Help needed analysing the effect of communal breeding on predation events (binomial data in mixed models) ?

#1
The research question I am dealing with is as follows: Do bird nests with helper individuals (in addition to the parents) have a decreased risk of predation?

I have observation-data of bird nests (European bee-eaters), including if helpers were present (yes/no)/ if predation events happen (yes/no) / brood size (number of chicks). I made three columns to feed the data into SPSS:

1) Two groups: without helpers: 1 / with helpers: 2

2) The occurrence of a predation event: yes: 1 / no: 0

3) The brood size: a number between 2 and 7 chicks

How can I test if the occurrence of predation events is affected by the presence of helpers and include that the brood size has an effect too (which I am not particularly interested in)?

I just don't understand how to do it, because the predation event (binominal) and the brood size (skewed) are not normally distributed - so I cant use a mixed linear model?

Does someone know which strategy/test to use to tackle this in SPSS?

Thanks for the help.
Best,
Saskia
 

Karabiner

TS Contributor
#2
You can predict predation event yes/no by multiple binary logistc regression, with the two predictors
"helper" (categorical, 2 levels), and "brood size".

With kind regards

Karabiner
 

Karabiner

TS Contributor
#4
Maybe it could be sensible to also include the interaction between brood size and helper
(if the amount of helper effect depends on brood size, or vice versa the effect of brood size
depends on whether a helper is present or not).

With kind regards

Karabiner
 
Last edited:
#5
Maybe it could be sensible to also include the interaction between brood size and helper
(if the amount of helper effect depends on brood size, or vice versa the effect of brood size
depends on whether a helper is present or not).

With kind regards

Karabiner
Yes, that sounds good! However, I am not sure how to do that.