The true temperature of a sample during isothermal crystallization is higher than the recorded differential scanning calorimetry program temperature because of two combined effects: the thermal resistance of the sample, which is mainly dependent on the thickness of the sample, and the release of the heat of crystallization. This is especially true for a low crystallization temperature, at which the heat of crystallization is released within a relatively short time. Isothermal crystallization is treated here as a nonisothermal process, and the values of the parameters obtained with single-mechanism equations (Avrami and Nakamura's) used in their description are compared. The validity of the procedure used for evaluating the temperature increase of a sample is analyzed and discussed (i.e., the evaluation of the thermal resistance of a sample as a function of its thickness). Some isothermal crystallization data obtained at relatively high supercoolings may be affected by errors if temperature corrections are not carried out. (C) 2003 Wiley Periodicals, Inc.