Using Insights from Cognitive Neuroscience to Investigate the Effects of Event-Driven Process Chains on Process Model Comprehension [original]

Using Insights from Cognitive Neuroscience to

Investigate the Effects of Event-Driven Process

Chains on Process Model Comprehension

Michael Zimoch1, Tim Mohring1, R¨udiger Pryss1, Thomas Probst2, Winfried

Schlee2, and Manfred Reichert1

1Institute of Databases and Information Systems, Ulm University, Germany

2Department of Psychiatry and Psychotherapy, Regensburg University, Germany

{michael.zimoch, tim.mohring, ruediger.pryss, manfred.reichert}@uni-ulm.de,

thomas.probst@psychologie.uni-regensburg.de, winfried.schlee@googlemail.com

Abstract. Business process models have been adopted by enterprises

for more than a decade. Especially for domain experts, the comprehen-

sion of process models constitutes a challenging task that needs to be

mastered when creating or reading these models. This paper presents

the results we obtained from an eye tracking experiment on process mo-

del comprehension. In detail, individuals with either no or advanced ex-

pertise in process modeling were confronted with models expressed in

terms of Event-driven Process Chains (EPCs), reflecting different levels

of difficulty. The first results of this experiment confirm recent findings

from one of our previous experiments on the reading and comprehension

of process models. On one hand, independent from their level of exper-

tise, all individuals face similar patterns, when being confronted with

process models exceeding a certain level of difficulty. On the other, it

appears that process models expressed in terms of EPCs are perceived

differently compared to process models specified in the Business Process

Model and Notation (BPMN). In the end, their generalization needs to

be confirmed by additional empirical experiments. The presented expe-

riment continues a series of experiments that aim to unravel the factors

fostering the comprehension of business process models by using met-

hods and theories stemming from the field of cognitive neuroscience and

psychology.

Keywords: Business Process Model Comprehension, Event-Driven Process

Chains, Eye Tracking

1 Introduction

Many enterprise repositories comprise large collections of process models, which

represent business processes serving to achieve specific goals with their corre-

sponding actors, tasks, and decisions. Usually, process models vary in respect

to their quality and level of granularity. As a consequence, these models face a

wide range of challenges affecting model comprehensibility and error probabi-

lity. However, the comprehensibility of process models is crucial for enterprises

to enable an overall understanding of respective processes for all involved actors.

A process model may be represented either as a textual or a graphical docu-

mentation, whereas the latter provides specific advantages compared to textual

descriptions [1]. Focusing on the graphical specification of processes, there exist

various modeling languages like, for example, EPCs [2], BPMN [3], and Flow

Chart [4]. Each of these languages is defined through its syntax and contains a

set of graphical symbols for documenting processes. Many studies have shown

that the use of graphical symbols foster process model comprehension [5].

Putting an emphasis on EPCs, this paper presents an experiment on process

model comprehension using eye tracking. Usually, processes expressed in terms

of EPCs consist of three different elements, i.e., functions, events, and logical

connectors. In general, an EPC is a chain of alternating events and functions.

Their specification, in turn, primarily describes the business logic of the process,

thus having a positive impact on process model comprehension [2].

The presented experiment is part of a series of experiments using a conceptual

framework that incorporates concepts from cognitive neuroscience and psycho-

logy for process model comprehension [6]. Regarding the influence of process

modeling expertise on model comprehension, the results are similar compared

to a priorly conducted experiment addressing BPMN process models. However,

the results additionally revealed that process models expressed in terms of EPC

are perceived differently than BPMN models. Due to the fact that the obtai-

ned results may be considered as preliminary, however, they provide promising

insights with respect to the reading and comprehension of process models.

The paper is organized as follows: Section 2 describes the context of the

experiment. The experimental setting and operation is introduced in Section 3.

In Section 4, the obtained results are analyzed and hypotheses are tested for

statistical significance. Finally, Section 5 discusses related work and Section 6

summarizes the paper.

2 Context of the Experiment

The complex biological and cognitive processes in the head of an individual,

whether being conscious or subconscious, ultimately decide how the environ-

ment is perceived and, hence, influence the decisions on the further actions to

be taken [7]. In the domain of process modeling, the application of cognitive

neuroscience and psychology entails auspicious prospects [8–10]. Focusing on is-

sues related to the comprehension of process models, we currently conduct a

series of experiments, utilizing advantages of different concepts from cognitive

neuroscience and psychology. Among others, we strive for a comparison bet-

ween existing process modeling languages to yield the perceived pros and cons

of respective languages. In order to achieve these objectives, we make use of a

conceptual framework we developed [6]. Table 1 presents concepts, for which we

already conducted experiments using the conceptual framework. Furthermore,

Table 1 illustrates the number of involved subjects and analyzed process models.

Table 2, in turn, depicts the evaluated process modeling languages and relates

them to the mentioned concepts, these languages have been evaluated with. In

more detail, we observed and measured changes in the electrodermal activity as

well as in the heart rate of subjects while confronting them with different mo-

deling related issues (e.g., level of difficulty, used process modeling languages)

with respect to the process models [11, 12]. Furthermore, for maybe lowering

the needed amount of mental effort and to may reduce the perceived difficulty,

while reading and comprehending process models, we applied the Cognitive Load

Theory and the Construal Level Theory in this context [13, 14].

Concepts No. Subjects No. Models

Cognitive Load Theory 52 588

Construal Level Theory 136 262

Eye Tracking 41 492

Electrodermal Activity 8 128

Heart Rate 7 168

Table 1: Experiments using the Conceptual Framework

Concepts BPMN EPK Petri Net eGantt Flow Chart UML AD

Cognitive Load Theory • • • • • •

Construal Level Theory •

Eye Tracking • • • • •

Electrodermal Activity • • • • •

Heart Rate • • • • •

Table 2: Process Modeling Languages evaluated with the Conceptual Framework

3 Experimental Setting

The expertise in process modeling might be a factor influencing the comprehen-

sion of process models. This leads us to the following research question:

Research Question

Does expertise in the domain of process modeling has a positive effect on

reading and comprehending process models expressed in terms of EPCs?

To investigate this research question, an experiment using eye tracking is

conducted. Note that the experiment is conducted as a quasi-experiment since

the subjects are assigned by judgment with respect to the level of expertise a

subject has in process modeling [15]. Generally, eye tracking is a reliable measu-

rement method to determine differences between subjects while comprehending

a visual stimulus (e.g., picture) [16]. Furthermore, the use of eye tracking re-

veals insights about cognition as well as cognitive development of a subject and

provides non-invasive indices of brain functions. By using this measurement met-

hod, we want to obtain insights into how subjects are comprehending process

models, while, for example, monitoring attention shifts. In our context, the eye

movements of the participating subjects were recorded, while three EPC process

models had to be comprehended, along with a set of comprehension questions.

In the experiment, the recorded types of eye movements were fixations,sacca-

des, and gaze paths [17]. Fixations are very slow eye movements at a specific

point in a stimulus, whereas saccades represent fast eye movements. To be more

precise, a saccade constitutes a change of fixation of the eyes in a stimulus. The

chronological order of fixations and saccades creates a gaze path. It is found in

eye tracking experiments that experts comprehending a stimulus are more likely

to have a smaller number of fixations, saccades, and consequently a shorter gaze

path length compared to novices [18].

3.1 Hypothesis Formulation

The following six hypotheses were derived to investigate whether or not exper-

tise in process modeling has a positive impact on process model comprehension.

More precisely, we focus on the question whether intermediates (i.e., individuals

with more expertise in process modeling) are more effective regarding the com-

prehension of process models expressed in terms of EPCs compared to novices:

H0,1: Intermediates need not less duration time for process model comprehension compared to

novices.

H1,1: Intermediates need significantly less duration time for process model comprehension com-

pared to novices.

H0,2: Intermediates do not achieve a better score for answering the questions compared to novices.

H1,2: Intermediates achieve a significantly better score for answering the questions compared to

novices.

H0,3: Intermediates do not have a better response time for answering the questions compared to

novices.

H1,3: Intermediates have a significantly better response time for answering the questions compared

to novices.

H0,4: Intermediates do not have less fixations in process model comprehension compared to novi-

ces.

H1,4: Intermediates have significantly less fixations in process model comprehension compared to

novices.

H0,5: Intermediates do not have less saccades in process model comprehension compared to novi-

ces.

H1,5: Intermediates have significantly less saccades in process model comprehension compared to

novices.

H0,6: Intermediates do not have a shorter gaze path in process model comprehension compared to

novices.

H1,6: Intermediates have a significantly shorter gaze path in process model comprehension com-

pared to novices.

3.2 Experimental Setup

This section describes the subjects and objects of the experiment, together with

the independent and dependent variables.

Subjects. There were no prerequisites for participating in the experiment.

Therefore, subjects with diverse backgrounds (i.e., students, academics, and pro-

fessionals) were invited to participate in the experiment. Subjects were informed

that the experiment takes place in the context of process model comprehension

and anonymity was guaranteed for all subjects. As done in other scientific fields,

the categorization into groups (i.e., intermediates and novices), in turn, was ac-

complished by a median split, i.e., based on time spent on process modeling

provided by self-reporting of the subjects.

Objects. In the experiment, three process models reflecting different levels

of model difficulty (i.e., easy, medium, and hard) are presented to the subjects.

In particular, the used process models were created in collaboration with several

experts in the domain of process modeling. The models were expressed in terms

of EPCs. The easy process model comprises only basic modeling elements (i.e.,

events and functions). With rising level of difficulty, new EPC elements were

introduced and the total number of elements was increased. The eye movements

were recorded throughout these comprehension tasks. After comprehending a

process model, four true-or-false comprehension questions, referring solely to

the scenario semantics, had to be answered by the subjects. The questions were

used to evaluate whether or not the process models were correctly interpreted.1

In addition, experts and novices in the domain of process modeling, who were not

participating in the experiment, ranked the used process models with respect to

their level of difficulty. Moreover, they were asked to compare the EPC models

with the priorly used BPMN models to ascertain a comparability between these

two modeling languages [6].

Independent variables. The experiment contains two independent varia-

bles; i.e., the 1

level of difficulty and 2

expertise level of subjects.

Dependent variables. Regarding the level of difficulty, the dependent va-

riables include the 1

duration needed for comprehending a process model, 2



achieved score based on the comprehension questions, and 3

needed response

time for answering the questions. In the context of eye tracking, the dependent

variables include the 4

number of fixations,5

number of saccades, and 6

length

of the gaze path. Fig. 1 summarizes the research model of the experiment.

In general, since the experiment consists of pure comprehension tasks, the

results may be considered as preliminary, i.e., their generalization needs to be

confirmed by additional experiments.

3.3 Experimental Design

The experimental setting is based on the guidelines set out by [19]. The ex-

periment was conducted in a designated lab at Ulm University. Prior to the

1Material downloadable from:

https://www.dropbox.com/sh/th6wc0761ajlxcw/AABs_LXE8mh-ufzSp95lT66za?dl=0

F: Theoretical Factor

O: Operationalization of Factor

Process Model Related Factors

F: Process Model Complexity

O: Level of Difficulty

Personal Related Factors

F: Expertise in Process Modeling

O: Hours Spent with Process Modeling

Perception of Process Models

F: Process Model Comprehension

O: Comprehension Duration

Answering Score

Response Time

Fixation Number

Saccade Number

Gaze Path Length

Fig. 1: Research Model of the Experiment

experiment, three pilot studies with 12 subjects were conducted to improve the

experimental design and material as well as to eliminate potential ambiguities,

e.g., optimization of the used process models. Fig. 2 shows the procedure of the

experiment: 1

An introduction was given, 2

subjects had to sign a consent

form, and 3

demographic data (e.g., expertise in process modeling in general,

familiarity with particular modeling languages) were collected. Subsequently, 4



the eye tracking appliance was calibrated and 5

subjects completed a tutorial

in order to familiarize them with the functionality of the eye tracker and the

procedure of the experiment. Therefore, an exemplary task based on the actual

experiment was shown to the subjects. The experiment could be done either in

English or German. After completing these mandatory steps, 6

subjects needed

to comprehend three EPC process models. First, the process model reflecting an

easy level of difficulty was presented, followed by the medium and, finally, the

difficult model. After studying each of the models, subjects had to answer four

comprehension questions related to the respective model. The questions could

be answered with ’true’, ’false’, or ’uncertain’. When answering the questions,

the process models were not visible. The pure comprehension of process models

(i.e., without any guidance) is uncommon, but we wanted to deliberately disclose

the approaches for the pure comprehension on EPC process models. 7

Finally,

subjects could provide textual or oral feedback.

Instrumentation and data collection procedure. For eye tracking, we

used the SMI iView X Hi-Speed system at a sampling rate of 240 Hz.2The

tracking appliance was placed in front of a monitor that provides the process

models to subjects. Subjects used a keyboard with three predefined keys provi-

ding the options for answering the comprehension questions. Eye tracking data

collected during the experiment were analyzed, visualized, and exported with

SMI BeGaze software [20]. Demographic data and qualitative feedback were

gathered with questionnaires.

2http://www.smivision.com/en/gaze-and-eye-tracking-systems/products/iview-x-hi-speed.html

Introduction Consent Form Calibration

Demographic

Questionnaire

Feedback

PM Process Model | CQ Comprehension Question

True-or-False Questions

Conducted in ENG or GER

Describe the overall

procedure of the experiment

Quasi -Experiment

PM #3 CQ 1-4

PM #2 CQ 1-4

PM #1 CQ 1-4

Tutorial

PM #0 CQ 1-2

Fig. 2: Experimental Design

3.4 Data Validation

Overall, data from 36 subjects (12 female participants) were collected, i.e., 20

students, 12 academics, and 4 professionals participated. Specifically, 19 of them

were computer scientists; additionally, 4 psychologists, 4 economists, and 4 social

workers participated. Finally, 5 subjects did not provide a statement. The median

of hours spent for process modeling, which is used to divide the participants in

intermediates and novices, was 20.5 hours. This resulted in a number of 21

intermediates and 15 novices. Furthermore, eye tracking data from 4 subjects

were excluded due to invalidity, i.e., eye movements were not recorded properly.

Finally, the group of novices included 14 subjects and the one of intermediates

consisted of 18 subjects.

4 Data Analysis and Interpretation

Table 3 presents mean and standard deviation (i.e., STD) for novices and inter-

mediates. It shows the process model comprehension duration (in ms) as well as

the achieved answering scores. Thereby, specific values to each answering option

were assigned, i.e., ’true’ = 1, ’false’ = -1, and ’uncertain’ = 0. Furthermore,

response times for answering related questions (in ms), number of fixations as

well as number of saccades and, finally, length of the gaze path (in px) are listed

in Table 3 (i.e., theoretical factor and operationalization of factor).3

Generally, all values, except the answering scores, increase with rising level

of difficulty, as expected by us. For the process model with an easy level of

difficulty, the results reveal that intermediates are more effective in terms of

process model comprehension compared to novices. The comprehension duration

is shorter and, in average, intermediates made less mistakes in answering the

3Sample images downloadable from:

www.dropbox.com/sh/th6wc0761ajlxcw/AABs_LXE8mh-ufzSp95lT66za?dl=0

Theoretical

Factor

Operation.

of Factor

Both Novices Intermediates

Mean STD Mean STD Mean STD

Difficulty

Easy

Comprehension

Duration 36398 23034 43481 29257 28304 7980

Score 3.57 0.82 3.19 0.98 4 0

Resp. Time 4933 1326 5254 1497 4575 1039

Eye Tracking

Fixations 104 57 120 71 85 26

Saccades 90 44 100 56 79 22

Gaze Path 15149 12081 17927 15599 11974 4948

Difficulty

Medium

Comprehension

Duration 54360 18325 58228 21385 49940 13490

Score 2.63 1.4 2.81 1.22 2.43 1.6

Resp. Time 7985 2138 7931 1843 8045 2504

Eye Tracking

Fixations 171 44 178 51 164 34

Saccades 154 39 156 48 151 28

Gaze Path 26197 7421 26666 8990 25661 5385

Difficulty

Hard

Comprehension

Duration 90355 30750 98771 36306 80737 20039

Score 1.73 1.57 1.88 1.5 1.57 1.7

Resp. Time 8358 2719 8528 3307 8163 1949

Eye Tracking

Fixations 279 107 299 136 256 57

Saccades 253 101 261 126 243 64

Gaze Path 40511 17599 43566 21957 37020 10491

Table 3: Obtained Experimental Results

questions. Furthermore, they needed less fixations and saccades, resulting in

a shorter gaze path. Concerning the process model with the medium and the

one with the highest difficulty, the results between novices and intermediates

are approaching a similar level and only slight differences are observable. It

appears to be that novices and intermediates perform equally regarding the

comprehension of EPC process models. However, it is noteworthy that the results

do not differ significantly considering the fact that few novices have had no

experience in EPCs at all. Figs. 3 - 6 show selected results of the experiment.

Fig. 3 indicates that the time needed for process model comprehension increases

with rising level of difficulty. Fig. 4 illustrates that the answering scores are

decreasing with rising level of difficulty. Moreover, for the process models with

the medium and highest level of difficulty, novices achieve a slightly better score.

The response times for answering the questions increase with rising level of

difficulty (cf. Fig. 5). Especially between the easy and medium process model, a

difference is discernible. The fixation number for intermediates is always lower

than the number for novices (cf. Fig. 6). Altogether, EPC process models seem

to be fairly comprehensible without previous knowledge.

4.1 Hypotheses Testing

The stated hypotheses (cf. Sect. 3.2) are tested for statistical significance using

aStudent’s t-test. Particularly, we refer to the rule of thumb, according to the

Level of Difficulty

Comprehension Duration

100

Easy Medium Hard

Expertise

Int Nov

Fig. 3: Comprehension Duration

Level of Difficulty

Answering Score

2.0

2.5

3.0

3.5

4.0

Easy Medium Hard

Expertise

Int Nov

Fig. 4: Answering Score

Level of Difficulty

Response Time

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

Easy Medium Hard

Expertise

Int Nov

Fig. 5: Response Time

Level of Difficulty

Fixation Number

100

150

200

250

300

Easy Medium Hard

Expertise

Int Nov

Fig. 6: Fixation Number

literature, that the use of the t-test is appropriate at a sample size of n > 30

[21]. In this context, a successful t-test (with p<p0at risk level α= 0,05) will

reject a null hypothesis [22]. Only for hypothesis H1,2for the easy process model,

a significant result emerged. In addition, tendencies are discernible in H1,1and

H1,4. However, the results confirm our observations from the first experiment

that process model comprehension not necessarily correlates with expertise in

this domain [6]. This is also evident in the results obtained from the process

models with a higher level of difficulty.

4.2 Threats to Validity

Any experiment bears risks and, hence, its levels of validity need to be checked

and discussed. The level of difficulty reflected by the respective process models

Theoretical

Factor

Operationalization

of Factor

Level of Difficulty

Easy (p)Medium (p)Hard (p)

Comprehension

H1,1- Duration 0.062 0.210 0.100

H1,2- Score 0.005* 0.473 0.610

H1,3- Resp. Time 0.162 0.890 0.712

Eye Tracking

H1,4- Fixations 0.084 0.365 0.262

H1,5- Saccades 0.181 0.714 0.631

H1,6- Gaze Path 0.165 0.710 0.300

Table 4: Hypotheses Testing Results

constitutes a threat to validity. The gap between the single difficulties might

be not alike. Furthermore, the comprehension of process models without any

concrete task or guidance is uncommon in practice. Moreover, process models

must be memorized for answering the comprehension questions. Consequently,

the risk arises that process models are wrongly memorized. Furthermore, pro-

cess scenarios are perceived differently based on various factors (i.e., familiarity).

Therefore, the considered scenarios constitute an additional risk. Finally, split-

ting and defining novices and intermediates, based on the hours spent on process

modeling, is another threat to validity. Obviously, an individual with a high num-

ber of hours spent with process modeling can be considered as an expert, but it is

questionable whether 20.5 hours are sufficient to denote an intermediate. Moreo-

ver, the representativeness of the results is limited by the relatively small sample

sizes. The sizes of the samples also limit the statistical power and there might

be significant differences between novices and intermediates, which we could not

show in this experiment, but which might become apparent in larger samples.

Regarding the statistics, it has to be mentioned that multiple t-tests were per-

formed and no correction for multiple testing was applied. We justify this by

the fact that this experiment was an explorative one instead of an experiment

aiming to replicate findings of a previous experiment on EPC.

4.3 Discussion and Comparison with prior Results

Prior to this experiment, a similar experiment (i.e., same setting and operation)

was conducted using process models expressed in terms of BPMN 2.0 [6]. Furt-

hermore, special care was taken that the used process models are comparable

between the two modeling languages as well as their levels of difficulty. The

main differences between both experiments were the various process scenarios

captured in the used process models. However, prior experiments showed that

a familiar or unfamiliar scenario does not have an influence on the comprehen-

sion of process models [23]. In general, two common effects could be observed

in both experiments with respect to process model comprehension. First, the

performance of subjects decreases with rising level of difficulty and, second, the

performance of novices and intermediates approximates with each other as well

with rising level of difficulty. It is noteworthy that the overall performance of

subjects confronted with EPC process models indicates better results compared

to BPMN models. Descriptively, comprehension duration and response times for

answering related questions are almost the same, but the final answering scores

are substantially higher for EPCs, despite expertise in process modeling and le-

vel of difficulty. It appears that subjects cope better with EPC process models

than BPMN models. However, in the end, the stated observations need to be

investigated by inferential statistics and by further research either through re-

plication or similar experiments. Finally, based on a set of stated categories that

can foster experiments on process model comprehension, and with the use of the

conceptual framework, further experiments will be subject of future work [23].

5 Related Work

With a focus on process model comprehension, [24] evaluates different process

modeling languages, whereas [25] presents factors that influence the comprehen-

sion of process models. The influence of process model complexity on related

model comprehensibility is investigated in [26]. In [27], factors that affect the

comprehension of process models are discussed. A state-of-the-art report on em-

pirical research on process model comprehension can be found in [28].

Regarding cognitive aspects in process modeling, [29] discusses how the com-

prehension of conceptual models is influenced by a reduced cognitive load. [30]

shows the difficulty of comprehending different relations between the elements

in a process model. [31], in turn, discusses principles for the design of a cogniti-

vely effective visual modeling language, whereas [32] attempts to operationalize

perceptual properties of modeling languages to improve their cognitive effecti-

veness. Moreover, [9] discusses individual preferences for process representations

based on their cognitive style.

In line with eye tracking, [33] concludes that a higher mental effort can be

measured by the change of pupil dilation for task-based process models. Furt-

hermore, [34] presents results on how eye tracking leads to a more fine-grained

understanding of process models. [35] studies the factors that influence process

model comprehension using eye tracking.

Comparing process models expressed in terms of BPMN and EPC respecti-

vely, [36] describes an experiment in which subjects ranked the modeling langua-

ges according to their subjective comprehension difficulty. The results conclude

that BPMN models are easier to comprehend. Finally, [37] investigates the diffe-

rences between BPMN and EPC process models, yielding that no final message

can be made regarding a positive perception.

6 Summary and Outlook

This paper investigates whether or not expertise in the domain of process mo-

deling has a positive impact on process model comprehension. Particularly, an

eye tracking experiment, using process models in terms of EPCs and related

results were presented. The preliminary results indicate that expertise in pro-

cess modeling not necessarily implies a better comprehension of process models.

In detail, intermediates as well as novices are struggling similarly once process

models exceed a certain level of difficulty. The obtained results are in line with

results from a priorly conducted experiment using BPMN process models. Ho-

wever, it appears that EPC process models are easier to read and comprehend.

Further, with broader and more detailed investigations, we attempting to con-

firm respective generalization. The presented experiment is part of a series of

experiments, in which a conceptual framework is used that incorporates met-

hods and theories from cognitive neuroscience and psychology [6]. Thereby, eye

tracking constitutes only one measurement method from the pool of existing

methods in this context. By using the conceptual framework, additional con-

cepts from neuroscience and psychology (e.g., electrodermal activity, Cognitive

Load Theory) will be used in future experiments with a particular focus on how

to foster the reading and comprehension of process models.

References

1. Ottensooser, A., Fekete, A., Reijers, H.A., Mendling, J., Meicstas, C.: Making Sense

of Business Process Descriptions: An Experimental Comparison of Graphical and

Textual Notations. In: J of Systems and Software. Volume 85. (2012) 596–606

2. van der Aalst, W.M.P.: Formalization and Verification of Event-driven Process

Chains. In: Inf & Soft Tech. (1999) 639–650

3. OMG: Business Process Management & Notation 2.0 (2017) www.bpmn.org, last

visited on 2017/02/27.

4. Schultheiss, L.A., Heiliger, E.: Techniques of Flow-Charting. In: Proc 1963 Clinic

on Library Applications of Data Processeing. (1963) 62–78

5. Johansson, L.O., W¨arja, M., Carlsson, S.: An Evaluation of Business Process

Model Techniques, Using Moodys Quality Criterion for a Good Diagram. In:

CEUR Workshop Proceedings. Volume 963. (2012)

6. Zimoch, M., et al.: Cognitive Insights into Business Process Model Comprehen-

sion: Preliminary Results for Experienced and Inexperienced Individuals. In: Proc

BPMDS’17. (2017) 137–152

7. Schwarz, N.: Emotion, Cognition, and Decision Making. (2000) 433–440

8. Zugal, S., Pinggera, J., Weber, B.: Assessing Process Models with Cognitive Psy-

chology. In: EMISA. Volume 190. (2011) 177–182

9. Figl, K., Recker, J.: Exploring Cognitive Style and Task-Specific Preferences for

Process Representations. In: Requir. Eng. (2014) 63–85

10. Recker, J., Reijers, H.A., van de Wouw, S.G.: Process Model Comprehension: The

Effects of Cognitive Abilities, Learning Style, and Strategy. Volume 34. (2014)

199–222

11. Prokasy, W.: Electrodermal Activity in Psychological Research. Elsevier (2012)

12. Camm, A.J., et al.: Heart Rate Variability: Standards of Measurement, Physiolo-

gical Interpretation and Clinical Use. Volume 93. (1996) 1043–1065

13. Sweller, J., Ayres, P., Kalyuga, S.: Cognitive Load Theory. Springer (2011)

14. Trope, Y., Liberman, N.: Construal-Level Theory of Psychological Distance. Vo-

lume 117. (2010) 440–463

15. Cook, T.D.: Quasi-Experimental Design. Wiley (2015)

16. Gegenfurtner, A., et al.: Expertise Differences in the Comprehension of Visualiza-

tions: A Meta-Analysis of Eye-Tracking Research in Professional Domains (2011)

17. Salvucci, D.D., Goldberg, J.H.: Identifying Fixations and Saccades in Eye-Tracking

Protocols. In: Proc 2000 Symp on Eye Tracking Research & App. (2000) 71–78

18. Raney, G.E., Campbell, S.J., Bovee, J.C.: Using Eye Movements to Evaluate the

Cognitive Processes Involved in Text Comprehension (2014)

19. Wohlin, C., Runeson, P., H¨ost, M., Ohlsson, M.C., Regnell, B., Wesslen, A.: Ex-

perimentation in Software Engineering - An Introduction. Kluwer (2000)

20. SMI: iView X Hi-Speed (2016) http://www.smivision.com/en/gaze-and-eye-

tracking-systems/products/iview-x-hi-speed.html, last visited on 2017/02/27.

21. Hogg, R.V., Tanis, E.A.: Probability and Statistical Inference. Macmillan New

York (1977)

22. Sirkin, M.: Statistics for the Social Sciences. Volume 3. Sage (2005)

23. Zimoch, M., et al.: Eye Tracking Experiments on Process Model Comprehension:

Lessons Learned. In: Proc BPMDS’17. (2017) 153–168

24. Kiepuszewski, B., Hofstede, A.H.M.T., Bussler, C.: On Structured Workflow Mo-

delling. In: Proc 12th Int’l Conf on Adv Inf Systems Engineering. (2000) 431–445

25. Melcher, J., Mendling, J., Reijers, H., Seese, D.: On Measuring the Understanda-

bility of Process Models. In: BPM Workshops: LNBIP. (2009) 465–476

26. Mendling, J., Reijers, H.A., Cardoso, J.: What Makes Process Models Understan-

dable? In: Proc 5th Int’l Conf on Business Process Mgmt. (2007) 48–63

27. Mendling, J., Strembeck, M., Recker, J.: Factors of Process Model Comprehension-

Findings from a Series of Experiments. In: Decis. Support Syst. (2012) 195–206

28. Figl, K.: Comprehension of Procedural Visual Business Process Models–A Litera-

ture Review. Volume 59. (2017) 41–67

29. Moody, D.L.: Cognitive Load Effects on End User Understanding of Conceptual

Models: An Experimental Analysis. In: Adv in Data and Inf Sys. (2004) 129–143

30. Figl, K., Laue, R.: Cognitive Complexity in Business Process Modeling. In: Ad-

vanced Information Systems Engineering. (2011) 452–466

31. Moody, D.: The ”Physics” of Notations: Toward a Scientific Basis for Constructing

Visual Notations in Software Engineering. In: Trans. Softw Eng. (2009) 756–779

32. van der Linden, D., Zamansky, A., Hadar, I.: How Cognitively Effective is a Visual

Notation? On the Inherent Difficulty of Operationalizing the Physics of Notations.

In: Proc BPMDS’16. (2016) 448–462

33. Dobesova, Z., Malcik, M.: Workflow Diagrams and Pupil Dilatation in Eye Tracking

Testing. In: Proc 13th Int’l Conf on Emer eLearning Tech & App. (2015) 59–64

34. Hogrebe, F., Gehrke, N., N¨uttgens, M.: Eye Tracking Experiments in Business Pro-

cess Modeling: Agenda Setting and Proof of Concept. In: Proc 4th Int’l Workshop

on Enterprise Modelling and Information Systems Architectures. (2011) 183–188

35. Petrusel, R., Mendling, J.: Eye-Tracking the Factors of Process Model Compre-

hension Tasks. In: Proc 25th Int’l Conf on Adv Inf Sys Eng. (2013) 224–239

36. Gabryelczyk, R., Jurczuk, A.: The Diagnosis of Information Potential of Selected

Business Process Modelling Notations. In: Information Systems in Management.

Volume 4. (2015) 26–38

37. Recker, J.C., Dreiling, A.: The Effects of Content Presentation Format and User

Characteristics on Novice Developers Understanding of Process Models. In: Com-

munications of the Association for Information Systems. Volume 28. (2011) 65–84