Hi Philips
I would start by drawing your attention to the fact that only a limited
reliance can be placed on the results of Pearson’s chi-square test when
sample sizes are small (less than 50). In such situations, I would
reccomend the Fisher's exact test for testing 2 proportions. For higher
order proportions, 3 and above as with your case, and extension of the
Fisher's exact test, attributable to Freeman and Halton (1951) has been
proposed. The Freeman Halton test is not that well known (and probably
better known as the Fisher-Freeman-Halton test) however it is implemented
in software [SPSS] as an exact test. Check SPSS for details and contact me
if you have any difficulties performing the test. Hope this is helpful
best wishes
Alianu
On Thu, Feb 2, 2012 at 11:32 AM, Philippa Hinks
<[log in to unmask]>wrote:
> Hi I have been asked to do a test to see if there is a difference in the
> proportions with the incidence of a particular infection between three
> different categorised groups.
>
> I am aware of the two-proportion z test and chisquared test but understand
> that they cover the comparison of two proportions, but I have three
> categories that I want to compare the proportions of.
>
> Is anyone aware of a test that will accommodate this?
>
>
>
> You may leave the list at any time by sending the command
>
> SIGNOFF allstat
>
> to [log in to unmask], leaving the subject line blank.
>
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|