Master Thesis defense by Steven Dassen


BPMN 2.0 in practice: The usefulness of extended constructs in a business environment.


Many studies are devoted to the theoretical capabilities of the workflow language BPMN, but not much is known about the actual practical use of BPMN in a business environment. Recker’s research (2010) concluded that business analysts barely use extended constructs in their business models, so they only use a core set of constructs (basic constructs). He questioned whether the extended constructs have an added value when used in industrial business process models. The research
objective of this study is to examine the influence of the extended constructs in BPMN 2.0 with regard to the understandability and correctness of a process model for practitioners in the healthcare domain.

To investigate this objective a theoretical analysis and an experiment were conducted. The theoretical analysis was based on the peri-operative process in hospitals. NICTIZ1, an organisation which focuses on process optimisation in the healthcare domain, delivered the textual description (guidelines) of this process. A part of the process (pre-operative process) was already modelled by business practitioners with BPMN (NICTIZ model). To examine the influence of the different
constructs, two different versions of the peri-operative process were modelled by the researcher, a model with only basic constructs (basic version) and a model with the full set of constructs (extended version). To measure and compare the understandability of the models, the complexity metrics for business process models were used (Gruhn & Laue, 2006). The correctness (level of correct behaviour) of the models was analysed by reproducing the state spaces of the models and illustrating
the behaviour of the models via tokens. These state spaces were compared with the guidelines.

The theoretical analysis showed that the models with the full set of constructs were more understandable than the models with only basic constructs based on the results of the complexity metrics. Moreover, the state spaces showed how the extended constructs can improve the correctness of a model. The experiment measured the actual effectiveness of the various constructs in BPMN based on the Method Evaluation Model (Moody, 2003). The experiment contained a test that consisted of
different questions related to the extended constructs to assess the factors of interest empirically. The purpose of the test was to examine the participants’ understandability and modifiability (ability to modify a model correctly) regarding the constructs (basic/extended). The focus in the experiment was solely on the constructs, not the process. The participants of the experiment were business analysts from the healthcare domain. A workshop, which explained the behaviour of several constructs, was constructed to examine whether training improved the results of the participants. Two different versions of the test were constructed to see whether the results were not test specific. The results of the tests were analysed and hypotheses were tested using the independent t-test. The experiment showed that the results of the test were not test specific (hypothesis 1). Also, it showed that the workshop did not have a significant effect on the results of the participants (hypothesis 2). Furthermore, it showed that practitioners do not encounter more difficulty understanding (hypothesis 3) or modifying (hypothesis 4) models with extended constructs compared to similar models with only basic constructs.

Online Access

Get the PDF from the TU/e library