To determine the heat transfer coefficient during dropwise condensation, two models are necessary: a heat transfer model through a single drop and a model of drop-size distribution.To model the distribution of the drop size, most studies dissociate the drop population into two distinct parts. A semiempirical model is then used to evaluate the drop-size distribution of "large" drops (i.e., typically greater than few micrometers), while the drop-size distribution of "small" drops is modeled using a statistical approach based on population balance. Currently, no accurate data are available to validate this latter distribution. In the present study, the statistical approach is compared to an individual-based numerical approach. This numerical approach takes into account the entire lifecycle of a drop: nucleation, growth, coalescence, and disappearance by sweeping of moving drops (jumping droplets are not considered in this paper). The drop-size distributions of large drops obtained thanks to this model are very similar to those obtained from the classical law in all configurations studied. Nevertheless, the distributions of "small" drops are notably different between the two types of modeling. In the configurations considered in the present study, an analysis of the main hypotheses used in the statistical approach (in particular, the assumption of a constant removal characteristic time τ irrespective of the range of drop size) revealed that the main mechanism for surface renewal is not the sweeping of the surface by moving drops. From these results, a modification of the statistical model is proposed and discussed.