On 25 Jun 2003 23:50:17 -0700, [EMAIL PROTECTED] (Felix Bach) wrote:

> Hello,
> I would like to ask you if it is possible to do a repeated measurement
> anova (or a MANOVA) for the following situation: A person is answering
> 10 times to a  questiion. The answersa allow three different types of
> grammar types (either A or B or C). Is it possibble to count the
> number of used grammar types and then use a repeated measurement ANOVA
> or MANOVA to look which grammar is used most? A multinomial regression
> sounds more plausibe but I haven't got a module which accounts for

I assume that many people each gave 10 answers,
and the count is what is interesting.  Do you have groups
that you want to discriminate?  - that could be a 
discriminant function with 2 or 3 variables.  Two dummy
variables are enough, if the sum is always 10.

My guess is that you have groups.  However, you don't 
seem to have given a problem, so far as I can see.
You can say that one person used Type B (say) most
often, if that was most frequent out of 10.  And then?


-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
"Taxes are the price we pay for civilization."  Justice Holmes.
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to