Modulated DSC, Dynamic DSC and steady-state DSC are three frequently used methods to determine the heat capacity. This paper investigates how the C-p-values obtained by these methods are influenced by the different thermal conductivity of the material used for calibration and as sample. A simple mathematical model of the DSC-instrument is derived and the obtained C-p-values are compared for different operating conditions of the instrument and for different samples. Experiments with different samples are also performed and compared with the simulation results. The results clearly show that the temperature distribution in the sample puts a higher limit on the useful frequencies. Recommendations for operating conditions for accurate C-p measurements are given.