What Chi Square test should I use?

Question

I am attempting to assess the ecological validity of an animal testing procedure as follows:

I have notated x number of events that occur in the wild. During each event, the speed of the animal is recorded in categorical terms (walk, run, sprint). eg. The animal walked 35 times, ran 70 times, and sprinted 3 times.

I then placed different animals in a lab and notated events again. Let's say this time the animal then walks 20 times, runs 30 times, and sprints 5 times.

How can I calculate if the distribution of running gaits within the lab fits the distribution of the running variables seen in the wild? I'm assuming I should use a chi-square test of some kind, but I'm not sure if I should use a chi-square goodness of fit test or a chi-square test of independence.

I am looking to undertake the statistical procedure using R if that changes anything.

Answer 1

If you did a chi-squared test the obvious one would be a test of homogeneity of proportions (which will be equivalent to testing independence), but if your alternative is a shift up or down your ordered categories, the ordinary chi-squared test would generally be somewhat less powerful than a test that focused on an ordered alternative.

While power is certainly a consideration, it may also be worth considering the expectations/familiarity of your intended audience (e.g. a test of homogeneity of proportions may be easier to get through a refereeing process and need less explanation).