Multi-speaker experimental designs: Methodological considerations

Abstract

Research on language use has become increasingly interested in the multimodal and interactional aspects of language – theoretical models of dialogue, such as the Communication Accommodation Theory and the Interactive Alignment Model are examples of this. In addition, researchers have started to give more consideration to the relationship between physiological processes and language use. This article aims to contribute to the advancement in studies of physiological and/or multimodal language use in naturalistic settings. It does so by providing methodological recommendations for such multi-speaker experimental designs. It covers the topics of (a) speaker preparation and logistics, (b) experimental tasks and (c) data synchronisation and post-processing. The types of data that will be considered in further detail include audio and video, electroencephalography, respiratory data and electromagnetic articulography. This overview with recommendations is based on the answers to a questionnaire that was sent amongst the members of the Horizon 2020 research network ‘Conversational Brains’, several researchers in the field and interviews with three additional experts.