Hi,
I would like to have an opinion about Zero-inflated models used with presence-absence data. It's reports on the literature that Zero-inflated models are only used with count data (poison or negative bionomial distribution).
In my case, I have binomial data (presence-absence) with many absences.
Is there a way to use zero-inflated model ?
Thank
Hi Genevieve,
I understand you cannot use ZeroInfl. As you said, that is for count data. For presence/absence you use GLM binomial() family. It is a very simple function.
I am sending you a doc attached about ZeroInfl. I hope it helps.
Cheers :)
Hi Marcela Paz Espinaze
so you are suggesting a glm () with family=binary is enough? Because that was my initial model and I got feedback from co-authors saying I need to consider negative binomial distribution/zero inflation models.
Cheers
Saskia
Hi. If you have presence/absence data you have 1 and 0. Is that the type of data you have? In that case you have to use glm() binomial.
ZeroInfl is just for count data (with excess of zeros).
Cheers :)
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude