Enabled is to execute the evaluation of a speech efficiently while securing the validity of the evaluation result. From the recognition result data created by a voice recognizing unit (6) on the basis of the voice data of a speech, a selecting unit selects the recognition result data to be evaluated, from a data storage unit (7) with reference to an evaluation-noted point managing table (2) through an evaluation-noted point managing unit and in accordance with a selecting action made to correspond to the noted point information. An evaluating unit performs the display of the recognition result data and the reproduction of the voice data with reference to the evaluation-noted point managing table (2) through the evaluation-noted point managing unit and in accordance with the data outputting action made to correspond to the noted point information.