Dear List,
I am using a cognitive testing paradigm whereby participants choose between
a pair of simultaneously presented stimuli over a series of trials. The 20
stimuli can be divided into 4 meaningful categories with 5 stimuli per
category. Over a fixed series of 150 such pair trials it is possible to
observe the number of selections in each category relative to the other
categories. When viewing the data descriptively, I can make intuitive
conclusions such as Pp#1 chooses category A>B>C>D or that Pp#1 chooses
category A more than Pp#2.
My problem is that the forced choice element of the task means that each
category total is dependent/relative to the other category totals (i.e. +1
choice in one category = -1 choice in the remaining 3 categories) and
therefore I am unable to test for differences using ANOVA.
Can anyone advise of an appropriate method that would enable analysis of
these data to test for differences between categories, between
administrations and between subjects?
Graham Finlayson
Division of Psychology
School of Life Sciences
Glasgow Caledonian University
|