Talk:Contingency table

WikiProject Statistics (Rated Start-class, High-importance)

This article is within the scope of the WikiProject Statistics, a collaborative effort to improve the coverage of statistics on Wikipedia. If you would like to participate, please visit the project page or join the discussion.

Start  This article has been rated as Start-Class on the quality scale.
High  This article has been rated as High-importance on the importance scale.

Bivariate tables

Please add redirect from here from bivariate tables. ~~ —Preceding unsigned comment added by 139.135.197.201 (talk) 05:06, 8 May 2008 (UTC)

Why? Which direction? They seem entirely different ideas. Contigency tables always contain counts and there are specific statistc rtechniques relating to these, while a "bivariate tables" is much more general ... for example a usual form of presenting tables of the F-distribution is in the form of a bivariate table , with the two degrees of freedom used to enter the table. Thus some contigency tables (the 2D ones) may be special cases of bivariate tables, and vice versa. Without seeing what might go intobivariate tables even the question of links in one or both directions should be left open for now. Melcombe (talk) 09:16, 8 May 2008 (UTC)
(Note: I created a redirect from bivariate table before Melcombe's reply above, but couldn't leave a note then due to connection problems.) I don't see the harm in a redirect myself. For at least one meaning, bivariate tables are a special case of contigency tables, and providing the redirect will help the reader. Yes, it would be better to have an article on bivariate tables, or perhaps on statistical tables, that covers all the different meanings, but the redirect will be some use in the meantime. -- Avenue (talk) 18:00, 8 May 2008 (UTC)

Example is great

..but it would be nice to see a bit more, about expansion beyond two variables, and how the table relates to measures of association. Theblindsage (talk) 08:49, 26 November 2013 (UTC)

Does not describe testing for homogeneity

I looked here because I wanted to reference this page to show someone that you could use ${\displaystyle 2\times n}$ tables in testing whether proportion differed in a number of different trials (say we toss 10 coins and record the numbers of heads and tails for each, and we want to test whether the heads probabilities differ). But the presentation here does not hint that the table might be used for that -- it is solely concerned with the table as a summary of covariation of two variables. Felsenst (talk) 17:48, 24 October 2008 (UTC)

Quality of writing

Needs to be completely rewritten. Lots of random information and jargon tossed about. — Preceding unsigned comment added by 138.88.98.187 (talk) 02:40, 16 February 2012 (UTC)

The range of ${\displaystyle \phi }$

I think that ${\displaystyle \phi }$ varies from 0 to 1 (can not have negative values). Because both ${\displaystyle \chi ^{2}}$ and ${\displaystyle N}$ have positive values.

--CogitoErgoSum.AgeRemigo (talk) 04:47, 12 May 2012 (UTC)

The phi coefficient can have negative values. It cannot, however, be less than -1.Iss246 (talk) 15:27, 12 May 2012 (UTC)

Contingency table vs confusion matrix

What is the difference between a contingency table and a confusion matrix? Do both terms refer to the same thing? If so, perhaps a merge should be considered. pgr94 (talk) 11:35, 13 June 2013 (UTC)

I get the impression that a confusion matrix is a special kind of contingency table, where the same set of values (often just false and true) appears in columns and rows (rows for real values, columns for observed values), but I don't think they should be... confused. -- RFST (talk) 06:38, 28 March 2016 (UTC)