Calorimetry is a powerful tool for the study of chemical reactions and of biological systems. Nevertheless, reaction calorimeters and calorimeters may have high time constants. In this study, the correction brought by two types of deconvolution methods using electrical signals has been compared. The distortion of the calorimetric signals has been evaluated and its effect on kinetic parameters determination has been quantified. For this purpose, relative errors on activation energy, pre-exponential factors and kinetic exponents have been computed by reference to measured signals whose thermokinetic parameters were perfectly known. 1st and 2nd order deconvolutions were found to be more suitable for kinetic analyses. Integral methods allow obtaining the most accurate results with model-fitting and model-free kinetic methods, while differential methods are more sensitive to noise and submitted to numerical instability. The correction proposed allows slightly improving the accuracy in the estimation of kinetic parameters and to adapt the correction to the kind of kinetic method used. The deconvolution method has been applied to isothermal data, but the method is easily transposable to nonisothermal measurements. (C) 2015 Elsevier B.V. All rights reserved.