WORKFLOW AND ORGANIZATIONAL UNIT: AN EMPIRICAL COMPARISON OF ANALYSIS PERSPECTIVES

Abstract

Many processes, techniques, tools, methodologies, and approaches claim to facilitate the process of information systems development, but little empirical validation in support of these claims has been publicly reported. This research addresses this shortcoming in two ways. First, it develops and applies a promising experimental design for the comparison of systems analysis techniques. The design's objective was to external validity of experimental findings while maintaining high degrees of control and comparability. Secondly, our design, the "transcript experiment," was used to evaluate two versions of an analysis procedure. This paper both presents and evaluates the transcript experiment as a research design and reports the results of an actual experiment. The study we report investigated the impact of a particular factor in the systems analysis process, which we term analysis perspective. After elaborating a (partial) theory of systems analysis enabling us to predict the impact of different analysis perspectives on (1) the analysis process, (2) the content of reports it produces, and (3) the utility of the analysts’ recommendations, we compared the influences of two particular perspectives, the workflow perspective and the organizational unit perspective. We observed significant differences in subject behavior in acquiring information during the analysis process, but the data were inconclusive with respect to our predictions concerning the content of reports and the utility of subjects’ recommendations. Finally, we noted a strong negative correlation between the number of recommendations produced by a subject and the degree to which he documented the current system. We term this correlation the descriptive/prescriptive tradeoff, and feel it deserves further study, as it may invalidate a number of widely-held assumptions concerning the systems design process.