Evaluating the Comprehensibility of
Graphical Business Process Models:
An Eye Tracking Study
Michael Zimoch1, R¨udiger Pryss1, Thomas Probst3, Winfried Schlee3, Georg
Layher2, Heiko Neumann2and Manfred Reichert1
1Institute of Databases and Information Systems, Ulm University, Germany
2Institute of Neural Information Processing, Ulm University, Germany
3Department of Psychiatry and Psychotherapy, Regensburg University, Germany
{michael.zimoch, ruediger.pryss, georg.layher, heiko.neumann,
manfred.reichert}@uni-ulm.de, thomas.probst@psychologie.uni-regensburg.de,
winfried.schlee@googlemail.com
Problem. Process models provide detailed information about tasks, decisions,
and actors involved in various business processes. Graphical representations pro-
vide tangible benefits regarding process model comprehension compared to tex-
tual documentations. Many unresolved issues regarding the factors thwarting
the understanding of process models, e.g., process model quality, exist. Here,
we use eye tracking to monitor selective attention shifts and serial groupings of
semantically meaningful chunks in process model comprehension.
Method. 36 subjects (23 male) had to study 12 different process models expres-
sed in BPMN, eGantt, EPC, and Petri Net by conducting a reading comprehen-
sion task. Further, subjects answered a questionnaire with questions related to
the process described in the models. Subjects’ scanning saccade patterns and
relative fixation durations were recorded with SMI iView X Hi-Speed system at
240 Hz.
Results and Conclusion. We observed specific eye-movement patterns (e.g.,
targeted search, back-and-forth saccade jumps) as well as unique strategies for
reading different process model representations. Additionally, scan path pattern
and fixation time variabilities indicate different levels of cognitive load and re-
veal potential stumbling blocks in the context of graphical business process mo-
del comprehension. The results, in turn, enrich the development of a conceptual
framework, targeting at the comprehension of business process models.
2
Evaluate Results and Derive RulesExperimental Setting for Process
Model Comprehension
Process Model Characteristics
Subject
A
B
Categorize Model
Difficulty based on
Experiment Data
A
B
Create Better
Reference Process
Models
directly influences
2 3 4
Cognitive Neuroscience and Psychology
Measurements and Theories
Derive Process
Modeling Rules
1.
2.
3.
4.
Business Process
Modeling Expert
Data Scientist
a b
Statistical and
Empirical Evaluation
Reference Process Models
in Different Notations
1
Experiment
Designer
Business Process
Modeling Expert
...
BPMN 2.0 Petri Nets
EPCs
^v
X
Y
Z
A XVS
X Y
Fig. 1: Conceptual Framework for Process Model Comprehension
