Thanks everyone.
I have tried 5 models based on your suggestions:
Poisson:
Poisson with a log link
Hurdle Poisson to accommodate for the large amount of observation at and close to zero:
Negbinomial:
Zero inflated Negbinomial to account for some of the zeros in our data being apes that did not approach (N = 6):
It seems like the hurdle Poisson fits the data best?
Ultimately, I want to use expected predicted values to see if the time the apes stay in the zone differs across conditions (using emmeans contrasts).
For example like this:
Thanks for your help! I’m relatively new to choosing very specific distributions for my data, so I appreciate others sharing their experience!