Talk:D'Agostino's K-squared test

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Statistics (Rated B-class, Low-importance)
WikiProject icon

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.

B-Class article B  This article has been rated as B-Class on the quality scale.
 Low  This article has been rated as Low-importance on the importance scale.


The early notation needs to be changed so that it distinguishes properly between μ for population values and m for sample values, similarly for σ. Also what version of the sample variance is to be used? (talk) 10:28, 10 September 2008 (UTC)

Test statistic?[edit]

I'm a little confused about the test statistic. I've applied this test to a vector of random numbers, generated by MathCAD as being normally distributed. Assuming my algebra is correct I reject the nul hypothesis if I compare K^2 against the Chi^2 distribution (95% significance with 2 degrees of freedom). This obviously troubles me. I fail to reject the nul hypothesis (quite convincingly, P value 0.77) if I compare K against the the Chi^2 distribution.

I have K^2=8.7 (K=2.95), Critical value(0.95,2)=5.991 - so it's an unlikely event.

Clarification is requested. (talk) 15:17, 30 October 2009 (UTC)

I just ran this test, for a sample of size n = 100 the test statistic K² has mean of 2.014 and standard deviation of 2.261 (versus theoretical 2.0 and 2.0) over B = 100,000 simulations. For a sample size n = 1000 i obtained mean = 1.995 and standard deviation = 2.023 (with 10,000 simulations). So you either made an error in coding, or got very unlucky.  … stpasha »  00:41, 20 December 2009 (UTC)
An online paper at may serve to validate your results. Melcombe (talk) 12:10, 21 December 2009 (UTC)