1
$\begingroup$

I am trying to complete a binomial GLMM but I don't think that the residuals look right when I compare them to the years. Is this too much of a pattern in the residuals?

My model looks like this: binary response ~ carbon + I(carbon^3) + nitrogen + class + (1|Year). In my model, carbon and nitrogen are scaled isotope values, and class is the age class of the animal. The years are between 2003 to 2011. I have tried adding interactions, transforming the isotope values, and fitting it as a glm. All of these haven't improved the fit. I am using glmer in R. How can I go about improving the model? Or do I need to?

output from plot(model)Year vs Model residuals

Improve this question
$\endgroup$
1

1 Answer 1

Reset to default
1
$\begingroup$

Welcome to the site Nicole! You don't need to change anything given what I can see in those residuals, although it doesn't preclude the presence of extra-binomial variation. Whenever you are using a generalised linear model that does not have normal errors (i.e. you fit a poisson, binomial, etc. family of errors) you are looking for different things in the residuals, which are often unique to the distribution in question. If you fit a binomial family of errors then it is quite normal to see a residual plot like yours, and on its own does not suggest you need to change your fit at all. If you follow @Florian Hartig advice above then you can perform further checks as to whether or not your model needs to be adjusted. Feel free to post the results here if you have anymore questions.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.