Consistency, context and confidence in judgements of affective communication in adults with profound intellectual and multiple disabilities.

Twenty-four service providers rated 12 video samples of four service users with whom they were familiar for affective behaviour (i.e. 'like'/'dislike') and confidence (i.e. 'certain'/'uncertain') in their judgement. Each video sample had been recorded as part of a stimulus preference assessment during which a wide range of specific stimuli were presented to each service user. Each video sample was presented twice in a counterbalanced design either with contextual information, i.e. what the presented stimulus was (C) or without such information, i.e. context free (CF). The observers showed considerable individual variation in their judgements, largely uninfluenced by the availability or otherwise of contextual information. However, as a group, observers significantly distinguished between video samples with regard to affective communication (determined through multiple analyses of variance) and the pattern of judgements, i.e. the relative judgement of positive or negative affect, from one sample to another. This showed a good level of consistency between observers (determined through principal components analysis). The impact of contextual information was not apparent for all video samples. However, contextual information significantly influenced judgements in four samples, typically making them more extreme; for example, a response indicative of positive affect in the CF situation became more positive when contextual information was provided, indicating that the stimulus was one that the participant was thought to like.