I am currently trying to run a zero inflated mixed effects model in R using the package glmmTMB following a significant test of zero-inflation (using the function testZeroInflation() in the package DHARMa). The summary output gives p-values pertaining to each of the variables assigned to the conditional and zero-inflation models however I’m not sure if this is the correct method of determining significance. I have also tried the Anova.glmmTMB function and tried sequential removal of variables from the model and model comparison (via the anova() function) however these methods only give a single p-value (I’m presuming looking at the significance of the variable overall). Is the summary output sufficient to determine significance or do further tests need to be completed? If the latter is the case how would I do this? Apologies for the long winded explanation as I’m relatively new to R.

  • 1
    $\begingroup$ I recommend using bayesian brms package. Google for "brms zero_inflated_binomial" examples. You can compare different models with loo() function, thus you can choose, which variables to include or exclude. Also, I recommend to split data first and do two analyses: logistic regression for zeros and negative binomial for positive values. This allows you to test which variables should predict zeros and which high values. $\endgroup$
    – st4co4
    Aug 7 '20 at 15:21
  • 3
    $\begingroup$ What are you seeking the significance of ? What is your research question ? Are you interested in inference or prediction ? $\endgroup$ Aug 7 '20 at 16:21
  • $\begingroup$ @RobertLong I am looking to determine whether or not insects change their behaviour in response to both the magnitude of and direction of change in weather parameters e.g. temperature over a period directly prior to completing said behaviour. This is to help me understand some of the abiotic cues insects use to direct their behaviour. $\endgroup$ Aug 21 '20 at 21:38
  • $\begingroup$ Please add new information as an edit to the post! Not everybody reads comments ... $\endgroup$ Nov 15 '20 at 11:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.