I'm doing an experiment where I'm imaging 3 groups of 5 rabbits (total = 15 rabbits) using 4 different machines at 4 different anatomical locations to see if a radio-opaque material is visible in all of these anatomic areas across multiple imaging machines. The outcome we record is basically, visible or not visible (1 or 0).
Obviously we have 240 data points of 1 and 0s. I'm trying to see if there's any significant difference in visibility across the three groups. I'm trying to take an approach where I count up the number of visible cases in each of the 3 groups and then do a chi-square test / mcnemar test or something similar to that.
What assumptions do I need to make this approach valid? Or is it just impossible to count up like that and perform chi-square/mcnemar test?