e-space
Manchester Metropolitan University's Research Repository

    Public appeals, news interviews and crocodile tears: An argument for a multi-channel annotation scheme

    Archer, DE and Lansley, C (2015) Public appeals, news interviews and crocodile tears: An argument for a multi-channel annotation scheme. Corpora, 10 (2). pp. 231-258. ISSN 1749-5032

    [img]
    Preview

    Available under License In Copyright.

    Download (602kB) | Preview

    Abstract

    In this paper, we propose a Six Channel Analysis System (SCAnS) for the (semi-)automatic investigation of potential deception across all communication channels. SCAnS builds on our current system: Six Channel Analysis in Realtime (SCAnR). SCAnR users are trained to code – as Points of Interest (PIns) – relevant occurrences of twenty-seven criteria relating to the six channels, when they appear to point to inconsistencies with respect to the speaker's account (the story they are trying to convey), their apparent baseline and the context. Our experiences to date confirm the view that multi-channel approaches have the potential to lead to higher accuracy rates of deception detection than is possible when using individual methods of detection and/or when focussing on one communication channel independently (Vrij et al., 2000: 257), especially when combined with cognitive elicitation strategies. However, we recognise the importance of (in)validating the relevance of the twenty-seven criteria through ongoing research. SCAnS will provide a (semi-)automated means of achieving this. Given our audience, we focus on the usefulness of content-analysis tools like Wmatrix (Rayson, 2008) for this purpose. More broadly, our research has implications for the analysis of data in forensic contexts, across all available channels of communication, and for the coding of (para)linguistic features.

    Impact and Reach

    Statistics

    Activity Overview
    6 month trend
    801Downloads
    6 month trend
    406Hits

    Additional statistics for this dataset are available via IRStats2.

    Altmetric

    Repository staff only

    Edit record Edit record