An call evaluation device pairs an utterance of the first speaker and an utterance of the second speaker, which are associated with the same screen identification information, to be a question and answer, obtains a totalized value of the number of question and answers for each screen identification information and an utterance time of the second speaker for each question and answer, and generates display information to cause an output device to display the totalized value of the number of question and answers for each screen identification information and the utterance time of the second speaker for each question and answer.