I ran a Poisson mixed-effect model (glmmTMB) to examine the eye fixations on words while reading. Based on my study design, I built a Poisson model and got the residual plots as beneath. "CONDITION" is a between-subject categorical variable with 3 types, and "fixation_count" is a count variable. It seems that my model has no issue with dispersion but the KS test and outlier tests were significant.
Q1. May I know how to understand the significant KS test and outlier tests? How to identify the outliers and remove them? Q2. May I know how to understand the residual vs. predicted plot? Is there something wrong with the plot? Q3. How to understand the red vertical lines in the DHARMa plot and the outlier test plot? Should the line be at the center of the histogram?
model<- glmmTMB(fixation_count~ CONDITION+ (1|PARTICIPANT) +
(1| WORD) + offset(log(TRIAL_TIME_1)),
data = my_data, ziformula=~1, family = poisson)
summary (model)
res <- simulateResiduals(model, plot = TRUE)
testOverdispersion (model)
testResiduals(res)
