Primary nucleation is a stochastic process; hence, detection times are statistically distributed even though experiments are repeated at the same conditions. This contribution aims at defining and discussing a method to perform an accurate statistical analysis of detection times; it focuses on three main aspects. First, we develop an accurate experimental protocol and set criteria on temperature, T, and supersaturation, S, to accept measurements as part of the experimental series. Applying such protocol to the isonicotinamide–ethanol system at several different supersaturations, we perform multiple series of isothermal nucleation experiments, calculating for each series the associated empirical cumulative distribution function (eCDF) of the detection times. Second, exploiting a set of statistical tools, we investigate whether the repetitions of the experiments at the same supersaturation conditions belong to the same stochastic process, assessing also the possibility to combine them in one single distribution. Third, we estimate for the eCDFs the values of its associated nucleation rate and derive an analytical expression to calculate the propagation of the uncertainty from the eCDF to the nucleation rate estimate.