Why diagrams are (sometimes) six times easier than words: Benefits beyond locational indexing

Abstract

By building computational models, Larkin and Simon (1987) showed that the effects of locational indexing give an explanation of 'Why a diagam is (sometimes) worth ten thousand words', to quote the title of their seminal paper. This paper reports an experiment in which participants solved three versions of Larkin and Simon's simple pulley system problem with varying complexity. Participants used a diagrammatic, tabular or sentential representation, which had different degrees of spatial indexing of information. Solutions with the diagrams were up to six times easier than informationally equivalent sentential representations. Contrary to predictions derived from the idea of locational indexing, the tabular representation was not better overall than sentential representation and the proportional advantage of the diagrammatic representation over the others did not increase with problem complexity. This suggests that the advantage of diagrams goes beyond the effects that locational indexing has on the processes of searching for items of information and the recognition of applicable rules. A possible explanation resides in the specific problem solving strategies that the participants may have been using, which depended on the structure of the representations and the extent to which they supported solution path recognition and planning.