Is Chi-square appropriate for a single measurement variable with repeated measures?

#1
Hi!

I am looking at how many burrows are occupied by a skink species at eight sites across a property. Counts of the number of occupied burrows have been taken over three years and I want to know if the change in the number of occupied burrows at a site is significant over the three years ie. is there a decreasing or increasing trend in occupancy at a particular site.

I am having difficulty working out which test to use. Given the data is categorical it seems a chi-square test would be best but the same burrows are sampled each year so the samples are not independent. I have also looked at a nested ANOVA but it seems that this would tell me if there is significant differences between the sites not changes within the sites.

Any help would be greatly appreciated!
 

Karabiner

TS Contributor
#2
Re: Is Chi-square appropriate for a single measurement variable with repeated measure

it seems a chi-square test would be best but the same burrows are sampled each year so the samples are not independent.
So you have n burrows, each measured 3 times, with a yes/no outcome? If you pool the data from all sites (or, if you analyse each site separately), then you can use Cochran's Q as global test, and McNemars test for pairwise comparisons.

With kind regards

Karabiner