This is fair comment, but the original posting was also asking for an
estimate of the difference in magnitude between two groups, which seems
a reasonable question even when the shape of the distribution is
different in the two groups. You might get this using visual analogue
scales, for example, where the shape of the distribution depends on
where the mean is.
Martin
Allan Reese (Cefas) wrote:
> The range of advice is impressive in its erudition but hardly inspiring as an example of statistical methods applied to a real-world problem.
>
> Monica expresses that problem as, "I am looking for an appropriate test to compare several ordinal outcomes
> (scores with bounded range) for two groups of patients with very different distributions."
>
> Which prompts me to ask, "testing what?" The answers in the summary all appear to come from Mathsworld, that area of scholastic effort where everything is viewed as "a sum" with the implication that you "do the maths and get an answer."
>
> If the two groups have obviously "very different distributions" what is the point of any test, except to get a p-value to satisfy some innumerate editor or reviewer? Conversely, you can perform many tests that might show no significant difference; I quoted one such example on a graph where one experimental group did not get a star because of its large variance although its mean lay clearly with the others that got all three!
>
> I'm writing because the problem as posed begs the question that seems to dog much statistics teaching: the assumption that in any situation there is a "best test" and the job of the statistician is to pick a name from the firmament rather than look into the murky data for valid meaning.
>
> Allan
>
>
> ***********************************************************************************
> This email and any attachments are intended for the named recipient only. Its unauthorised use, distribution, disclosure, storage or copying is not permitted. If you have received it in error, please destroy all copies and notify the sender. In messages of a non-business nature, the views and opinions expressed are the author's own and do not necessarily reflect those of the organisation from which it is sent. All emails may be subject to monitoring.
> ***********************************************************************************
>
>
--
***************************************************
J. Martin Bland
Prof. of Health Statistics
Dept. of Health Sciences
Seebohm Rowntree Building Area 2
University of York
Heslington
York YO10 5DD
Email: [log in to unmask]
Phone: 01904 321334
Fax: 01904 321382
Web site: http://martinbland.co.uk/
***************************************************
|