Extracting reliable reaction kinetics from nonisothermal calorimetric results can be difficult. The reaction model, activation energy, and frequency factor make up a "kinetic triplet" for a particular reaction and define the reaction kinetics. One expects a good correlation between data and the predictions of the reaction model for a variety of experiments, provided the reaction tripler has been well determined. Such a correlation is expected for the results of accelerating rate calorimeter (ARC) and differential scanning calorimeter (DSC) experiments. As an example, the reaction of LixCoO2 in nonaqueous electrolyte (as is important in Li-ion battery safety) has been studied with both DSC and ARC. Comparing the shape of ARC profiles to those predicted theoretically limits the choice of reaction model. The activation energy is determined from the shift of the DSC profile with heating rate or from the change in the initial self-heating rate of ARC samples as a function of temperature. The frequency factor is then chosen to give the correct DSC peak temperature and correct self-heating rate. Calculated DSC and ARC curves fit experiment well for several related reaction models.