The advent of experimental techniques for the time-course monitoring of gene expression at the single-cell level has paved the way to the model-based study of gene expression variability within- an across-cells. A number of approaches to the inference of models accounting for variability of gene expression over isogenic cell populations have been developed and applied to real-world scenarios. The development of a systematic approach for the validation of population models is however lagging behind, and accuracy of the models obtained is often assessed on a semi-empirical basis. In this paper we study the problem of validating models of gene network dynamics for cell populations, providing statistical tools for qualitative and quantitative model validation and comparison, and guidelines for their application and interpretation based on a real biological case study.