We introduce a method for the analysis of nucleation using mean first-passage time (MFPT) statistics obtained by molecular dynamics simulation. The method is based on the Becker-Döring model for the dynamics of a nucleation-mediated phase change and rigorously accounts for the system size dependence of first-passage statistics. It is thus suitable for the analysis of systems in which the separation between time scales for nucleation and growth is small, due to either a small free energy barrier or a large system size. The method is made computationally practical by an approximation of the first-passage time distribution based on its cumulant expansion. Using this approximation, the MFPT of the model can be fit to data from molecular dynamics simulation in order to estimate valuable kinetic parameters, including the free energy barrier, critical nucleus size, and monomer attachment pre-factor, as well as the steady-state rates of nucleation and growth. The method is demonstrated using a case study on nucleation of n-eicosane crystals from the melt. For this system, we found that the observed distribution of first-passage times do not follow an exponential distribution at short times, rendering it incompatible with the assumptions made by some other methods. Using our method, the observed distribution of first-passage times was accurately described, and reasonable estimates for the kinetic parameters and steady-state rates of nucleation and growth were obtained.