User:Doncram/Arbitration-JustForFun
A: Here's an estimate: 294 articles
- In this diff, Dudemanfellabra provides information (thanks!) that in fact I created 784, about 18 percent of a now-reduced total (4386 articles). That's a bit higher than the 699 number that the sample statistics suggested was the high end of a 90 or 95 percent confidence interval for the true number. And two editors do have more than me. Change some numbers but everything else I wrote here stands, including that 18 percent is a lot lower than you'd expect, and that my working through 784 is not too large for me to address in a relatively short time. --doncram 22:39, 21 April 2016 (UTC)
The above is an answer to the "just for fun" question: How many articles, out of the current 4,412 "NRIS-only" NRHP stub articles, would you guess were created by Doncram? An estimated 294 articles. Honestly, isn't 294 articles, less than 7 percent of the total, a lot lower than you'd expect, if you are at all familiar with the 2012-2013 brouhaha?
And, based on the sample, 588 articles have been created by others since 2013 and survive that way now (recent creations of NRIS-only articles by other editors are twice the number of my old ones in the sample). And three editors have more than me in the sample (and I am sure that at least two of these have more than me in the entire population).
An expanded answer: 294 articles is the estimate, and with 90 or 95 percent probability the true number is in fact below 699. The estimate is a significant number, but is not too large for me to address in a relatively short time, which I would be inclined to do if I weren't banned from doing so.
Statistics of the estimate: The estimate is 294 articles are mine, or 6.67 percent of the 4,412 short "NRIS only" NRHP articles existing as of 4/13/2016. The estimate comes with an error margin of +/- 9.2 percent, or 405 articles. So with 90% certainty (or maybe 95%?) the number created by me is under 699. The estimate is based on my drawing a random sample of 30 numbers out of integers 1 to 4,412, and checking the history of the corresponding numbered articles. There were 2 created by me in the sample, 6.67 percent of the 30. A 90 percent confidence interval around the 6.67 percent fraction is +/- 9.2 percent. A sample size of 30 or more seems to be necessary for the statistical assumptions to be valid. I will happily email an Excel spreadsheet including all of the sample information, my calculations, and links to the two statistics webpages I relied upon, to anyone interested.
Why did I create the two articles found in the sample? I don't know why the other 28 short articles were created, but one of mine was created in 2009 to support the Peterson Farm disambiguation page. A Nebraska state historical nomination document has since become available so it could now be developed a bit. The other one was to support the Philip T. Shutze page about an architect, which was created by me in this version as of 21 September 2012, and now is linked from 18 mainspace articles. Creating this yielded a link to this MPS document which gives information that could/should have been used in the Shutze article: that Shutze was once regarded as "the nation's foremost living classical architect"; that he and his firm designed seven of the mansions on Atlanta's West Paces Ferry Road; that his landscape design for the "Swan House" was important to the overall success of the home's English Renaissance Revival architecture. And now an additional one page description and data sheet is available to develop the Thornton house article. I'd like to expand both.
Is 294 a significant fraction of the articles that I have created? Yes and no. I have created about 9,340 mainspace articles. 294 is 3.1 percent of that. For all or most of these, NRHP nomination documents were not then available on-line, but I knew they were notable topics and I felt there was good reason to create them (such as to head off contention about disambiguation, disambiguation needed for the NRHP list-system to work). For most of these, NRHP nomination documents are now available, and I'd like to expand them.
Is it lower just because others have expanded the articles? Yes and no. The number I created as "NRIS-only" stubs (for good reasons IMO) is larger than 294. I later expanded some myself, and others have arrived and expanded many others that I left that way. But also I arrived at and expanded many other "NRIS-only" stubs created by others, too. I am fairly positive that I expanded more belonging to others (usually by adding an NRHP nomination document source and a bit more) than the number of mine expanded first by others or still left unexpanded.
Other thoughts It's good all around that some of my stubs were created as stubs and then expanded by others, once sources became available. Some editors prefer to create new articles from scratch. Many others are happier to find an existing stub, especially ones already linked to corresponding architects' and other articles, and already having inbound connections too, like many of the ones I created. And/or having an NRHP nomination document source added by me. I have gotten "thanks" on my work many times when others were developing articles originally started by me. Maybe my topics got expanded, because they had been started, while non-started topics were too inconvenient for others to start plus expand. Maybe my stubs got expanded while others' stubs did not, because mine were better targeted (more needed, because their topics inter-connect more to other topics such as architects) and better started. It would require some better programming than I know how to do, to assess these comparisons, though. :) The bottom-line about "NRIS-only" stubs is that I too would like to see them developed, and I would be part of the solution to that if allowed.