sorry - stats question

R

Hi,
I am struggling for help.
All I know is - it's within subjects because all of the participants judged these four stimuli and I have a mean percentage correct response for all the participants to all the stimuli. In a paper I read which did a similar study their output looked like chi squared. But I am not sure if this is the right test?

For reaction time data on the same experiment I used a one-way within subjects anova...

is this right and what test should I use for the correct responses data? I want to know if one stimuli was easier to get right than another?

Help much appreciated....

C

I'd use a one way ANOVA.

percentage correct scores are interval/ratio data and therefore suitable for ANOVA.

You'd get them same answer (as to whether it is significant) with a CHI Square (though you'd have to use the raw scores and not percentages)but you can easily do post hoc tests using the ANOVA to find out which of the four condition differened from each other.

A

Just a quick caveat - make sure your data fulfill the requirements for a parametric test such as an ANOVA, non-parametric tests (e.g. Kruskal Wallis) may be more appropriate.

Your best bet is to consult a statistician - you should be able to do this within your Uni, most Maths/Stats faculties offer stats clinics where research staff and postgrad students can bring along their data for professional guidance on analysis. Good luck

7505