The goal of this research is to construct conversational robots that can stimulate users’ motivation to talk with them in non-task-oriented dialogue, where it is required to keep up the dialogue. The non-task-oriented dialogue involves exchanging subjective opinions between speakers. This paper aims at investigating how the user’s dialogue motivation is influenced by the attribution of opinions to the conversational android. We examined the influence by testing various kinds of the android’s opinions in a questionnaire survey. As the result, it is clarified that not only the users’ interest in the android’s opinions but the attribution of the subjective opinions to the android influence their motivation for dialogue. This result suggests that there is a problem when the conversational robot makes the utterances based on human-human dialogue database that includes the opinions which are hardly attributed to it. In a design of conversational robot, it is necessary to take account of whether users can attribute the subjective opinions included in the dialogue contents to the robot in order to promote their motivation of dialogue.