How Social Distance of Process Designers Affects the
Process of Process Modeling:
Insights From a Controlled Experiment
Jens Kolb, Michael
Zimoch
Ulm University, Germany
{jens.kolb, michael.
zimoch}@uni-ulm.de
Barbara Weber
Quality Engineering Research
Group, University of
Innsbruck, Austria
barbara.w[email protected]
Manfred Reichert
Ulm University, Germany
manfred.reichert
@uni-ulm.de
ABSTRACT
The increasing adoption of process-aware information sys-
tems (PAISs) by enterprises has resulted in large process
model collections. Usually, process models are created ei-
ther by in-house domain experts or external consultants.
Thereby, high model quality is crucial, i.e., process mod-
els should be syntactically correct and sound, and also re-
flect the real business processes properly. While numerous
guidelines exist for creating correct and sound process mod-
els, there is only little work dealing with cognitive aspects
affecting process modeling. This paper addresses this gap
and presents a controlled experiment using construal level
theory. We investigate the influence the social distance of a
process designer to the modeled domain has on the creation
of process models. In particular, we are able to show sig-
nificant differences between high and low social distance in
respect to model quality and granularity. The results may
help enterprises to compose adequate teams for creating or
optimizing business process models.
Categories and Subject Descriptors
H.1.2 [User/Machine Systems]: Software Psychology
Keywords
business process modeling, construal level theory
1. INTRODUCTION
Due to the increasing adoption of process-aware informa-
tion systems (PAIS), more and more enterprise repositories
comprise large process model collections [1]. Usually, process
models vary in respect to their quality and level of granu-
larity, and contain a wide range of problems that impede
upon their understandability [2]. However, high quality of
process models is crucial for enterprises to guarantee proper
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise, to
republish, to post on servers or to redistribute to lists, requires prior specific
permission and/or a fee.
SAC’14 March 24-28, 2014, Gyeongju, Korea.
Copyright 2014 ACM 978-1-4503-2469-4/14/03 ...$15.00.
process execution in PAIS [3]. As a prerequisite for the lat-
ter, process models should reflect the real-world processes
properly and at the right level of granularity [4]. To ad-
dress this issue, considerable work on criteria that ensure
process model quality and comprehensibility has been con-
ducted [5, 6]. Further, different modeling guidelines (e.g.,
GoM, 7PMG) exist, supporting process designers in creat-
ing process models of high quality [7, 8]. However, there
is only little work evaluating the influence of cognitive as-
pects on the process of process modeling [9] and their effects
on the resulting process models [10, 11]. However, if we do
not understand the cognitive aspects affecting process model
quality, process modeling projects might not deliver the re-
quired results or even fail. This paper investigates a fun-
damental factor presumably influencing the process of pro-
cess modeling—social distance [12]. The latter is addressed
by the Construal Level Theory (CLT) [13] and constitutes
an important part of psychological distance. In particular,
studies have shown that human thinking and acting are both
strongly influenced by psychological distance [14]. Accord-
ing to CLT, we experience only the here and now, and form
an abstract mental construal of distant objects [12, 14]; e.g.,
when thinking about a concert we plan to visit, it is impor-
tant to know which bands will be playing, but details about
the trip are not in our mind set yet. In turn, just before the
concert, it is important to know with whom to visit or how
to get there, i.e., planning is at a more fine-grained level.
Similarly, in the process of process modeling, various actors
having different distance to the modeled business process
and its environment may be involved. While certain process
models are directly designed by process participants, oth-
ers are modeled by external consultants or in-house experts
(e.g., from the quality assurance department) not involved in
the process (i.e., people having a high social distance to the
process). A relevant question in this context is how social
distance influences resulting process models. Taking CLT
as theoretical basis, we conduct a controlled experiment in-
vestigating the influence social distance has on the process
of process modeling as well as the quality and granularity
of the resulting process models. Our results show a signifi-
cant influence of social distance on the granularity of process
models. Regarding process model quality, significant effects
can be observed as well. Respective insights allow companies
when composing process modeling teams. In particular, our
results indicate that it is advisable to involve process design-
ers with low social distance in any process modeling team.
The experiment has been conducted and results are used in
the context of the proView project1, which utilizes experi-
mental results aiming at user-centered business process sup-
port. In particular, proView provides techniques enabling
personalized process models (i.e., process views [15, 16]) as
well as user-friendly process model visualizations (e.g., dia-
grams and trees [17]) and interactions [18, 19].
Sect. 2 provides backgrounds on CLT. Sect. 3 introduces our
research question and defines the experiment. Sect. 4 deals
with experiment preparation and its execution. Results are
presented and analyzed in Sect. 5. Finally, Sect. 6 discusses
related work and Sect. 7 summarizes the paper.
2. FUNDAMENTALS ON CONSTRUAL
LEVEL THEORY
Construal Level Theory (CLT) describes the effects psy-
chological distance has on objects or events [12, 14]. Gen-
erally, CLT states that increasing psychological distance af-
fects our mental representation of these objects or events.
In turn, this influence on human perception has a strong
impact on our actions and thoughts [13], i.e., everything be-
ing distant to us creates more abstract mental reflections.
The reason behind this phenomenon is the level of construal
(LOC), which describes how individuals interpret and per-
ceive objects and events. Increasing psychological distance
affects the cognitive abilities of individuals and leads to a
change in their perception of an object or event.
CLT describes two levels of thinking: high- and low-level
construal.High-level construals are abstract, imprecise, co-
herent, and superordinate representations compared to low-
level construals. If an object is further away, we think about
it in terms of high-level construals. However, the smaller
our distance to objects or events is, the more we think in
low-level construals. Moreover, these two levels of constru-
als are influenced by psychological distance. While objec-
tive distance describes the quantitative spatial distance in
the real world, psychological distance describes our feelings,
thoughts and emotions in relation to an object or event. In
turn, an object or event is defined as psychological distant,
if it is not experienced physically. For this case, a mental
representation must be constructed. Psychological distance
can be further subdivided into social, spatial,temporal, and
hypothetical distance.Social distance, on which we focus in
this paper, describes our relation to other individuals (cf.
Fig. 1); e.g., whether choosing a more distant seat in a bus
from another individual or not is directly reflected by social
distance [20]. Other psychological distances included in our
work, but not considered in this paper are spatial,temporal,
and hypothetical distance [13, 21, 22, 23].
Low DistanceHigh Distance
abstract, imprecise
specific, precise Person
Figure 1: Construal Level Theory - Social Distance
3. RESEARCH QUESTION AND EXPERI-
MENT DEFINITION
This section deals with definition and planning of our ex-
periment for measuring the influence of social distance on
1www.dbis.info/proView
the process of process modeling and resulting artifacts. Sect.
3.1 explains the context of the experiment and defines its
goal. Sect. 3.2 introduces the hypothesis considered for
testing and Sect. 3.3 presents the experimental setup. Sect.
3.4 explains the design of our experiment. Finally, Sect. 3.5
discusses factors threatening the validity of results.
3.1 Context Selection and Goal Definition
In companies, processes are typically either modeled by
in-house teams or external consultants. Respective process
designers are responsible for interviewing process partici-
pants and capturing gathered knowledge in process models.
However, these designers are often not directly involved in
the process to be modeled; e.g., they may be member of the
quality assurance department. In other cases, due to limited
resources, companies assign such modeling tasks to external
consultants. However, it is not well understood how such an
increased social distance affects the quality and granularity
of resulting process models. To close this gap, this paper
investigates the following fundamental research question:
Is the process of process modeling, i.e., quality and
granularity of the process models resulting from it,
affected by the social distance process designers have on
respective business processes?
Despite the large number of work on process model qual-
ity [2, 8, 24, 25, 26] and granularity of process modeling [27],
there exists only little work dealing with cognitive aspects
of process modeling [10, 11, 28, 29]. In particular, it is not
well understood how these factors actually lead to minor
process quality, i.e., deficiencies regarding syntactic, seman-
tic, and perceived quality, and how they impact granularity
of process models. What has not been considered so far,
is the social distance process designers have to the process
models created by them. According to CLT, however, ex-
isting social distance to objects influences the way we act
and therefore presumably also impacts the way how process
models are created. Based on a controlled software exper-
iment, this paper investigates the influence social distance
has on the process of process modeling and its outcomes.
The experiment varies social distance to learn whether it
has any influence on the quality and granularity of the re-
sulting process models. The goal may be formulated as:
Analyze process models
for the purpose of evaluating
with respect to their level of construal
from the point of view of the researchers
in the context of students and research staff.
3.2 Hypothesis Formulation
Based on the goal of our experiment, the following hy-
potheses are derived. The experiment investigates whether
social distance influences the level of construal during the
process of process modeling, and thus the quality and gran-
ularity of the resulting process model. In total, we have
derived four hypotheses, one for the level of granularity and
three for the quality dimensions, i.e., syntactic, semantic,
and perceived quality:
Does social distance lead to an increase of the
level of granularity when modeling a process?
H0,1: There are no significant differences in the level of granular-
ity when modeling processes with low social distance compared to
high social distance.
H1,1: There are significant differences in the level of granularity
when modeling processes with low social distance compared to
high social distance.
Does social distance lead to an increase of the
syntactic quality when modeling a process?
H0,2: There are no significant differences in the syntactic quality
when modeling processes with low social distance compared to
high social distance.
H1,2: There are significant differences in the syntactic quality
when modeling processes with low social distance compared to
high social distance.
Does social distance lead to an increase of the
semantic quality when modeling a process?
H0,3: There are no significant differences in the semantic quality
when modeling processes with low social distance compared to
high social distance.
H1,3: There are significant differences in the semantic quality
when modeling processes with low social distance compared to
high social distance.
Does social distance lead to an increase of the
perceived quality when modeling a process?
H0,4: There are no significant differences in the perceived quality
when modeling processes with low social distance compared to
high social distance.
H1,4: There are significant differences in the perceived quality
when modeling processes with low social distance compared to
high social distance.
3.3 Experimental Setup
This section describes subjects,object, and response vari-
ables of our experiment as well as its instrumentation and
data collection procedure.
Subjects. Ideally, process designers in enterprises are mod-
eling experts. But, often they only obtain basic training and
have limited process modeling skills [30]. Hence, from sub-
jects we require that they are at least moderately familiar
with process modeling, but we do not require expert level.
Object. The object is a process model using the Business
Process Model and Notation (BPMN). To ensure familiarity
of subjects and that differences in quality and granularity are
not caused due to a lack of familiarity, but rather due to dif-
ferences in social distance, we choose a well-known scenario.
More precisely, the modeling task deals with the process
of visiting the university canteen. We created task descrip-
tions in two versions reflecting different social distances (cf.
Tab. 1). To ensure that the gap between the two distances
is big enough, a pilot study is performed. In the context of
low social distance, visiting the canteen with a good friend
shall be modeled. Regarding high social distance, visiting
the canteen with a foreign (i.e., unknown) student shall be
modeled. Note that task description are rather abstract to
give subjects the possibility to model as detailed as they like.
Low Social Distance:A good friend of you starts to study at
your university. Since it is essential for students to know the
canteen, model the process of a characteristic visit of the canteen.
Start the process at the point, the student enters the canteen.
High Social Distance:A foreign student visits your university.
Since it is essential for students to know the canteen, model the
process of a characteristic visit of the canteen. Start the process
at the point, the student enters the canteen.
Table 1: Task Descriptions
Factor and factor levels. The factor considered in our
experiment is social distance with levels high and low. Ac-
cordingly, the task description is adjusted to vary social dis-
tance: i.e., modeling the canteen process either for a foreign
student (i.e., high) or good friend (i.e., low social distance).
Response variable. As response variable, we consider the
level of construal which cannot be directly measured. Con-
sidering the level of construal, everything being distant from
us is created more abstract (cf. Sect. 2). Hence, we as-
sume that the level of construal impacts quality and granu-
larity of the resulting process model. Therefore, high so-
cial distance may lead to course-grained, more abstract,
and imprecise process models (i.e., reflecting a low qual-
ity/granularity), while low social distance may result in more
fine-grained and precise process models (i.e., reflecting a
high quality/granularity). Hence, process model quality is
characterized by three dimensions, i.e., syntactic,semantic,
and perceived quality, making use of semiotic theory [31].
Syntactic quality of a process model is measured by count-
ing syntactical errors (i.e., syntactical rule violations) of the
applied modeling language (i.e., BPMN) [32]. In turn, se-
mantic quality covers correctness, completeness, relevance,
and authenticity of a process model. Correctness expresses
that all elements in a process model are correct and rele-
vant. Completeness implies that relevant aspects about the
domain are missing. Further, relevance signifies that all el-
ements in the process model are relevant for the process
(i.e., superfluous elements are considered as well). Finally,
authenticity expresses that the chosen representation gives
a true impression of the domain. All four dimensions of
semantic quality are rated by two modeling experts in a
consensus building process from each other and based on a
7-point Likert scale [33]: i.e., between 0 (strongly disagree)
and 6 (strongly agree). Finally, perceived quality depends
on the degree to which a subject agrees with the result-
ing model [34]. Perceived quality can be further subdivided
into agreement,missing aspects,accurate description,mis-
takes, and satisfaction. It is rated by each subject on a
5-point Likert scale, i.e., between 0 (strongly disagree) and
4 (strongly agree), after finishing the modeling task. Agree-
ment expresses to which degree the process model matches
with the real-world business process. Missing aspects rates
whether significant aspects are missing in the resulting pro-
cess model. In turn, accurate description expresses how ac-
curate the process model matches the real-world process.
Mistakes corresponds to the subject rate indicating whether
there are serious mistakes in the resulting process model.
Finally, satisfaction expresses the degree of satisfaction a
subject has with the process model created by him or her.
Process granularity is measured through the complexity of
the resulting process models, i.e., number of activities, edges,
gateways, and execution paths. Fig. 2 summarizes the vari-
ables considered in a research model.
3.4 Experimental Design
We apply guidelines for designing experiments as described
in [35] and conduct a randomized balanced single factor ex-
periment. The experiment is randomized since subjects are
assigned to groups randomly. Further, only a single factor
varies, i.e., the level of construal. Fig. 3 illustrates the setup.
Instrumentation and data collection procedure. To
precisely measure response variables in a non-intrusive man-
ner, we use the Cheetah Experimental Platform (CEP) [9].
CEP provides a BPMN modeling environment that records
all modeling steps and their attributes (i.e., timestamp, type
Psychological Distance
F: Syntactic Quality
O: Number of Syntactical Errors
Quality
Legend:
F: Theoretical Factor
O: Factor Operationalization
F: Semantic Quality
O: Correctness
Authenticity
Relevance
Completness
F: Perceived Quality
O: Agreement
Missing Aspects
Accurate Description
Mistakes
Satisfaction
Granularity
F: Granularity
O: Number of Activities
Number of Edges
Number of Gateways
Number of Elements
Number of Exec. Paths
F: Social Distance
O: Level of Social
Distance
Figure 2: Research Model
Low Social
Distance
Process Model
n Subjects Factor
.
.
.
Subject 1
Subject n/2
High Social
Distance
Process Model
.
.
.
Subject n/2+1
Subject n
Object
Task
Subject
Group 1
Subject
Group 2
Object
Figure 3: Experiment Design
of modeling action). Further, resulting process models are
stored. Finally, demographic data and qualitative feedback
is gathered from subjects based on questionnaires.
3.5 Risk Analysis
Generally, any experiment bears risks that might affect its
results. Thus, its validity or—more precisely—its levels of
validity need to be checked: internal validity (“Are effects
caused by independent response variables?”) and external
validity (“May results be generalized?”).
Risks to internal validity.Risks that might influence the
modeling outcome include process modeling experience of
subjects involved and uneven distributions of subjects over
two groups. Furthermore, post data validation ensures that
in both groups subjects are at least moderately familiar with
process modeling (cf. Sect. 4.3). It is assured that both
groups show about the same familiarity level (i.e., median
is 5.0 and 4.5 for groups). The chosen modeling task consti-
tutes another risk to internal validity. To ensure familiarity
of subjects and to guarantee that differences in quality and
granularity are due to social distance, we choose a scenario
all subjects know well (i.e., the campus has only one canteen,
which is used by all subjects). Particularly, this prevents
faulty models due to lack of domain knowledge. To further
ensure that subjects are not negatively influenced by tired-
ness, boredom, or hunger, the experiment is conducted at a
time of the day for which the mentioned frame of mind can
be excluded. Finally, expected duration of modeling tasks
is 15 min. This should prevent faulty models due to lack of
motivation. Subjects are recruited on a voluntary basis.
Risks to external validity.The subjects have academic
background (i.e., students and research staff), which might
limit generalizability of results. However, the subjects rather
have profound knowledge in process modeling (cf. Sect. 4.3).
Hence, we may consider them as proxies for professionals
who have obtained basic training so far. As another risk for
external validity, process model quality may depend on the
appropriateness of the chosen modeling languages and tools.
To mitigate this risk, both groups use an easy-to-use process
modeling tool as well as an established process modeling
language. Finally, a potential risk for external validity is
that we measure social distance with one modeling task. To
mitigate this and to improve generalizability, experiments in
different environments or conditions one may conduct.
4. EXPERIMENT OPERATION
Based on the provided experiment definition, Sect. 4.1
summarizes the experiment preparation. Sect. 4.2 describes
the execution of the experiment, while Sect. 4.3 introduces
the validation of the data collected during the experiment.
4.1 Experiment Preparation
Students and research staff familiar with process modeling
are invited to join the experiment. Subjects are not informed
about aspects intended to investigate. However, they know
the experiment takes place in the context of a thesis. For
all subjects, anonymity is guaranteed. Before conducting
the experiment, a pilot study is performed whose results
are used, to eliminate ambiguities and misunderstandings
as well as to improve the description of the modeling task.
Additionally, it is checked whether social distance between
the tasks is sufficiently large. Finally, an evaluation sheet is
created to assess the level of construal by analyzing quality
and granularity of resulting process models.
4.2 Experiment Execution
The experiment is executed in a computer lab at Ulm Uni-
versity. All in all 44 students and research staff participate.
Due to spatial constraints, up to 12 subjects may partici-
pate at the same time and several sessions within a period
of four weeks are offered. Each session lasts about 15 min-
utes and runs as follows: The procedure of the experiment is
explained and worksheets with task description are handed
out. Thereby, subjects are randomly assigned to one of the
subject groups (cf. Sect. 3.4). Then, subjects fill out an
initial questionnaire capturing their actual modeling experi-
ence. This information is used to test whether the subjects
are familiar with process modeling. Finally, subjects model
the process based on the textual description from the work-
sheets. After finishing the task, subjects provide their rating
for perceived quality (cf. Sect. 3.3). At the end, they may
give feedback. All results are stored in the CEP database.
4.3 Data Validation
In total, data is collected from 44 subjects. Two of them
are excluded due to invalidity of the process models ob-
tained; i.e., one contains only one start node and the other
one differs substantially from the postulated task descrip-
tion. Finally, 42 subjects are considered for data analysis;
i.e., 32 subjects are students and 10 are research staff; 5 sub-
jects are female. In addition, we screened the subjects for
familiarity with BPMN (process modeling language we use
in our models), since our research setup requires subjects to
be familiar with BPMN. On a Likert scale from 0 to 6 the
median value for familiarity with BPMN is 5.0 (i.e., above
average). For confidence in understanding BPMN process
models, a median value of 5.0 is obtained. Perceived com-
petence in creating BPMN models has a median value of
4.0. Prior to the experiment, subjects analyzed 38 process
models and created 18 in average. Since all values range
above average and subjects are familiar with process mod-
eling, we may conclude that the participating subjects fit to
the targeted profile. The full data set can be found in [36].
5. DATA ANALYSIS & INTERPRETATION
Sect. 5.1 presents descriptive statistics for data gathered
during the experiment. Sect. 5.2 discusses whether a data set
reduction is needed. Sect. 5.3 tests hypotheses from Sect. 3.2.
5.1 Data Analysis and Descriptive Statistics
Figure 4 shows box plots (i.e., median, min, and max val-
ues as well as 1st and 3rd quartiles) of all measures related
to granularity of process models, i.e., number of activities,
edges,gateways,total process elements, and possible paths.
Low HighLow High LowHigh
0
20
10
30
40
#Edges #Gateways #Elements #Paths
High
#Activities
50
60
513
77
Low HighLow
Figure 4: Measurements for Granularity
Figure 4 shows, all measured values are higher in the con-
text of low social distance. In particular, differences in the
numbers of possible paths are especially large. Note that low
social distance results have a median of 21 possible paths,
while high distance only leads to 3.50 possible paths.
Next, Figure 5a displays box plots of measurements for
syntactic and semantic quality.
As shown in Figure 5a, process models with low social dis-
tance seem to give a better account of the domain; however,
subjects make more syntactical errors (median of 3 for low,
2 for high). Further, Figure 5b gives an overview of mea-
sures regarding perceived quality of subjects. It shows that
subjects with low social distance believe that the description
of their designed process models matches the real-world pro-
cess more accurately (median of 3 for low, 2 for high). It is
noteworthy that in conjunction with low distance, satisfac-
tion leads to a median of 1, while high distance leads to a
median of 2. Our observations are merely based on descrip-
tive statistics. For a more rigid investigation, hypotheses
will be tested for statistical significance in Sect. 5.3.
5.2 Data Set Reduction
Generally, results of any statistical analysis depend on the
quality of its input data, i.e., faulty data might contribute to
incorrect conclusions. Therefore, it is important to identify
outliers and evaluate whether they shall be excluded, i.e., a
Low High Low High Low High Low HighLow High
0
2
1
3
4
Correct Relevance Complete Authentic
5
6
#Errors
10
SemanticSyntactic
(a) Measurements for Syntactic and Semantic Quality
Low High Low HighHigh Low
0
2
1
3
4
Missing
Aspects Description Mistakes Satisfaction
Low High Low High
Agreement
(b) Measurements for Perceived Quality
Figure 5: Measurements for Quality Metrics
data set reduction might become necessary. Note that the
latter might be also critical due to potential loss of informa-
tion. In our experiment, several outliers can be identified.
We decided not to remove them since we consider them as
correct, i.e., they are not a result of wrong modeling. One
subject modeled a process model with 513 possible execution
paths. Another subject modeled a task with 10 syntactical
errors. Removing them would bias results obtained. Fur-
thermore, we do not have to remove results from any sub-
ject due to missing process modeling knowledge—since all
subjects are familiar with process modeling (cf. Sect. 4.3).
5.3 Hypothesis Testing
Sect. 5.1 indicates differences regarding low and high so-
cial distance. In the following, we test whether observed
differences are statistically significant. We test our response
variables based on a non-parametric two-sample u-test [37].
A successful u-test (with p < p0at risk level α= 0,05) will
reject a null hypothesis. Table 2 shows results of hypothesis
testing (cf. Sect. 3.2).
In summary, hypotheses H1,1and H1,3can be accepted.
In turn, hypothesis H1,4is only partially supported and thus
it cannot be accepted. Further, hypothesis H1,2must be re-
jected. From this, we can conclude that low social distance
has a positive impact on the granularity (H1,1) and seman-
tic quality (H1,3) of resulting process models. We may also
assume that low social distance has a positive impact on the
perceived quality (H1,4); however, since it is partially sup-
ported, it cannot be generalized. Regarding syntactic quality
(H1,2), no statistically significant difference is observed.
6. RELATED WORK
This paper investigates the impact of social distance on
the quality and granularity of process models. Hence, it is
related to existing research regarding the quality and gran-
ularity of process models. There are different frameworks
and guidelines in respect to process model quality. Among
Level of Granularity H1,1
Response Variable p-value Significant?
Number of activities 0.001 (<0.05) Yes
Number of edges 0.001 (<0.05) Yes
Number of gateways 0.001 (<0.05) Yes
Number of elements 0.001 (<0.05) Yes
Number of paths 0.001 (<0.05) Yes
Syntactic Quality H1,2
Response Variable p-value Significant?
Number of syn. errors 0.087 (>0.05) No
Semantic Quality H1,3
Response Variable p-value Significant?
Correctness 0.006 (<0.05) Yes
Relevance 0.001 (<0.05) Yes
Completeness 0.001 (<0.05) Yes
Authenticity 0.001 (<0.05) Yes
Perceived Quality H1,4
Response Variable p-value Significant?
Agreement 0.017 (<0.05) Yes
Missing aspects 0.347 (>0.05) No
Description 0.034 (<0.05) Yes
Mistakes 0.038 (<0.05) Yes
Satisfaction 0.084 (>0.05) No
Table 2: Results of Hypothesis Testing
others, the SEQUAL framework uses semiotic theory for
identifying various aspects of process model quality [26],
whereas GoM (Guidelines of Process Modeling) describes
quality considerations for process models [7] and 7PMG (Seven
Process Modeling Guidelines) characterize desirable proper-
ties of a process model [8]. Moreover, significant research on
factors affecting process model comprehensibility and main-
tainability exists. The influence of model complexity on pro-
cess model comprehensibility is investigated in [5]. In turn,
[38] analyzes the effect of modularity on process understand-
ing. The influence of grammatical styles for labeling activ-
ities on model understanding is discussed in [39], and an
experiment investigating the impact of secondary notations
is presented in [40]. The impact of different quality metrics
on error probability is discussed in [41].
[42] provides prediction models for true usability and main-
tainability of process models. Effects of how and at which
level of granularity a designer models a particular process is
described in [27]. Regardless, in the context of process mod-
eling there exists little work taking cognitive aspects into ac-
count. [29] presents the effects of reducing cognitive load on
end user understanding of conceptual models. Understand-
ing complex models quickly reaches cognitive limits. [11]
describes the cognitive difficulty of understanding different
relations between model elements.
Common to all these works is the focus on the resulting
process model (i.e., the product of process modeling), while
little attention has been paid on the process of the process
modeling itself. The Nautilus project complements these
works by investigating the process of process modeling for
tracing model quality back to different modeling strategies
resulting in process models of different quality [43, 44].
7. DISCUSSION AND CONCLUSION
This paper investigates whether social distance affects pro-
cess modeling and quality as well as granularity of the re-
sulting process models. In particular, an experiment is con-
ducted showing that for social distance there is a significant
difference depending on whether a process designer has a low
or high social distance to the modeled domain. For enter-
prises, our results indicate that process designers showing a
high social distance to a particular business process tend to
create a more coarse-grained and abstract process model. In
addition, process models, whose designers show a high social
distance, reflect a lower semantic model quality. In particu-
lar, these models tend to be more incomplete and less correct
with respect to the domain to be modeled. Further, process
designers with low social distance are more self-confident
about the process models they create. In the context of an
enterprise, it is thus recommended to involve process de-
signers being more confident with corresponding business
processes; e.g., to achieve a high process model quality.
Based on our results, in general, one can assume that lower
social distance leads to more precise and fine-grained process
models. However, while results look promising, their gener-
alization needs to be confirmed by additional experiments;
i.e., in order to obtain more accurate results allowing such a
generalization, additional studies are needed either through
replication or similar studies in other environments to in-
vestigate the influence of social distance on the process of
process modeling.
Furthermore, our experiments related to other psycholog-
ical distances (i.e., spatial, temporal and hypothetical dis-
tance) will be subject of future papers. Combining experi-
ment results for all psychological distances enables us to ex-
tract guidelines on how modeling teams in enterprises should
be put together and optimal process models can be achieved.
Finally, experiments with practitioners are planned to vali-
date results in a real-world setting.
8. REFERENCES
[1] Weber, B., Reichert, M., Mendling, J., Reijers, H.A.:
Refactoring Large Process Model Repositories.
Computers in Industry 62 (2011) 467–486
[2] Mendling, J.: Metrics for Process Models: Empirical
Foundations of Verifiation, Error Prediction, and
Guidelines for Correctness. Springer (2008)
[3] Reijers, H., Mendling, J.: A Study into the Factors
that Influence the Understandability of Business
Process Models. Systems, Man and Cybernetics, Part
A: Systems and Humans, IEEE Transactions on
(2011) 1–14
[4] Moody, D.: Theoretical and Practical Issues in
Evaluating the Quality of Conceptual Models: Current
State and Future Directions. DKE 55 (2005) 243–276
[5] Mendling, J., Reijers, H.A., Cardoso, J.: What Makes
Process Models Understandable? In: Proc. 5th BPM.
(2007) 48–63
[6] Mendling, J., Strembeck, M.: Influence Factors of
Understanding Business Process Models. In: Proc.
11th Int’l Conf. Business Inf. Systems. (2008) 142–153
[7] Becker, J., Rosemann, M., Uthmann, C.V.: Guidelines
of Business Process Modeling. In: Proc. 1st BPM.
Volume 1806., Springer-Verlag (2000) 30–49
[8] Mendling, J., Reijers, H.A., van Der Aalst, W.M.:
Seven Process Modeling Guidelines (7PMG). Inf. &
Software Techn. 52 (2010) 127–136
[9] Pinggera, J., Zugal, S., Weber, B.: Investigating the
Process of Process Modeling with Cheetah
Experimental Platform. In: Proc. 1st Int’l WS
Empirical Research Proc.-Oriented Inf. Sys.,
Hammamet (2010)
[10] Figl, K., Weber, B.: Individual Creativity in Designing
Business Processes. In: Advanced Information Systems
Engineering Workshops, Springer (2012) 294–306
[11] Figl, K., Laue, R.: Cognitive Complexity in Business
Process Modeling. In: Advanced Information Systems
Engineering, Springer (2011) 452–466
[12] Trope, Y., Liberman, N., Wakslak, C.: Construal
Levels and Psychological Distance: Effects on
Representation, Prediction, Evaluation, and Behavior.
Journal of Consumer Psychology 17 (2007) 83–95
[13] Todorov, A., Goren, A., Trope, Y.: Probability as a
Psychological Distance: Construal and Preferences. J
Experimental Social Psychology 43 (2007) 473–482
[14] Trope, Y., Liberman, N.: Construal-level Theory of
Psychological Distance. Psychological Review 117
(2010) 440–463
[15] Kolb, J., Kammerer, K., Reichert, M.: Updatable
Process Views for User-centered Adaption of Large
Process Models. In: Proc 10th Intl. Conf. on Service
Oriented Computing (ICSOC’12), Shanghai (2012)
[16] Kolb, J., Reichert, M.: Data Flow Abstractions and
Adaptations through Updatable Process Views. In:
Proc 27th Symposium On Applied Computing
(SAC’13), Coimbra, Portugal (2013)
[17] Kolb, J., Reichert, M.: Using Concurrent Task Trees
for Stakeholder-centered Modeling and Visualization
of Business Processes. In: Proc S-BPM ONE 2012,
CCIS 284. (2012) 237–251
[18] Kolb, J., Rudner, B., Reichert, M.: Towards
Gesture-based Process Modeling on Multi-Touch
Devices. In: Proc 1st Int’l Workshop on
Human-Centric Process-Aware Information Systems
(HC-PAIS’12), Gdansk, Poland (2012) 280–293
[19] Kolb, J., Rudner, B., Reichert, M.: Gesture-based
Process Modeling Using Multi-Touch Devices.
International Journal of Information System Modeling
and Design (IJISMD) 4(2013) (to appear)
[20] Pronin, E., Olivola, C.Y., Kennedy, K.a.: Doing Unto
Future Selves As You Would Do Unto Others:
Psychological Distance and Decision Making.
Personality & Social Psych. Bulletin 34 (2008) 224–36
[21] Day, S., Bartels, D.: Representation over Time: The
Effects of Temporal Distance on Similarity. Cognition
106 (2008) 1504–1513
[22] Liberman, N., Sagristano, M.D., Trope, Y.: The Effect
of Temporal Distance on Level of Mental Construal. J
Exp. Social Psychology 38 (2002) 523–534
[23] Fujita, K., Henderson, M.D., Eng, J., Trope, Y.,
Liberman, N.: Spatial Distance and Mental Construal
of Social Events. Psych. Science 17 (2006) 278–82
[24] Moody, D.L.: The ”Physics” of Notations: Toward a
Scientific Basis for Constructing Visual Notations in
Software Engineering. Software Eng. 35 (2008)
756–779
[25] Siau, K., Rossi, M.: Evaluation Techniques for
Systems Analysis and Design Modelling Methods: a
Review and Comparative Analysis. Information
System Journal 21 (2011) 249–268
[26] Krogstie, J.: Model-Based Development and Evolution
of Information Systems. Springer (2012)
[27] Holschke, O., Rake, J., Levina, O.: Granularity as a
Cognitive Factor in the Effectiveness of Business
Process Model Reuse. LNCS 5701 (2009) 245–260
[28] Figl, K., Derntl, M.: The Impact of Perceived
Cognitive Effectiveness on Perceived Usefulness of
Visual Conceptual Modeling Languages. In: Lecture
Notes in Computer Science, Springer (2011) 78–91
[29] Moody, D.: Cognitive Load Effects on End User
Understanding of Conceptual Models: An
Experimental Analysis. In: Advances in Databases
and Information Systems, Springer (2004) 129–143
[30] Wolf, C., Harmon, P.: The State of Business Process
Management 2012. BPTrends Report (2012)
[31] Lindland, O.I., Sindre, G., Solvberg, A.:
Understanding Quality in Conceptual Modeling. IEEE
Software 11 (1994) 42–49
[32] OMG: Business Process Management Notation
(BPMN) 2.0 (2010)
[33] Recker, J., Safrudin, N., Rosemann, M.: How Novices
Model Business Processes. In: Proc. 9th Int’l Conf
Business Process Management (BPM’10), New York,
USA (2010) 29–44
[34] Rittgen, P.: Quality and Perceived Usefulness of
Process Models. In: Proc. ACM SAC 2010, New York,
New York, USA, ACM Press (2010) 65—-72
[35] Wohlin, C., Runeson, P., H¨
ost, M., Ohlsson, M.C.,
Regnell, B., Wesslen, A.: Experimentation in Software
Engineering - An Introduction. Kluwer (2000)
[36] Zimoch, M.: Experiments on Influence of Construal
Level During Process Modeling. Bachelor Thesis, Ulm
University (2012)
[37] Siegel, S., Castellan, J.: Nonparametric Statistics for
the Behavioral Sciences. McGraw-Hill Editions (1988)
[38] Reijers, H., Mendling, J.: Modularity in Process
Models: Review and Effects. Proc. 5th BPM (2008)
20–35
[39] Mendling, J., Reijers, H.A., Recker, J.: Activity
Labeling in Process Modeling: Empirical Insights and
Recommendations. Inf. Sys. 35 (2010) 467–482
[40] Schrepfer, M., Wolf, J., Mendling, J., Reijers, H.A.:
The Impact of Secondary Notation on Process Model
Understanding. Proc. PoEM 09 (2009) 161–175
[41] Mendling, J., Verbeek, H.M.W., van Dongen, B.F.,
van der Aalst, W.M.P., Neumann, G.: Detection and
Prediction of Errors in EPCs of the SAP Reference
Model. Data and Knowledge Eng. 64 (2008) 312–329
[42] Aguilar, E.R., Sanchez, L., Carballeira, F.G., Ruiz, F.,
Piattini, M., Caivano, D., Visaggio, G.: Prediction
Models for BPMN Usability and Maintainability.
Proc. CEC 09 (2009) 383–390
[43] Pinggera, J., Soffer, P., Zugal, S., Weber, B., Weidlich,
M., Fahland, D., Reijers, H., Mendling, J.: Modeling
Styles in Business Process Modeling. Proc. BPMDS
’12 (2012) 151–166
[44] Pinggera, J., Soffer, P., Fahland, D., Weidlich, M.,
Zugal, S., Weber, B., Reijers, H., Mendling, J.: Styles
in business process modeling: an exploration and a
model. Software & Systems Modeling (2013) 1–26
