Wikipedia talk:Criteria for speedy deletion/Analysis

This is the page to discuss analyzing various aspects of the speedy deletion process.

single-administrator deletions
What I suggest doing is taking a sufficient sample from the deletion log to get 200 single-admin. deletions of articles (images etc are another project) on the basis of speedy criteria other than purely administrative. Id do this by starting with perhaps 100 items at different times of a day--say tomorrow, Friday July 6, copying them to a spreadsheet, sorting out the ones by different criteria, and copying the appropriate ones to a subpage. This count alone would be useful.

As a second step, two or three eds. should check to see if a/the reasons given were valid speedy reasons and b/if it made even approximate sense. I could certainly do that first sort tomorrow. and have it ready for Saturday. If it proves too cumbersome to do as a list or table here, we could use Google spreadsheets. DGG 03:51, 6 July 2007 (UTC)


 * That means admins need to volunteer; mortals such as myself cannot see the history (to see if the deleted article was tagged) or the previous contents (to see if they made sense). &mdash; Coren (talk) 03:54, 6 July 2007 (UTC)
 * I'm checking to see how much is visible and you seem to be right. Here's a random example:

"01:39, July 6, 2007 Jimfbleak (Talk | contribs | block) deleted "Bill Whatcott" (unsourced nn content was: 'Bill Whatcott is a prominent Canadian social conservative and executive director of the Christian Truth Activists and an anti-gay and anti-abortion ac...' (and the only contributor was '[[Special:Contributions/Superfleet|Superfl) (Restore)" Not until I click restore do I see the page history which says: " 01:14, July 6, 2007 . . Superfleet (Talk | contribs | block) (2,332 bytes) 01:04, July 6, 2007 . . Superfleet (Talk | contribs | block) (2,276 bytes) (←Created page with 'Bill Whatcott is a prominent Canadian social conservative and executive director of the Christian Truth Activists and an anti-gay and anti-abortion activist.(Archie...')" This indicates that the article was not previously tagged.

However, enough should have been visible to see that the reason given: "unsourced nn content" is not a correct exact reason for Speedy, and interpreting it as A7, then in fact the article did assert notability in the few words included in the summary. Upon displaying the entire article, it is furthermore clear that there are 8 newspaper sources, including several national newspapers. There are also some BLP concerns, but the article is sourced. Interesting that this should be the first one I tried. I jafd no reason to select this particular admin.

So maybe we should instead simply look for improper deletion rationales on their face, and then see if they are single-admin deletions. There will be one problem--this was an indicative deletion log summary--not all of them are. In particular, for attack pages we are required not to leave anything in the summary to indicate the problematic content. But some admins dont leave any at all in any case. DGG 06:59, 6 July 2007 (UTC)

DGG 06:56, 6 July 2007 (UTC)


 * Would one or more admins be able to do some genuinely random sampling, perhaps? Then record some proper, objective coded data: which CSD seems to have been use, previously flagged, maybe something else. Who did it for completeness, but that shouldn't be used in initial analysis. Then get a couple of opinions on each for the subjective question - was it justified? I suppose there's more questions there. If we can get some figures on this, I'm happy to do some basic analysis on it (the analysis being something I've some experience with). SamBC 12:58, 6 July 2007 (UTC)


 * Once upon a time, there was a proposal for an alternate form of deletion that made blank pages appear as redlinks and thus preserved edit history for deleted pages (except when Oversight'd, BLP'd, etc., of course). See Pure wiki deletion system.  Never quite understood why that proposal failed, to be honest, and it would certainly solve any (potential) problem of transparency in the CSD process.  Jouster  (  whisper  ) 15:16, 6 July 2007 (UTC)


 * I can certainly see lots of reasons to have issue with that proposal. I shan't get into them, however, as this relates to a different and seperate proposal, that is simpler and deals with a much narrower issue. SamBC 16:06, 6 July 2007 (UTC)


 * Meh, okay, I didn't really see any disadvantages in a quick glance-over, and given that it has failed, I see no real need to dig them up now. Thought I'd point you towards something that might be of use.  Toodles!  Jouster  (  whisper  ) 17:08, 6 July 2007 (UTC)


 * This analysis seems like a good idea to me. I will try to do a version of it in the next week or so -- it will take soem time to set up and do. DES (talk) 06:44, 7 July 2007 (UTC)

Any luck with the analysis? &mdash; Coren (talk) 05:46, 15 July 2007 (UTC)