# Talk:Cook's distance

Jump to navigation Jump to search
WikiProject Statistics (Rated Start-class, Low-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.
Low  This article has been rated as Low-importance on the importance scale.

## Correct?

I believe the second equation to be incorrect, atleast when compared to my textbook .. can any one confirm ? —Preceding unsigned comment added by Thedreamshaper (talkcontribs) 17:51, 4 July 2010 (UTC)

It looks ok to me. There are probably several algebraically equivalent expressions, so your book may give a different one without either being incorrect. Qwfp (talk) 11:01, 5 July 2010 (UTC)
The second Eq. agrees with a source I have. Check whether you have a "studentized residual" rather than the "raw residual" here. As well as there being several algebraically equivalent expressions there are also several different modified versions of "Cook's distance". Melcombe (talk) 14:36, 5 July 2010 (UTC)
I can confirm that the 2nd equation is equivalent to that presented in Cook's original ref. However, the latter is a simpler formula and possibly preferable here:
D = (ri^2 / p) * (hi / (1-hi))
Where:
hi is the leverage of the ith obs.
p is the rank of the model
ri is the internally studentized residual ( ri = ei / (sigmahat*sqrt(1-hi)) )
[where ei is the raw residual, and sigmahat is the estimated residual sd of the model) — Preceding unsigned comment added by 130.126.55.118 (talk) 20:50, 13 March 2012 (UTC)

## Not Correct?

It's the third equation that's incorrect: the denominator should be p s^2, not (1+p) s^2. Primrose61 (talk) 18:52, 11 November 2013 (UTC)

The denumerator in the second equation should be p * s^2, and in the terms of MSE, s^2 = MSE * n / (n - p), where n is the number of observations, and p is the number of parameters, so the whole equation should be:

$D_{i}={\frac {e_{i}^{2}\ (n-p)}{p\ n\ \mathrm {MSE} }}\left[{\frac {h_{ii}}{(1-h_{ii})^{2}}}\right],$ Keskival (talk) 12:30, 5 December 2014 (UTC)