Metallurgical processes of steel production are strongly connected with knowledge of the thermo-physical and thermodynamic properties of liquid metal during individual technological steps. Critical factors describing behaviour of steel in the process of its solidification and then important to a correct setting of conditions of steel casting and solidification of steel, which significantly affect the final quality of the as-cast steel, are liquidus (T-L) and solidus (T-S)temperatures.Experimental determination of these temperatures (or generally additional phase transformation temperatures) can lead to the refinement of all these phase transformation temperatures as well as to a more precise determination of the interval of steels' solidification. Utilization of this approach, e.g. in the process of optimizing the technology of ingot or continuous casting, could lead to achieve more accurate results moreover in direct connection with numerical simulations, and thus a better knowledge of the impact of optimized technological parameters on the resulting quality of cast steel.The paper is devoted to the application of method of Design of Experiment (DOE) to analyse and clear quantification of the impact of experimental conditions on the experimental results (measured data) of direct thermal analysis method (DirTA). First of all, it is a way to find a new approach to quantify and verify a statistical significance of experimental effects (influence of experimental conditions) on obtained phase transformation temperatures. The developed method allows filtering relevant data from individual thermo-analytical experiments for final calculation of high temperature phase temperatures more precisely than it was possible ever before. This is crucial for proper liquidus and solidus and liquidus temperatures' recommendation to real conditions.