Talk:Reduced chi-squared statistic

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Expert-subject template[edit]

I added the expert-subject template because I think the article needs considerable attention regarding a proper statement of the model, what "variance" is being estimated, and the different assumptions made at various stages in the article, as these seem to change ... the formulae quoted can't all be true under the same assumptions. Melcombe (talk) 17:18, 30 January 2009 (UTC)[reply]

The mean square weighted deviation (MSWD) is widely used in the field of isotopic dating. Wendt and Carl (1991) provide an "expert" publication on this statistic. Note however that MSWD is directly equivalent to reduced chi squared statistic. Wendt, I., and Carl, C., 1991,The statistical distribution of the mean squared weighted deviation, Chemical Geology, 275-285. Mark Brandon, Yale University

Massive cancellation[edit]

Some of the formulas in this article (those involving sum(x)^2 - sum(x^2)) are potentially vulnerable to massive cancellation and applying them to floating point numbers may be ill-advised. — Preceding unsigned comment added by 18.93.4.100 (talk) 16:54, 10 January 2014 (UTC)[reply]