Main Article Content
This paper describes the R package cold for the analysis of count longitudinal data. In this package marginal and random effects models are considered. In both cases estimation is via maximization of the exact likelihood and serial dependence among observations is assumed to be of Markovian type and referred as the integer-valued autoregressive of order one process. For random effects models adaptive Gaussian quadrature and Monte Carlo methods are used to compute integrals whose dimension depends on the structure of random effects. cold is written partly in R language, partly in Fortran 77, interfaced through R and is built following the S4 formulation of R methods.