Is Chi square the appropriate test in this scenario?

[P2324]

Hi everyone,

First time here with a (hopefully) basic question. I'm interested in studying Burrowing Owls, and one of the questions I'm looking at is whether occupied burrows tend to have a relatively higher number of satellite burrows within 10 m than what would be expected at nearby, unused sites (say, 100 m away). I would end up in my study having a total count of burrows within 10 m of X occupied burrows and a total count of burrows within 10 m of X random, unused sites (i.e., sample sizes in the two groups will be identical). Is it legit to use a chi-square test here? I see this being a question of whether something (i.e., satellite burrows) is distributed evenly between two categories. Am I on the right track?

Best,
Nick

 

[obh]

Hi P2325,

How do you distinguish between the main burrow and the satellite burrow?
Will the satellite be counted as a main burrow later? and the main as satellite?
Or do you compare the number of burrows in a cluster with at least one occupied burrow, to the number in a cluster with no occupied burrow?

When comparing averages I would first think of the t-test, (if meets the assumptions)
You can do a lot with the non-parametric chi-squared test. but I would start to check the t-test assumptions

 

[gianmarco]

Hello,
do you have spatial information about those burrows (i.e., geographic coordinates)?
Just wanted to provide you with another perspective: the question could be framed "spatially"; in other words, one could wonder if satellite burrows tend to be closer in space to occupied burrows than to random locations. Formal spatial tests exist, but you got to have info about the actual location of burrows.

Best
Gm