0
$\begingroup$

I am fitting a poisson GLMM of the type

glmer(captured ~ treatment + offset(log(Effort)) + (1|Trip_Code), data=x, family="poisson", glmerControl(optimizer = "bobyqa"))

where captured is a count and treatment is a categorical variable with two levels (control and experiment). The dependent variable contains mainly zeros (Fig 1).

I am running some diagnostic tests using the DHARMa package and I did not detect any singularity, overdispersion or zero inflation.

When I check the quantile residuals plots though, using two different methods a) and b) I obtain two different results.

#a) 
simulationOutput <- simulateResiduals(fittedModel = m1)
plot(simulationOutput)

#b) 
testQuantiles(simulationOutput)

With method a) no significant quantile deviation is detected (Fig 2, right plot), while it is detected with method b) (Fig 3). However, I was expecting a) and b) to give the same results since the R documentation for testQuantiles {DHARMa} explains:

the quantile test is automatically performed in

plot(simulationOutput)
plotResiduals(simulationOutput)

I understand that method a) plots rank transformed model predictions, while b) doesn't but I wouldn't think this affects the p-value, am I correct? I also understand that method a) automatically adds some noise to the residuals to maintain a uniform response (see: integerResponse parameter). Does method b) do the same? If it doesn't, could that be the reason for the different results?

Many thanks

$\endgroup$

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.