Propositional Idea Density: Computerized analysis to determine effects of presence and severity of aphasia

Ferguson, Alison and Spencer, Elizabeth and Bryant, Lucy and Craig, Hugh and Colyvas, Kim and Worrall, Linda (2013) Propositional Idea Density: Computerized analysis to determine effects of presence and severity of aphasia. [Clinical Aphasiology Paper]

PDF Propositional_Idea_Density_Computerized_analysis_to_determine_effects_of_presence_and_severity_of_aphasia.pdf Download (122kB)

Abstract

This paper presents research that aimed to extend the available analyses of informativeness of aphasic discourse. A ‘proposition’ can be defined as a linguistic relation and its associated arguments (Kintsch & Keenan, 1973; Turner & Greene, 1977), and has been used as an index of informativeness in research on language and aging. The proportion of propositions in a text (Propositional Idea Density – PD) has been found to be a sensitive index of age-associated cognitive impairment and dementia (Riley, Snowdon, Desrosiers, & Markesbery, 2005). The research on PD has primarily used manual analysis methods, noting high training needs for raters to ensure adequate inter-coder and intra-coder reliability, as has also been found in analyses of informativeness in the field of aphasia (Nicholas & Brookshire, 1993; Oelschlaeger & Thorne, 1999; Yorkston & Beukelman, 1980). The development of a computer program, Computerized Propositional Idea Density Rater known as CPIDR (Brown, Snodgrass, & Covington, 2007; Brown, Snodgrass, Kemper, Herman, & Covington, 2008) has made the process of calculating PD accessible to untrained individuals. The benefits of a computer-based program are further seen in reliability, with 100% consistency when re-counting a single sample, and inter-rater reliability of 97% when compared to manual calculations which is more reliable than most human coders (Brown, et al., 2008). The present research made use of this computerised analysis of PD to investigate the effects of aphasia on informativeness. It was hypothesised that information content, as measured by PD, would be significantly reduced in the oral discourse of people with aphasia when compared to non-aphasic controls, and that PD would decrease with increasing aphasia severity as determined by Western Aphasia Battery - Aphasia Quotient (Kertesz, 2006).

Actions (login required)

View Item