An experimental comparison of abstract and concrete representations in systems analysis

Abstract

The process of information requirements determination requires effective communication between systems analysts and users of the system to be developed. The analyst's ability to discover user requirements is partially determined by the analyst's familiarity with and ability to communicate in the user's domain of knowledge and discourse. One such aspect of the user knowledge domain is concrete terminology versus more abstract, conceptual understanding. This paper documents the results of an experiment which compared knowledge representation used by analysts in a systems development discovery task. We hypothesized that the discovery task would be more effective when the analyst's representation was biased toward the concrete. We found that systems analysts whose initial representation was a physical data flow diagram (concrete) made more correct modifications and fewer errors than systems analysts who started with a logical data flow diagram (abstract). The two groups used the same amount of time for each of the sub-tasks. These results indicate that analyst knowledge and use of concrete terms in the user knowledge domain is of more utility in the discovery task than abstract, conceptual domain knowledge.