The Poisson model of the number of young raised by helmeted honeyeater pairs is very simple (Box 4.1). Its simplicity is embodied in the single parameter that needs to be estimated. The model based on a categorical distribution is more complex, requiring seven parameters to be estimated (Box 4.2). A key question is whether the extra parameters are justified. We will make this comparison more formally in the next section where the trade-off between model simplicity and model fit is considered.
Was this article helpful?