Wikipedia talk:Featured list criteria

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Lists (Rated Project-class)
WikiProject icon This page is within the scope of WikiProject Lists, an attempt to structure and organize all list pages on Wikipedia. If you wish to help, please visit the project page, where you can join the project and/or contribute to the discussion.
 Project  This page does not require a rating on the project's quality scale.
 

Completeness[edit]

Two points on completeness, to avoid violating WP:SYNTH:

  1. If a list is complete, it should
    • say it is complete, and
    • include a reference for its completeness
    case in point was Talk:List of Premier League hat-tricks#Multiple hat-tricks
  2. If a list is not complete, or uses inclusion criteria specific to Wikipedia, it should not include bogus statistics; e.g. List of awards and nominations received by Bradley Cooper "Bradley Cooper has won 21 awards from 70 nominations".

jnestorius(talk) 17:26, 30 May 2016 (UTC)

That's not a bad idea at all, as long as incomplete lists are not prejudiced against. How do we avoid that? The Rambling Man (talk) 17:31, 30 May 2016 (UTC)
Hi there, I'm the nominator of the Bradley Cooper list. I just saw the comment on the talk page, and I've reworded the opening line to omit SYNTH. When it comes to possibly incomplete lists, I think it may be a good idea to began with a generalized statement such as "Bradley Cooper has received several awards and nominations for his film and theatrical performances", instead of stating the exact number of nominations. Famous Hobo (talk) 01:39, 31 May 2016 (UTC)
Certainly, some introductory statement is necessary to set the context for the list. However, if it's too generic readers may be annoyed by the low information content. OTOH loading too much information might stray into WP:PEACOCK. The balance is tricky. jnestorius(talk) 11:10, 31 May 2016 (UTC)
As regards how to avoiding prejudice against incomplete lists: not including bogus statistics should be easy; it's not something you do, it's something you don't do. If anything my suggestions tilt the balance away from complete lists, because the additional reference requirement raises the bar for those. jnestorius(talk) 11:10, 31 May 2016 (UTC)
The only reservation I can think of at the moment is that most "complete list" sources are usually primary, e.g. from the Nobel Foundation or from AMPAS or from BAFTA etc. These lists are often out of date, sometimes months, yet other reliable sources, such as the BBC or The Daily Telegraph may report on new additions to such lists. Under your proposal, lists would need to be delisted until the official lists (if they are the only ones, of course) are updated. That seems a little radical. The Rambling Man (talk) 11:28, 31 May 2016 (UTC)
Well, in the examples you've listed, which are only updated once a year (or some other regular interval) I don't think it's a violation of synth to combine two references, the not-updated "complete" ref and the most-recent-addition ref, as it's reasonable to infer that nothing has fallen between the two or after the latter. If additions are irregular, in some cases it's plausible to find a ref that states "Foo is the first Bar since Baz", where Baz is the most recent entry in the "complete" ref. In principle such a most-recent-ref can't rule out that there has been yet another addition later than Foo; I suggest in such cases a bit of discretion is advisable before deciding that the most-recent-ref is too "stale" to prove completeness. For example, List of Major League Baseball perfect games is added to much less often than the aforementioned List of Premier League hat-tricks, so a month-old news story might suffice for the former but not the latter. (Analogy: a featured article about a living person cannot "prove" the person hasn't died.)
Alternative proposal: would it be acceptable to have a featured list that editors believe (but can't prove) to be complete, as long the list does not claim to be complete? I wouldn't favour that myself, but others may find it an acceptable compromise. jnestorius(talk) 15:13, 31 May 2016 (UTC)
Ok, it seems like you're trying to narrowly focus this on "lists that may (or may not) need periodic updates, as they can not be 100% proved to be "complete" (the person is still alive to maybe win more, there's no reason why another hat trick can't be scored, etc.). Though I will note that "Bradley Cooper has won 21 awards from 70 nominations" is technically true even if he has since had a 71st nomination, winning or not. Anyways.
By definition, a list that says "I am complete" is really saying "I am complete as of when this sentence was written". So the argument is really "can a list say just '78 super-home runs have been scored ever', or does it need to say '78 super-home runs have been scored ever as of May 2016' or '78 super-home runs are known to have been scored as of May 2016'. ('At least 78 super-home runs have been scored ever' manages to be both accurate and imply that it's not true even if it is, and I don't think it's a good option).
I personally don't see the problem as WP:SYNTH; the examples given are plain counting up what is given in the article, which is fine. The problem is a list falling out of date (a new super-home run is scored) and if it's acceptable for a list to present itself as up-to-date even though there will by definition be a gap between when the run is scored and the list is updated. It's not a big deal on popular lists, but it's certainly reasonable to expect that a lesser-updated list might take months or longer to get updated, and be confidently stating something that is not true in the meantime. --PresN 17:50, 31 May 2016 (UTC)
Quite. Do we have an endemic issue with inaccurate lists or lists that our readers have been misled by? Is this a solution looking for a problem or is it something that is a genuine concern? PresN makes a fundamental point, i.e. that all articles on Wikipedia will have a shelf life, and some lists will become inaccurate or outdated almost daily, while some will be fundamentally perfect for years. It would seem that this proposal now requires every single list to note "This list may be incomplete." Even those with a comprehensive list of entries sourced (not independently per se) and even worse, with a comprehensive list plus a few other articles that may seem to be complete. Perhaps that's even worse. The Rambling Man (talk) 20:59, 31 May 2016 (UTC)
Some comments
  • I would support updatable lists saying "complete as of <date>", and in many cases a single cite can confirm that. (Although the relevant date is that at which the source was published, rather than the later date it was checked; so "as of when this sentence was written" is not correct.) I strongly believe this is an important service to the many readers who might otherwise assume that because an article is instantly updatable it is therefore up-to-date, especially if it is marked as "featured". I concede this places a burden on article editors; but we are talking about featured articles.
  • I don't think a list needs explicitly to state "This list may be incomplete" unless there is a danger that readers might otherwise infer that it was complete. Certainly giving a specific number invites such an inference. Can someone who has won 22 awards truly be said to have won 21 awards? Mathematically yes, pragmatically no (see Maxim of Quantity).
  • I don't see any value in "plain counting up what is given in the article" unless the list purports to be complete; otherwise the number of items in the list is as meaningless as the number of bytes in the article. Conversely, if a count is given (as formerly in the Bradley Cooper list), it may give the misleading impression that the list does purport to be complete when in fact no such inference was intended.
  • This discussion is about the featured list criteria, not fixing "endemic" problems. The proposal doesn't apply to "every single list"; for starters it only applies to featured lists, and more specifically to the subset that are complete but not unchanging. If you want specific examples, there are a few problem cases in Low-importance Football FLs, where a few Wikipedians have just trawled the relevant statistics and checked each other's work. That would be fine as long as they don't claim their list is complete; to do so is obvious WP:OR.
jnestorius(talk) 22:14, 31 May 2016 (UTC)

Let me revise my proposal in light of the comments so far:

  1. If a list purports to be complete, it should
    • include a reference for its completeness, and
    • if it is liable to have future additions (or deletions) it should
      • give the date as of which it is complete (based on the publication date of the relevant references)
  2. If a list is not complete, or uses inclusion criteria specific to Wikipedia, its introduction should take care not to give the impression that it is complete (for example including a count of the number of items in the list might give such an impression)

jnestorius(talk) 22:14, 31 May 2016 (UTC)

Still problematic. For example, BAFTA Fellowship. The official organisation has a list, but others Fellows are sporadically created. These are usually reported in RS such as the BBC, but there is no guarantee that those being added plus the initial list is the superset. One or more Fellows may have gone missing on the way. Another example, List of current Premier League and English Football League managers, there is no reliable source I can see which lists all 92 managers. We all know there are 92 clubs in the top four tiers (the sky is blue) and we all know that each club has a manager, and that each manager is referenced. According to your proposal, this is insufficient. It appears that at a closer look, this is just a reiteration that Wikipedia is not perfect, and that as long as reviewers during the FLC process are satisfied that due diligence has been applied, no further action is needed. The Rambling Man (talk) 11:18, 1 June 2016 (UTC)
And just out of interest, why would we need more criteria since WP:WIAFL already speaks of comprehensiveness as: It comprehensively covers the defined scope, providing at least all of the major items and, where practical, a complete set of items. The Rambling Man (talk) 11:23, 1 June 2016 (UTC)
  • BAFTA Fellowship can just say "complete up to 5 July 2009". There is no guarantee that the later entries are complete and it is misleading to pretend otherwise. The statement "The most recent recipients were comedy writers Ray Galton and Alan Simpson" should be deleted. They are at the end of the list; reader can infer they are probably the most recent, which is as much as is justified.
  • List of current Premier League and English Football League managers
    • As regards completeness: we do need a ref that there are 92 clubs, but that's not exactly hard to find. Once you have a 92-item list and a ref that there are 92 clubs, WP:CALC says you're done. A fixed-size list is easier to complete than a constantly-increasing one.
      • No, you need a ref that says that the 92 clubs listed are current members of the Football League. You're using WP:SYNTH not WP:CALC. Just because there are 92 listed, that doesn't guarantee they're the right ones. The Rambling Man (talk) 15:02, 1 June 2016 (UTC)
        • In this particular case, it's still easy to find references ([1] [2] [3] [4] [5]). But more generally, this is the point: if you can't guarantee it then it should not be featured. jnestorius(talk) 13:47, 2 June 2016 (UTC)
    • As regards up-to-date: some of the 92 references work; e.g. Wenger from the arsenal website, so we can be confident the page will be updated as soon as he is gone. Others fail: e.g. Tisdale from a 2006 news story could and should be augmented with the club website to prove he's still there 10 years later.
      • No, we don't need to use primary sources. The Rambling Man (talk) 15:02, 1 June 2016 (UTC)
        • We need to use something more recent than a 2006 news story to provide adequate evidence for a 2016 manager. Do you mean "we don't have to use primary sources" or "we must not use primary sources"? In any case WP:PRIMARY says "A primary source may only be used on Wikipedia to make straightforward, descriptive statements of facts that can be verified by any educated person with access to the primary source but without further, specialized knowledge" which I think clearly applies in this case. jnestorius(talk) 13:47, 2 June 2016 (UTC)
  • "why would we need more criteria" -- the criteria are making explicit the pre-existing requirement for Wikipedia:Verifiability as it relates to the specific requirement of completeness. So it's not a new goal, but an explanation of how to achieve that goal.
  • The objections seem to be either that it's too hard to verify or that readers already know it might not be perfect; neither objection is appropriate for Featured content.
jnestorius(talk) 14:46, 1 June 2016 (UTC)
Frankly, adding that the list is complete to June 2012 when there are entries up to 2016 is actually going to create more confusion with the reader than not. Could you also wait for a consensus to emerge before you start to implement your own proposal? Thanks. And it's not "more criteria", your version is a variety of the criterion that already exists. You must see that? The Rambling Man (talk) 14:53, 1 June 2016 (UTC)
  • Fair enough about the complete-until change. I had no intention of charging off on an edit rampage; I felt a single edit to a single article would be illustrative rather than disruptive.
  • it's not "more criteria" -- you're quoting me, but I was quoting you. So we do agree that it's not more criteria, it's an elaboration of the existing criteria? I'm not sure whether you intend that to be a criticism or not? jnestorius(talk) 13:47, 2 June 2016 (UTC)
And for reference, Wikipedia is not perfect by design. You would need to start adding caveats to featured articles to let the reader know that some facts may have been omitted from the article. Or that some facts may be simply incorrect (because, much like Britannica, mistakes can be and often are made). Ramming this approach down on featured lists definitely needs a wider conversation before you start implementing it, and badly at that. The Rambling Man (talk) 15:04, 1 June 2016 (UTC)
I imagine that by "Wikipedia is not perfect by design" you mean, not "Wikipedia is intended to be imperfect", but rather "We should all know Wikipedia can never be perfect"; in which case, I agree. The WP:NOTPERFECT section of WP:EDITING is intended to reassure editors that any edit that makes a little improvement is fine. Again, I stress we are talking about Featured articles, not articles in general. The relevant quote from WP:EDITING is "Even the best articles should not be considered complete, as each new editor can offer new insights on how to enhance the content in it at any time." It doesn't say "Even the best articles may have mistakes". If an article leaves out something important, or lacks sufficient references, it should not have been featured in the first place. In a featured article, a statement may be incorrect if it is incorrect in the source or if the editor has misinterpreted the source; but the reader in principle can check the source and judge for themselves. If a statement is unsourced, the reader has to trust anonymous Wikipedians, and the article should not have been featured. jnestorius(talk) 13:47, 2 June 2016 (UTC)
So you are going to apply this to featured articles as well? Interesting. Perhaps start there and come back to the lists once you see how that flies. Anyway, I disagree that your change should be mandated, and in fact just to make the point, we can have {{dynamic list}} on every single featured list, just to cover your issue, that would cover your point nicely. Cheers. The Rambling Man (talk) 14:50, 2 June 2016 (UTC)
  • "So you are going to apply this to featured articles as well?" -- not sure what you mean by "this": could you elaborate? WP:VER is already part of WP:FACR, and a lot of WP:FAR is identifying unreferenced statements to either cite or delete.
  • {{dynamic list}} doesn't mean "entries are subject to change"; it's where the list has no cut-and-dry inclusion criterion (case #2 in my original proposal, whereas I think we have been disagreeing about case #1). Did you mean {{Complete list}} ?
jnestorius(talk) 16:27, 2 June 2016 (UTC)
You were the one to mention featured articles. And it's not about verification, you're talking about completeness. They are different. E.g. a FA may have a list of, say, international centuries, or tries, or the like. Are you now mandating (or attempting to mandate) that a single source covers those too? In which case you need to expand the discussion to FA as well. Oh, and I'm sorry, are you only worried about lists whose content remains static but elements of the list change in time? What about lists of alumni for example? They change, aren't covered by a single source, they aren't even accurate to a particular date. Sure, you could use {{complete list}} too, and that's fine under FL as long as inclusion criteria are defined (already one of the WP:WIAFL). These can be enhanced by adding further items as and when they are discovered. I don't think your proposal is adding going to add anything to FL, other than bloating lists with unnecessary references. For example, how do you date the "accurate" list of Football League managers? Presumably by the earliest of the individual citations for when a manager started? Isn't that what "accessdate" is all about? Plus I have yet to see anyone make any complaints about this, in particular where are the readers who are being misled moaning about it? Cheers. The Rambling Man (talk) 17:49, 2 June 2016 (UTC)

So, going back to the list of Football League managers, your proposed criteria suggests it must "include a reference for its completeness". Your suggested solution uses no fewer than five references to attempt to demonstrate part of its completeness, that's just the clubs that comprise the league. One which claims 92 clubs and then four which deal with each division. Now are you stating that we need to "date" the article to the earliest of those five references? After all, the leagues change (once a year, normally) and the four which deal with clubs in each division will change each year. But wait, not one ref covers all the managers either.... so we need a ref for the start date and a recent (primary?) source covering the fact each individual manager is still the manager? So instead of 92 refs, we need 184? And we need to "date" the list by the oldest? The Rambling Man (talk) 19:39, 2 June 2016 (UTC)

Another question, you said if you can't guarantee it then it should not be featured, where is that a requirement? No-one can guarantee anything on Wikipedia, we use reliable sources to verify our articles and lists. That's the best we can offer, no guarantees, nothing of the sort. Are you proposing we change Wikipedia to now offer "guaranteed" information? The Rambling Man (talk) 20:28, 2 June 2016 (UTC)

Further question, are you now seeking to delist featured lists that don't have your personal version of completeness? For instance, there's no one reference to cover List of 1930s jazz standards, or List of UK hit singles by footballers, or List of aircraft accidents and incidents resulting in at least 50 fatalities, or List of freshwater islands in Scotland, etc. The community and those hundreds of editors with lists that don't match your new approach need to be informed that you are intending to seek demotion of vast swathes of featured lists (although you curiously have avoided mentioning the featured articles which also include tables of information which don't match your new approach). The Rambling Man (talk) 20:58, 2 June 2016 (UTC)

Gee, we've come a long way from "That's not a bad idea at all". Maybe there are multiple independent subproposals that could usefully be discussed separately, ringfencing more controversial ones. I'm going to leave this a while as it's in danger of becoming a two-person shouting-match. Hopefully someone else might comment in the meantime. I'll check back in a week or so and respond then. Regards, jnestorius(talk) 17:10, 3 June 2016 (UTC)
Indeed, that's what happens when strawman arguments fall flat and the proposal seems to be far weaker and actually beyond reasonable implementation, on further inspection. C'est la vie. The Rambling Man (talk) 18:38, 3 June 2016 (UTC)
If you decide to re-ignite the discussion, it would be helpful for you to answer the various questions as a starter for ten. In particular I'd be interested to check that you're applying this "completeness" criterion to Featured Articles as well, because that's equally important, you have even mentioned that in one of your responses. I'd also like you to show me an example of (say) List of current English Football League managers with referencing that you would be satisfied with, i.e. with the 184 citations for "start" and "current" and the various citations to prove the 92 clubs are there and they are all correct, and then a logical "best data as of..." date alongside. That way we can then gauge the scope of change you would be requiring existing lists to undertake. Cheers. The Rambling Man (talk) 20:50, 3 June 2016 (UTC)

Jnestorius You didn't resolve my final queries. It's been more than a couple of weeks, are we done here? The Rambling Man (talk) 20:12, 19 June 2016 (UTC)

Completeness is a property of lists, not articles. I can't imagine in what sense a non-list article could be "complete". Lists may not claim to be complete, in which case obviously there is no need for a reference for completeness.
  • List of 1930s jazz standards specifically does not claim to be complete.
  • The section List of freshwater islands in Scotland#Larger islands claims to be complete but there is no reference for its completeness, and the reference for islands' areas is "Estimates based on Ordnance Survey maps and General Register Office for Scotland statistics unless otherwise stated"; the General Register Office reference only covers inhabited islands so it seems to me the author simply scanned OS maps for likely looking islands and estimated their area, which would obviously be OR. Maybe they did something more sophisticated, but if so they should specify what. The other tables do not claim to be complete; the linked OS gridref looks adequate for each row assuming the named islands are easily found from there.
  • List of aircraft accidents and incidents resulting in at least 50 fatalities has long sections of statistics prose, which are not referenced: are these generated from the table of entries? If so its completeness is vital to ensure the integrity of the statistics. Maybe completeness up to 2006 can be got from the 2006 Gero book, but the annual lists for 2007-2016 from https://aviation-safety.net/database/ only include airliners and may miss out some cargo or military. So it seems again we must trust the diligence of the authors' OR.
  • The best fix for List of UK hit singles by footballers is simply to delete the claim that it "contains every single recorded by a professional football team or individual player which spent at least one week in the UK top 75", as there is no way to verify that without trawling the entirety of the Guinness World Records: British Hit Singles and Albums and cross-referencing artists' names with football squad lists since 1955.
  • I'm not sure a permanently dynamic list like List of current Premier League and English Football League managers can ever be reliable enough to be featured. Wikipedia is not a directory. Editors could create List of 2015-16 Premier League and English Football League managers, which would be finished at the end of the season and could then be promoted. At a minimum, permanently dynamic lists need stricter criteria than other lists because the potential for staleness is so much greater.
The general principle is that Wikipedia can never be the first to publish any piece of information. We can aspire to be second by citing the first, but being first will always be OR. jnestorius(talk) 22:31, 19 June 2016 (UTC)
Then per your first response, we simply add a template to every single list in Wikipedia which states that the list may not be complete. Bingo. We should add a similar caveat to every featured article, as there's no guarantee that every featured article completely covers the topic it is addressing. And we should add the caveat to every article on Wikipedia because, as you know, Wikipedia is a work in progress. The Rambling Man (talk) 04:40, 20 June 2016 (UTC)
Yes check.svg already done In the bottom margin of every list (featured or not), every article (featured or not), indeed every page (including this one) there is a link "Disclaimers".
We have never claimed completeness except in a few specific instances; therefore, in the absence of any explicit statement that a list is complete, the implication is that it may be incomplete. --Redrose64 (talk) 08:56, 20 June 2016 (UTC)
@Redrose64: I partly agree; do you agree that List of UK hit singles by footballers and List of freshwater islands in Scotland#Larger islands do in fact contain an "explicit statement that a list is complete"? If so, do you think that claim needs to be backed up or taken on trust? I partly disagree: List of aircraft accidents and incidents resulting in at least 50 fatalities does not contain an explicit claim to be complete, but does contain the sentence "There have been 529 such incidents, including terrorist or other attacks. 195 of these occurrences have involved at least 100 fatalities, 33 have had at least 200 fatalities, 8 have had at least 300 fatalities, and 4 incidents have had over 500 fatalities" which has no explicit source, and seems to imply that the table is the source and ipso facto complete. jnestorius(talk) 10:47, 20 June 2016 (UTC)
You could add "at least" to the lead if you like. I'm sure no-one would really object. The Rambling Man (talk) 10:49, 20 June 2016 (UTC)
Done, thanks. Any suggestion for the islands? 10:57, 20 June 2016 (UTC)