On Fri, 19 Nov 1999 18:27:21 GMT, Jin Kim <[EMAIL PROTECTED]>
wrote:
> ...
> I have a question about the underlying statistical concept of Pearson
> chi^2 test.
> My question is:
>
> Does Pearson chi^2 test assume that both row and column margins are
> fixed?
>
> In other words, I wish to know whether Pearson chi^2 test is
> 'conditional' like Fisher's exact test.
(Here is an answer to confuse them.) - Theoretically, yes. In
practice, no.
That is, the Pearson chisquare, computed on the 2x2 table, has exactly
the same stated assumptions (of fixed marginal totals, both rows and
columns) as Fisher's exact test. But if you don't do Yates's
correction, then the chisquared value is quite a good estimate of what
you get when marginal totals *are* allowed to vary; Yates's
correction gives a better estimate of the Exact, fixed-margin
p-value, an improvement which is only noticable for small N.
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
