General Linear Models

#1
Hi,

I have an experiment where participants look at a picture and then have to choose a voice out of three for that picture. I want to find out how many participants get the right voice and whether participants' selection of the correct voice is significantly above chance.
From what I've read, I think General Linear Models is the way to go, but I am not sure. Has anyone had a similar problem before?
Also, could anyone with R experience give me a hand to code the analysis?

Many thanks,

Valentina
 
#3
Your problem does not need fitting a GLM. Infact I think you don't need a model at all.

There are three voices. Then the probability of choosing the correct voice by chance is 1/3.
Lets say you have Ntot participants and N of them schose the correct voice. The test will be whether proportion N/Ntot is significantly different from 1/3 (two-way test) or significantly greater than 1/3 (one-way test).