Skip to main content

Table 3 Results of the LR tests among nested count models

From: Modeling count data for health care utilization: an empirical study of outpatient visits among Vietnamese older people

Pair model

Differences in LR

1% critical value

HNB1a vs. NB1

667.718***

χ2(34) = 56.1

HNB2a vs. NB2

486.544***

χ2(34) = 56.1

ZINB2a vs ZIP

4229.739***

χ2(1) = 6.6

HPMa vs. PRM

471.855***

χ2nn = 56.1

  1. LR denotes the likelihood ratio, HNB represents the hurdle negative binomial, NB indicates the negative binomial, ZINB means the zero-inflated negative binomial, ZIP denotes the zero-inflated Poisson, HPM represents the hurdle Poisson model, and PRM means the Poisson regression model
  2. aindicates the preferred model in pair comparison; and χ2(.) means chi-square test and the number in the bracket refers to degree of freedom of each model considered
  3. *p < 0.05, **p < 0.01, and ***p < 0.001