Information Technology Reference
In-Depth Information
For conceptualizing model comprehensibility in greater detail we draw on
the notion that understanding of a fact in a BPM becomes more dicult if
the number of model elements that need to be attended to increases. This is
backed by the work on Cognitive Load Theory. Cognitive Load Theory builds
on the fact that the capacity of the working memory at a given point of time is
limited [13]. If the amount of information to be processed exceeds this capacity,
comprehension is affected negatively. It has been shown that an instructional
design that avoids an overload of the working memory makes understanding of
the instructional material easier [14]. Prior research on various visual languages
like entity-relationship models [15] or UML class diagrams [16] suggests that
reducing the cognitive load improves the understandability of visual models.
2.2
Influence Factors for Model Comprehensibility
To determine the relevant factors for the cognitive load involved in understanding
elements and their relations in a model, we draw on work on BPM metrics.
Relations between Elements.
Based on the similarity between structures
in software code and process models, research results on code comprehensibil-
ity can serve as a profound basis for analyzing BPM comprehensibility. A large
body of research exists on the cognitive complexity of different programming
elements. Different control structures demand e.g. different levels of effort for
understanding [17]. Little research has been undertaken to investigate the cog-
nitive diculty of different understanding tasks in process models. First efforts
have been made by Melcher et al. [18]. In an experiment with 42 students reading
a rather small BPM (containing 12 activities) they found that understandability
values for questions on the four aspects order, concurrency, repetition and exclu-
siveness are different. As this is the only study in this context, further empirical
research still needs to be done. Additionally, there is another strand of research
stemming from the area of cognitive psychology, relating control flow elements
and cognitive effort. Research on deductive reasoning has shown that systematic
fallacies (so called 'illusory inferences') can occur when individuals construct
or interpret mental models on premises concerning modeling-level connectives
(like conjunctions or disjunctions) [19]. This situation may also be present for
externalized visual BPM and may lead to higher error rates for understanding
specific control flow elements. The current body of literature on error analysis
of process models suggests for instance the existence of systematic reasoning
fallacies concerning routing elements as inclusive OR gateways [20].
Element Interactivity.
The cognitive load that a task imposes on a person is
represented by the number of elements that have to be attended to. This num-
ber is determined by the level of interactivity between the elements. Elements
interact if they are interrelated such that it is necessary to assimilate them si-
multaneously [21]. High interactivity leads to high cognitive load because each
element has to be processed with references to other elements. On the other
hand, cognitive load is low if the elements can be processed serially without
referring to other elements.
