better to talk about it and get consenus for its use (and the inclusion of the facts it is supporting) rather than just asserting that you are correct and pig-headedly putting it back into the article. David Underdown (talk) 14:59, 7 August 2012 (UTC)
I am responding to the request for comment on my talkpage. I would agree with others that statinng what the content that is being disputed would be helpful. In any case, I have temporarily at least removed the birthdates and names of Mr. Harrison and his children. This is per BLP and should NOT be restored without getting consensus here. BLP:Primary is very clear that WP does not use primary sources such as public documents (e.g birth records) for information about living people (which seems to be the case for some of the informationhere). I've left in Harrison's birth year, as that date is quoted in the autobiography. BLP policy is also very clear that we have to be extra cautious about including full dates of birth and names, which is also the case for obvious reasons.WP:DOB. This is most especially in the case of non-notable children. See WP:BLPNAME. Bielle and David explains some of the reasons why primary sources should not be used and why, per BLP, the information needs to stay out of the article until a different consensus is reached. See . This could come if secondary (non-public records) support for this material is found. Slp1 (talk) 15:43, 7 August 2012 (UTC)
- Regarding Harrison's family, including his children: he names them in his book License to Pawn: Page 34 refers to exnuptial pregnancy of RH's later 1st wife Kim, subsequent miscarriage, marriage. Page 38 refers to Corey's birth ca.1983, and birth of second son Adam 2 years later ca.1985. Page 39 refers to separation of RH and KIm ca. 1985. Page 40 identifies 2nd wife as Tracy from WV. Page 41 refers to their wedding. Page 42 identifies name of Corey's wife. Photo section between pages 84 and 85 idebtifies RH's brothers by name, mother's name, names and photos of Corey and Adam. Page 77 identifies RH and Tracy's son as Jake born in 2003. My point. If Harrison is willing to identify these (and more) details in his best-selling book, why the reluctance to do so in this article? As this book is certainly RS, then surely these details can be restored.smjwalsh (talk) 16:21, 7 August 2012 (UTC)
- The extra caution comes from the fact that Misplaced Pages provides a much more accessible and searcheable format than a book does, and of course that some of the information you wanted to include wasn't even in the book-or was given approximately -but was the result of your (probably correct, but nevertheless) original research through primary sources.
- As to the reason for the rules, you'd have to discuss that at WP:BLP, but let me give you a possible scenario. Your edits noted a different first name for his eldest. We don't know why the first name was dropped; maybe he hated the name; maybe he was teased unmercilessly at school about it; maybe his name was changed by deed poll because of this. Whatever the reason, he is a non-notable person who does not need to have his full name and age published so that it will be found whenever anybody googles his name when thinking about hiring him, dating him or whatever. The question is does it help our understanding of Mr. Harrison to know his son's birthname, and the answer is no. I think we can discuss whether it might helpful to have a bit more detail, saying that he had a son, born c. 1983 and another born c 1985 or similar, since it is in the book, but that's quite another matter from the stuff sourced to birth certificates. --Slp1 (talk) 16:42, 7 August 2012 (UTC)
- You can see I know nothing about this show! I see now that Corey is actually on the show so to my mind it is fine to name him at least. --Slp1 (talk) 17:05, 7 August 2012 (UTC)
- Thank you for your reasonable response. No-one is expecyed to know everything about every article. I realise some are on this page because of previous discussions about ancestry.com as a RS, rather than as folk interested in the subject of this article. Corey's birth name is available in various publically available RS, including newspaper articles. It is not a huge secret. Obviously he is a notable person (although without a WP article yet). I suspect the specific birth dates can be found online in verious RS. I have not looked (yet). Corey also contributes a chapter in his father's book that disloses (among other things) his drug addictions, wife's name and details, etc. Again, if they are willing to disclose this info in a book or in various interviews, surely it can be included in WP articles where relevant.smjwalsh (talk) 17:21, 7 August 2012 (UTC)
- (edit conflict) Can you do your searches and list the reliable sources you find here? That's the simplest method. If there are reliable secondary sources for all or some of the information, then the problem of using primary sources will disappear, and we only need to deal with potential privacy issues.
- If Corey is notable enough to need an article (and I am not convinced he is, based a quick search for reliable sources about him), then the info in that book would be fine for his article, subject to all the usual rules about Undue weight, BLP etc. --Slp1 (talk) 17:33, 7 August 2012 (UTC)
- Thank you for your reasonable response. No-one is expecyed to know everything about every article. I realise some are on this page because of previous discussions about ancestry.com as a RS, rather than as folk interested in the subject of this article. Corey's birth name is available in various publically available RS, including newspaper articles. It is not a huge secret. Obviously he is a notable person (although without a WP article yet). I suspect the specific birth dates can be found online in verious RS. I have not looked (yet). Corey also contributes a chapter in his father's book that disloses (among other things) his drug addictions, wife's name and details, etc. Again, if they are willing to disclose this info in a book or in various interviews, surely it can be included in WP articles where relevant.smjwalsh (talk) 17:21, 7 August 2012 (UTC)
- You can see I know nothing about this show! I see now that Corey is actually on the show so to my mind it is fine to name him at least. --Slp1 (talk) 17:05, 7 August 2012 (UTC)
First, thanks to all for joining this discussion. A couple of points:
Zad68, this has already been at RSN four times, and a total of about ten or eleven people indicated that the material on ancestry.com was user-generated, including one editor, Bielle, who said he was a contributor to that site. So it seemed pretty cut and dried. But when Smjwalsh disputed this, I felt I needed clarification on this, so I wanted a broader consensus. The fact that I often get one response from one editor before the thread dies and is archived on RSN is another reason why I opened up an RfC. I hope this was okay.
As for your point about asking more specific questions about what material from a source is being used for what portions of an article, I have amended the questions underneath the RfC banner above to be more specific.
David, I'm sorry if I misrepresented anything you said in one of those discussions; I tried my best, when examining the discussions, to represent each editor's position as accurately as I could. As to the matter 3RR, I do indeed stop reverting when it's clear that the other editor is not going to back down, and/or has a legitimate policy-based rationale for disputing or reverting. I saw that Smjwalsh indeed had such a line of argument here after my last revert; Had I seen his post here prior to my last revert, I would not have reverted.
Smjwalsh, I want to thank you again for editing these articles. Your dedication to adding all this material and sourcing everything definitely improves the project. I want to make clear that I will not oppose using ancestry.com if it appears that a preponderance of the editors here say that it's okay to do so. But let's be clear: My previous opposition was to ancestry.com, and not Rick Harrison's book. I don't think anyone has stated any opposition to citing his book. Nightscream (talk) 17:30, 7 August 2012 (UTC) (Addendum: I see now that you were referring to Slp1's removal of material sourced to his autobio. That should not have been done.)
- Thanks for your comments. I am impressed by your willingness to change your opinion based on more compelling evidence. Perhaps we have moved on from the discussion as to whether ancestry.com is a user-generated site or not. If not, here is what their own website indicates: "The foundation of our service is an extensive and unique collection of billions of historical records that we have digitized, indexed and put online since 1996. We have developed efficient and proprietary systems for digitizing handwritten historical documents, and have established relationships with national, state and local government archives, historical societies, religious institutions and private collectors of historical content around the world. These digital records and documents, combined with our proprietary online search technologies and tools, enable our subscribers to research their family history, build their family trees and make meaningful discoveries about the lives of their ancestors" (emphasies mine). It also discusses the sections that are user-generated, but clearly indicates that there are 10 billion historical records they have uploaded (ie the employees of ancestry.com), compared to 38 million family trees and 136 million user-contributed photographs, stories, and documents. In my mind, this clearly established that while ancestry.com has user-generated content, that the preponderance of its material is their collection of historical documents and records their employees have uploaded. I hope this clarifies this mattersmjwalsh (talk) 17:36, 7 August 2012 (UTC)
- To be honest, this discussion of ancestry.com is moot, and I think the RFC should be closed. Ancestry is a repository of public records and public records such as these cannot be used in biographies of living people such as this one. Because of the original research required with the primary records such as this, they are not the greatest source for any WP article, but in a BLP they are an absolute no-no. Find some secondary sources. --Slp1 (talk) 17:59, 7 August 2012 (UTC)
- It seems that the point is no longer whether it's user-generated, but whether to use public records as primary sources. Editors such as Slp1 say we can't. Nightscream (talk) 18:24, 7 August 2012 (UTC)
- A couple of corrections: the issue isn't whether to use public records as primary sources, but whether we can use primary sources such as public records in a BLP. And the answer is no. That is not my opinion but the very clear statement of our BLP policy. "Do not use trial transcripts and other court records, or other public documents, to support assertions about a living person. Do not use public records that include personal details, such as date of birth, home value, traffic citations, vehicle registrations, and home or business addresses.". As I said, if there are secondary sources out there, let's use them, but ancestry.com is a non-starter. --Slp1 (talk) 19:09, 7 August 2012 (UTC)
- It seems that the point is no longer whether it's user-generated, but whether to use public records as primary sources. Editors such as Slp1 say we can't. Nightscream (talk) 18:24, 7 August 2012 (UTC)
One scribbler's thoughts
Most of this was drafted many hours ago, so it may not engage later posts as well as it should. Also, I had intended to write five quick sentences, but things got away from me. I hope at least someone finds this useful...
I have no idea who Rich Harrison is -- haven't even looked at the article. I stumbled on this discussion by accident. These are my opinions based on six years of academic research in which I regularly consult Ancestry:
(1) The scans of census records and so on are exactly analogous to court records and are WP:PRIMARY. They may seem straightforward to interpret but they are not. The following example comes (with some simplifications) from my own research. A census return appears on its face to establish that, on the enumeration date ("Name of each person whose place of abode on April 1, 1880 was..."), person X was living with person Y at address A. But it turns out that X did not move to address A until some weeks after April 1 (census enumerators often didn't come by until several months after the enumeration date, and would get the information from whomever happened to be home, or even from neighbors, so exactly who was in residence on April 1 wasn't always accurately remembered). It also turns out that person Y was listed with a completely incorrect last name (I still don't know why that is, but there's no doubt). Establishing these errors required detailed knowledge of census-taking practices at the time, plus comparison with other sources from several states and several foreign countries (literally). After all this work is (God willing) published in (ahem) a highly prestigious academic journal, then an image of the census record, with dicussion of the above, might make its way into WP. But for someone to just upload the census record image to WP and write, "X and Y lived at A on April 1" would be utterly inappropriate and completely wrong.
- (1A) The only way I can imagine that a census record image would be usable in a WP article would be as an illustration of a fact established by the usual reliable sources. A good example is seen in Malcolm X. That Malcolm X's family lived at a certain address is established by sources cited in the text. Given that, there's nothing wrong (I won't go into potential copyright issues here) with fetching the corresponding census sheet off Ancestry as a fascinating image.
- (1B) However, if that image were to show an "M.L. King" living next door, but no source comments on that, it would be completely inappropriate to say anything about that in the article; in fact, given the liklihood that readers would jump to an unsupported conclusion in this case, an argument might be made for excluding the image from the article after all.
- (1C) I gather from comments above...
- For example, the assertion that Richard Kevin Harrison was born on March 22, 1965 in Lexington, Norh Carolina is appropriately sourced. It requires no mental cogitation on behalf of the WP editor. The record lists his parents and there are other details provided that ensure that the right person's records are accessed and used.
- ...that the question at issue is whether some birth certificate record or such can be used as a source on WP to establish this birthdate. In my opinion absolutely not. I wish it were so that official records, legal documents, and so on were always so straightforward that they can be confidently given their "obvious" facial interpretation, in isolation. Unfortunately it's just not so, as the census example I've already given shows.
- (1D) In fact, let me give a second example of the difficulty of interpreting primary records. Let's suppose:
- A scan of a real estate deed, on file in a Deeds Office somewhere, appears to show that Property P was bought by Person A on 1/1/1990.
- Another deed appears to show that Person B bought the property on 1/1/1999.
- A third scan, of a page from a Deeds index, apparently shows that there were no deeds affecting this property in between those two dates.
- Would we be be justified in writing, "Person A owned Property P on the date of the accident, 1/1/1995"? Absolutely not. Who owned something on a certain date is a very tricky business -- to be certain no other deed intervenes between 1990 and 1994 requires an extensive knowledge of state law and the particular recording and indexing practices of a given county (see Title insurance#Reason for existence) and there's no way for WP editors to effectively settle this kind of question.
- Of course, when we use a secondary source we are relying on that source to have done that complex work, or to have consulted reliable people to do it for him or her. It may even be that another secondary source comes to a different conclusion about who owned what. So haven't we merely substituted a new debate -- over which of the two secondary sources should be believed for their interpretation of the primary sources -- for original debate -- over the primary sources themselves? Yes, but we're still better off, and here's why: the technical minutiae of various primary sources differ widely, but the indicia of source reliability are very much the same across disciplines: reputation, writing style, internal consistency, internal sourcing, and so on. Furthermore, secondary sources often engage in debate with each other and criticize each others reasoning -- enhancing editors' ability to evaluate them -- while primary sources pretty much sit there naked and mute, for better or worse. Thus discussing the secondary sources is usually more productive than directly debating the underlying primary sources.
(2) Connormah is incorrect. Ancestry's family trees and so on are assembled in any number of ways, including user input and mass importation of other sources, often involving an OCR step of uncertain quality. They are useful for the sort of research that leads to the creation of what will eventually be reliable sources once published, but they are not themselves reliable sources for WP. Even if a tree entry cites a reliable source, on Ancestry you're not seeing that source -- you're seeing someone transcribing and interpreting that source -- and there's nothing to indicate that the "someone" is reliable.
(3) If Ancestry acts as a web gateway to e.g. the NYT archive database, as part of what you're paying for, then the NYT archive is the source, not Ancestry, and should be evaluated on its own.
(4) Ancestry's OCR scans of otherwise reliable sources (e.g. old encyclopedias, DNBs) are of uncertain quality and are of highly variable accuracy. If you find something in such a scan (or an index generated from such a scan), and consultation of the original source shows the scan was correct, then fine (and in this case your source is the original source, not Ancestry).
Again, the above are just my opinons, albeit based on extensive experience. I'd be interested to hear others' comments.
By the way, I hasten to add that I have found Ancestry to be absolutely invaluable and they do an excellent job of keeping their data quality high. But that doesn't make them a reliable, secondary source for WP purpose.
EEng (talk) 23:28, 7 August 2012 (UTC)