Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Re: RE: What does contribution to chi-square mean in tabchii


From   "Donald Spady" <dspady@ualberta.ca>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Re: RE: What does contribution to chi-square mean in tabchii
Date   Mon, 1 Dec 2003 09:53:35 -0700

Thanks for the response.  I will use it as a way to get insight into the
data, but for little else.

Don
----- Original Message ----- 
From: "Nick Cox" <n.j.cox@durham.ac.uk>
To: <statalist@hsphsun2.harvard.edu>
Sent: Saturday, November 29, 2003 4:35 AM
Subject: st: RE: What does contribution to chi-square mean in tabchii


> -tabchii- is a user-written command. A version
> which will work in Stata 6 through 8 is
> available from the -tab_chi- package on SSC
> and an earlier version which will work in
> Stata 5 is available from
> http://www.stata.com/users/njc/tab_chi/
>
> I am not sure what kind of answer Don
> wants here. The definition of contribution
> given in the help file is
>
> (observed frequency - expected frequency)^2
> / expected frequency
>
> and Don clearly understands that. The
> programmer's intent in allowing
> such contributions to be shown was that
> some users like to see that step, perhaps
> even as a check on hand calculations.
> Personally I much prefer to look at
> the (signed) square root of that
>
> (observed freq - expected freq)
> / sqrt(expected frequency)
>
> given the usual argument that the direction of
> deviation from expectation is often helpful
> in interpretation. These quantities are called Pearson
> residuals in -tabchii- and indeed
> elsewhere. And in turn a refinement
> on that, the adjusted residual,
> has a distribution in the null case
> that is closer to Gaussian(0,1).
>
> Nick
> n.j.cox@durham.ac.uk
>
> Donald Spady
>
> > Can you please tell me how to interpret the chi-square
> > 'contribution' to
> > chi-sqare as used in tabchii.  I find it useful in telling
> > me where the
> > total chi-square is coming from but is there more that one
> > can get out of
> > it? i.e. say the chi-square (total) is 100 and the
> > contribution of one cell
> > is 55, another is 35, another 5 and another 5.  Obviously
> > the first cell's
> > chi square counts a lot but can I say anything else, say in
> > regard to cell
> > 2's chi-square of 35.  Or is it just an interesting way of
> > getting an
> > insight into the data.
>
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index