Wikipedia:Reference desk/Archives/Mathematics/2023 February 5

= February 5 =

what is a cantellated great icosahedron?
Ya got yer
 * great icosahedron (regular)
 * truncated great icosahedron
 * great icosidodecahedron (rectified)
 * truncated great stellated dodecahedron (degenerate)
 * great stellated dodecahedron (regular)

But where are the cantellate (great rhombicosidodecahedron is something else) and the omnitruncate? Not on the usual lists of uniform polyhedra. Are they also degenerate? —Tamfang (talk) 06:32, 5 February 2023 (UTC)
 * The omnitruncate degenerates for basically the same reason the bitruncate does: truncating a {p/q} with even denominator. It would be a rhombicosahedron with twelve additional {10/2} faces.
 * The cantellate has coincident vertices, and becomes a combination of the small ditrigonal icosidodecahedron with the compound of five cubes. In some webpages it is called the small complex (ditrigonal) rhombicosidodecahedron. So four faces meet at each edge and vertices coincide (if not for this degeneracy it would indeed be uniform). It is explained in Coxeter et al.'s 1954 paper on uniform polyhedra (p. 416): Figure 8 (cf. Coxeter 1948, p. 111, Fig. 6.7B) shows the partition of a trirectangular spherical triangle (2 2 2) into fifteen (2 3 5)’s. The centre of the (2 2 2), being on an axis of trigonal symmetry, is a vertex of the dodecahedron 3 Double sharp (talk) 19:07, 5 February 2023 (UTC)
 * Thanks. —Tamfang (talk) 00:24, 6 February 2023 (UTC)
 * P.S. Richard Klitzing's webpages list all the cases from Wythoff's construction, including the degenerate ones. Your specific case is covered under "o5/2o3o (µ=7)". :) Double sharp (talk) 19:12, 5 February 2023 (UTC)
 * Fun fact: I am Klitzing's webhost. —Tamfang (talk) 05:02, 9 February 2023 (UTC)

Wikipedia on trial?
Copied, as suggested, from WP:RDH

An article for 4 February in The Guardian links to a recent paper User-Generated Content Shapes Judicial Reasoning: Evidence From a Randomized Control Trial on Wikipedia. This is a final version of the paper reported in The Irish Times last year. The authors, with help from their law students, generated 77 (out of around 80) articles linked at List of Irish Supreme Court cases, to see if actual text from WP (rather than just following the links) materially influenced Irish High Court decisions. Apart from the (perhaps deliberate) abysmal reffing standards of the articles, is anyone here competent to analyse their statistical methods? The 6-month period after the release of the articles on WP seems quite short compared to the 28 months previous. I have no expertise in this area. The histories of the articles are quite revealing. MinorProphet (talk) 00:53, 5 February 2023 (UTC) The Supreme Court of Ireland decisions are freely available here. MinorProphet (talk) 01:05, 5 February 2023 (UTC)


 * This could be suitable for Village pump rather than the reference desk as no question is asked? Apokrif (talk) 09:45, 5 February 2023 (UTC)
 * There is a literal question: "is anyone here competent to analyse their statistical methods?" There is an obviously implied pragmatic question: "If so, are their methods appropriate?". (I think, though, that WP:RDMA (mathematics, geometry, probability, and statistics ) is a better forum for this topic than WP:RDH (history, politics, literature, religion, philosophy, law, finance, economics, art, and society).) --Lambiam 10:09, 5 February 2023 (UTC)
 * I tried to read the article but got stuck on their linear regression model introduced in Section 4.1. Quoting from the article:
 * 106%
 * 106%
 * 106%
 * So $$\textit{Wiki}_{\,ij}$$ and $$\textit{After}_{\,t}$$ are, apparently, real-valued variables, but how are their values defined? Or are these the parameters to be estimated? Then the model is not linear.
 * The relatively short duration of six months (or seven? I did not immediately spot the upload date(s) but various figures with a time scale, such as Figure 2, have seven ticks for the "post-treatment" period) should not be an issue, as long as the number of data points collected is high enough. More than 7,200 observed citations, as they state, should be ample for the purpose of hypothesis testing. It is somewhat ironic that our article Statistics is among their cited references. --Lambiam 14:51, 5 February 2023 (UTC)
 * Thank you as usual for your thoughtful replies. The suggested RefDesk is probably a better place for this question, and I will copy it over there: I am no angel, and as a mathematical/statistical fool I fear to rush in and tread the hallowed halls of WP:RDMA. The possibilities for deliberately introducing weasel words into apparently innocent articles to influence legal decisions seem manifold. MinorProphet (talk) 16:00, 5 February 2023 (UTC)
 * Published version in Information Systems Research is available here: https://pubsonline.informs.org/doi/abs/10.1287/isre.2023.0034 Pinkoh1 (talk) 11:48, 7 April 2024 (UTC)