A meta-analysis of effectiveness studies on dialogue-based CALL, based on 17 studies and $k = 100$ individual effect sizes. We use Morris and DeShon’s (2002) formulas to compute comparable effect sizes across designs, and analyse them through a multilevel random-effects model. Results confirm that dialogue-based CALL practice has a significant medium effect size on L2 proficiency development ($d = 0.59$). Extensive moderator analyses on several learning outcomes of different types and features of dialogue-based CALL (type of interaction, modality, agent embodiment, etc.) confirm the effectiveness of form-focused and goal-oriented systems, system-guided interactions, corrective feedback provision, and gamification features. Significant effects for lower proficiency learners, and on vocabulary, morphosyntax, holistic proficiency and accuracy are established.

Dialogue systems allow a user to interact, orally or in writing, with an automated interlocutor, whether it is referred to as a chatbot, a robot, a conversational agent or an intelligent personal assistant. We discuss the different typologies of dialogue-based computer-assisted language learning (CALL), the natural language processing (NLP) technology operating those systems, and the issues of their instructional design. We review the scientific findings on the cognitive, behavioral, and emotional effects of dialogue systems. Finally, we provide recommendations for the use of dialogue-based CALL in foreign language learning and teaching, as well as for the development of new conversational applications.