This article describes the development of a method for determining phase change kinetics for multistep diffusion limited solid-state transformations from dilatation data. Since each step in a multistep reaction proceeds at a different rate, and the volume changes for the transformations are, in general, not equal, determination of the reaction kinetics from the dilatation data is not straightforward. Thus, a model is developed for the phase change process in which the transient dilatation is calculated based on the fractional extent of the various phases present. In this way, kinetic parameters are determined that allow the best match to the experimental data. However, both random and systematic experimental errors make reproduction of the experimental dilatation difficult. Therefore, a selfcalibration process is developed that uses portions of the dilatation data to obtain the density variation of the various phases with temperature to help correct for experimental uncertainties. This procedure also enables the model to be used in situations where accurate property data are not available. The model and procedures are applied to the formation of austenite in a pearlite/ferrite low carbon steel where the pearlite and ferrite regions transform at different rates. A single kinetic parameter set allows reproduction of transformation transients of significantly different heating rates. These parameters can then be used to describe the austenitization for any time-temperature path. Excellent agreement between the model and experimental data is shown.