|
List members,
I have a question that is puzzling me. I have a project that will be
investigating categories of self-talk used by subjects, Participants will
have their self-talk recorded during a task, and after the task they will
code their talk into 1 of 4 categories. Given that some people will talk a
lot, and other not so much, we were going to use the % of the total number
of statements in each category as DVs. For example, one subject may have
50% of their statements categorized as CAT1, 25% as CAT2, 5% as CAT3, and
20% as CAT4.
My questions relates to using these percentages as DVs in a MANOVA
(we are comparing different types of people and different tasks on the
types of self-talk used) . Since the total of the four percentages for each
subject add up to 100, will the analysis crash? If so, what are some of the
alteratives to handling the DVs?
Any advice would be appreciated.
Thanks.
Allen Cornelius
|