Empirical analysis of a proposed process granularity heuristic

Abstract

Choosing the adequate size of process activities (process granularity) is a problem during process design. Vanderfeesten et al. have proposed a heuristic based on a process granularity metric and postulated a hypothesis concerning error probability about its use. The heuristic prefers process designs with high cohesion and low coupling-a principle originating in software engineering. In this paper, we present an experimentation system consisting of a small web-based workflow engine for empirically analyzing the error probability hypothesis. Furthermore, the results of a conducted experiment with 165 students using this experimentation system are reported. The experiment does not support the hypothesis. Instead, an alternative error probability model explaining the results is suggested. © 2010 Springer-Verlag.