Information Technology Reference
In-Depth Information
As examples, Sedera et al. [21] used case study research and survey methods to
derive qualitatively a framework of factors that influence the success of process
modeling efforts. Amongst others, they found tooling and participation to be key
drivers. Participative model building was investigated by Persson [16]. She found
that it leads to enhanced model quality, more stakeholder consensus and more
commitment to results. The workshops are set up with a dedicated software
tool operator to channel participant knowledge and create a common picture
at the projector [23]. Rittgen developed software and guidance for modeling
workshops in which the participants themselves use the software to create the
model together [18]. Our approach uses intuitive tooling to remove software as
a barrier for individuals to participate.
For individuals, Recker found that modeler performance is influenced by the
complexity of the grammar [17], modeling experience and modeling background.
Controlled experiments with individuals have been conducted e.g. by Weber et
al. [24] to investigate the effect of events on process planing performance or by
Holschke [10] to investigate the influence of model granularity on reusability of
artifacts. To our best knowledge there is no controlled experiment that investi-
gates the presence of an intuitive mapping tool for business process modeling.
The setup and execution of our controlled experiment was guided by Creswell
[3] and Wohlin et al. [25]. We use literature from experimental software engineer-
ing [12] and statistics [6] to inform the structure of the paper and the level of
reporting.
3 Experiment Planning
We outline all planning activities in this section. We start by deriving our hy-
potheses, talk about setup, the actual measurement of the hypotheses and the
analysis procedures.
3.1 Goal and Hypotheses
The goal of this paper is to examine the effect of t.BPM on process elicitation
with individuals. Therefore we compare t.BPM to structured interviews which
are seen as the most effective requirements elicitation technique [4]. By 'effective'
we mean that it produces a 'desired or intended result' [22]. In requirements en-
gineering, more information is typically indicating more effective elicitation. But
it was already shown that the presence of visual representations does not neces-
sarily elicit more information [4]. We argue that effective process elicitation has
more aspects such as user engagement and validated results. Fig. 2 visualizes how
we refine our model towards hypotheses based on the following considerations:
User engagement is widely recognized as a key factor for success of IT
projects [21]. Our approach uses tangible media which is seen as a key factor
for task engagement, e.g. in HCI research [11]. In those cases, engagement is
typically measured as the time spent on a problem, e.g. by Xie et al. [26]. Since
tangible modeling consumes time to handle the tool itself (e.g, writing on tiles),
