Talk:Uncertainty coefficient

Was it introduced by Theil?
The article states "It was first introduced by Henri Theil" but there is no citation to the original source in Theil's published work, and no indication of just when it was introduced. In fact, the original source does not seem to be very easy to find so if it could be tracked down and properly documented this would be useful. The earliest reference I have found so far is from 1959 (not Theil, though). GH 143.234.212.144 (talk) 10:34, 19 October 2010 (UTC)

I first read about the measure in Numerical Recipes. It will have some earlier references. Maybe it can help us track down the source. Unfortunately I don't have the book here with me. Peteymills (talk) 04:32, 7 November 2010 (UTC)

The 2nd ed of Numerical Recipes mentions: Dunn, O.J., and Clark, V.A. 1974, Applied Statistics: Analysis of Variance and Regression (New York: Wiley). Norusis, M.J. 1982, SPSS Introductory Guide: Basic Statistics and Operations; and 1985, SPSSX Advanced Statistics Guide (New York: McGraw-Hill). Fano, R.M. 1961, Transmission of Information (New York: Wiley and MIT Press), Chapter 2. Of these, the Fano book might be worth checking as a possible source, maybe. GH 143.234.212.144 (talk) 12:28, 11 November 2010 (UTC)

Couldn't find anything relevant to this discussion in Fano's book. 143.234.212.144 (talk) 11:37, 6 January 2011 (UTC)

I think "Theil's U" Statistic should have its own article and not redirect here. Theil published two versions of the formula, references are here http://armstrong.wharton.upenn.edu/dictionary/definitions/theil's%20u.html --Viort (talk) 15:09, 14 October 2011 (UTC)

I've got a couple of points/rants/questions. Somebody has rated this article as a "1" for completeness. First of all, I can't stand the new page rating system. The whole idea of Wikipedia is that if somebody doesn't like a page, then they can change it. Their user-name, or at least their IP, is logged. With the page rating, there is no accountability. The low completeness rating as well as the "start class" ranking are silly. How long should this article be? It's just a single definition, large chunks of which are contained in other articles which I have referenced. Considering that it's also ranked "low importance"--"primarily of specialist interest"--surely it needn't be much longer. Peteymills (talk) 15:37, 28 March 2012 (UTC)


 * I'm starting to think that Theil is nothing to do with this. He's not mentioned in the Numerical Recipes citation. He did however introduce coefficients using the letter "U", but as Viort says above, those are very different and they're probably the same topic as Theil index? --mcld (talk) 07:51, 31 July 2012 (UTC)

Entropy with a minus sign
The definition of entropy should have a minus sign (see Entropy (information theory)). The conditional entropy does too, depending on how it's stated. I've corrected this in the article. The definition of U doesn't change since it's minus-signs all round. --mcld (talk) 14:15, 26 July 2012 (UTC)