The goal of this research is to develop a reliable method to derive nucleation rates and time lags from times of nucleation observed in sets of molecular dynamics simulations of supercooled clusters undergoing spontaneous freezing. What is taken explicitly into account, which was missing in prior analyses, is the kinetics of transient nucleation. Failure to consider this aspect of nucleation has led to substantially larger errors than the statistical errors associated with the stochastic Process. It is found that analyses simultaneously including the transient nucleation parameter as well as the nucleation time lag and the steady-state nucleation rate are too ill-conditioned to yield satisfactory results. A procedure to circumvent this problem and to estimate uncertainties is formulated.