Personal construct theory as a research tool for analysing user perceptions of photographs

Abstract

This paper describes a preliminary research project which applies Personal Construct Theory to individual user perceptions of photographs. The background to Personal Construct Theory and to the use of Repertory Grids is presented. The analysis provided by repertory grids is compared with that from various theoretical and practical approaches to indexing visual images, including ’ofness’ and ’aboutness’, facet analysis, iconology and MPEG7. A high level of consistency is found among the personal constructs which participants use to distinguish between photographs. The research concludes that repertory grids provide a useful method of collecting unbiased data about what users see in visual images and for comparing user perceptions with alternative retrieval vocabularies and methods. Incorporation of a participant’s constructs in automatic classification systems for visual images remains a major challenge.