The analysis of zero-inflated count data:
beyond zero-inflated Poisson regression

This webpage contains the supplementary material (data, R code & extra examples) for the paper
"The analysis of zero-inflated count data: beyond zero-inflated Poisson regression" (tutorial for the British Journal of Mathematical and Statistical Psychology).


Example: Modelling Unwanted Pursuit Behaviour

In this example, we use zero-inflated and hurdle Poisson models to investigate the impact of 'education level' and 'level of anxious attachment' on the number of unwanted pursuit behaviour (UPB) perpetrations in the context of couple separation trajectories. The data are part of the Interdisciplinary Project for the Optimization of Separation trajectories conducted in Flanders (IPOS; www.scheidingsonderzoek.be).

Data

Couple.txt

R-code

Analysis UPB.R


Artificial example 1

This example illustrates a common misinterpretation in the zero-inflated Poisson models concerning the effects of predictors on the excess zeroes. It is shown how hurdle models can make interpretation more straightforward by directly modelling the effects on all zero counts instead on the excess zero counts.

Example

Artificial example 1.pdf

R-code

Artificial example 1.R


Artificial example 2

This example illustrates the need for caution when interpreting interaction effects in the zero-inflated Poisson models and the hurdle models.

Example

Artificial example 2.pdf

R-code

Artificial example 2.R