scieee Science in your language
[en] (orig)
Ulm University
Institute of Databases
and Information Systems
2018
Dissertation
Johannes Schobel
A Model-Driven Framework for Enabling Flexible
and Robust Mobile Data Collection Applications
Titelbild:
c
Fabian Maier
Universität Ulm | 89069 Ulm | Germany Fakultät für Ingenieurwissenschaften,
Informatik und Psychologie
Institut für
Datenbanken und Informationssysteme
A Model-Driven Framework for
Enabling Flexible and Robust
Mobile Data Collection Applications
Dissertation zur Erlangung des Doktorgrades Dr. rer. nat.
der Fakultät für Ingenieurwissenschaften, Informatik und Psychologie der Universität Ulm
Vorgelegt von:
Johannes Schobel
geboren in Bregenz, Österreich
2018
“Begin at the beginning and go on till
you come to the end: then stop.”
(Lewis Carroll, Alice in Wonderland)
Amtierender Dekan: Prof. Dr. Frank Kargl
Gutachter: Prof. Dr. Manfred Reichert
PD Dr. Winfried Schlee
Tag der Promotion: 28. September 2018
ii
Vorwort
Die hier vorliegende Dissertation ist im Rahmen meiner beruflichen Tätigkeit als Dok-
torand und wissenschaftlicher Mitarbeiter am Institut für Datenbanken und Informa-
tionssysteme der Universität Ulm entstanden. In dieser Zeit habe ich viele Personen
kennengelernt, ohne deren Unterstützung diese Arbeit in diesem Umfang und dieser Form
nicht denkbar gewesen wäre – dafür möchte ich mich bedanken.
Zuerst möchte ich Herrn Prof. Dr. Manfred Reichert für die tatkräftige Unterstützung
während meiner Promotionszeit und das freundliche Aufnehmen im Institut – schon
damals während meiner Zeit als Student – danken. Noch mehr jedoch möchte ich mich für
deine kollegiale Art und dein Vertrauen in den letzten 6 Jahren mir gegenüber bedanken.
Darüber hinaus möchte ich Herrn Dr. Winfried Schlee für die vielen positiven Gespräche
und zahllosen Ideen in diesem Themenumfeld danken.
Besonderer Dank gilt meinen Kollegen am Institut, die nicht einfach “nur Arbeitskollegen”,
sondern gute Freunde geworden sind. Für die vielen hilfreichen Diskussionen, Blödeleien
zwischendurch, gemeinsamen Konferenzreisen und vieles mehr möchte ich mich herzlich
bei Marc Schickler, Michael Zimoch, Michael Stach und Kevin Andrews bedanken. Der
größte Dank gilt jedoch Herrn Dr. Rüdiger Pryss: Mit deiner aufopfernden, freundlichen
und hilfsbereiten Art hast du maßgeblich zu dieser Arbeit beigetragen. Dem ganzen
Team möchte ich für das freundliche Arbeitsklima, die gute Zusammenarbeit und die
aufbauenden Worte danken. Ich habe jeden Tag gerne mit euch am Institut verbracht
und bin froh, dass ich mit euch zusammen arbeiten durfte.
Meine Zeit im Institut war auch geprägt durch enge Kooperationen mit interdiszi-
plinären Fachbereichen und Arbeitsgruppen. Insbesondere möchte ich das Team um
Prof. Dr. Thomas Elbert und Dr. Martina Ruf-Leuschner der Universität Konstanz
nennen. Die spannenden Einblicke haben mich maßgeblich dazu motiviert, das in dieser
Arbeit beschriebene Rahmenwerk zu entwickeln und technisch umzusetzen. Weiter
möchte ich die Gruppe um Herrn Dr. Winfried Schlee der Universität Regensburg her-
vorheben. Durch die Mitarbeit an den “Track Your”-Projekten konnte ich viele spannende
Eindrücke in weitere Domänen sammeln. Darüber hinaus möchte ich mich bei Herrn
Prof. Dr. Thomas Probst für die hilfreichen Diskussionen, die Nachhilfe im Bereich
Statistik und die wirklich unkomplizierte Zusammenarbeit bedanken.
v
Weiter möchte ich “meinen” Studenten danken, die mich in zahlreichen Projekt- und
Abschlussarbeiten tatkräftig unterstützt haben. Besonderer Dank geht dabei an Steffen
Scherle, Juri Schulte, Karoline Blendinger, Arnim Schindler, Thomas Miholic, Fabian
Maier, Wolfgang Wipp, Robin Martin, Philipp Butz, Jens Winkler, Fabian Widmann,
Dominic Gebhardt und Lena Arndt.
Zuletzt möchte ich mich bei meiner Partnerin und meiner Familie für die Unterstützung,
Geduld und den Rückhalt bedanken; Ihr wisst, wie dankbar ich euch bin.
Danke für die schöne Zeit!
vi
Abstract
In the light of the ubiquitous digital transformation, smart mobile technology has
become a salient factor for enabling large-scale data collection scenarios. Structured
instruments (e.g., questionnaires) are frequently used to collect data in various application
domains, like healthcare, psychology, and social sciences. In current practice, instruments
are usually distributed and filled out in a paper-based fashion (e.g., paper-and-pencil
questionnaires). The widespread use of smart mobile devices, like smartphones or tablets,
offers promising perspectives for the controlled collection of accurate data in high quality.
The design, implementation and deployment of mobile data collection applications,
however, is a challenging endeavor. First, various mobile operating systems need to be
properly supported, taking their short release cycles into account. Second, domain-specific
peculiarities need to be flexibly aligned with mobile application development. Third,
domain-specific usability guidelines need to be obeyed. Altogether, these challenges turn
both programming and maintaining of mobile data collection applications into a costly,
time-consuming, and error-prone endeavor.
The Ph. D. thesis at hand presents an advanced framework that shall enable domain
experts to transform paper-based instruments to mobile data collection applications. The
latter, in turn, can then be deployed to and executed on heterogeneous smart mobile
devices. In particular, the framework shall empower domain experts (i.e., end-users)
to flexibly design and create robust mobile data collection applications on their own;
i.e., without need to involve IT experts or mobile application developers. As major
benefit, the framework enables the development of sophisticated mobile data collection
applications by orders of magnitude faster compared to current approaches, and relieves
domain experts from manual tasks like, for example, digitizing and analyzing the collected
data.
vii
Zusammenfassung
Getrieben durch die fortschreitende digitale Transformation nehmen mobile Technologien
einen immer größeren Stellenwert für das Erfassen großer Datenmengen ein. Insbesondere
in der Medizin, der Psychologie und den Sozialwissenschaften werden häufig strukturierte
Instrumente (beispielsweise Fragebögen) eingesetzt, um Daten in unterschiedlichen Szena-
rien und Studien mobil zu erfassen. Diese werden allerdings, trotz bekannter Nachteile
immer noch größtenteils, in papierbasierter Form durchgeführt. Die flächendeckende
Verbreitung mobiler Endgeräte (beispielsweise Smartphones oder Tablets) ermöglicht
visionäre Ansätze zur kontrollierten Erhebung großer Datenmengen in hoher Qualität.
Die Konzeption, Entwicklung und Verteilung mobiler Anwendungen zur kontrollierten
Datenerhebung ist allerdings aus mehreren Gründen herausfordernd. Erstens müssen für
eine breite Nutzbarkeit unterschiedliche mobile Betriebssysteme (beispielsweise Android
und iOS) adäquat unterstützt werden. Eine besondere Schwierigkeit bilden die relativ
kurzen Entwicklungszyklen dieser Plattformen. Zweitens müssen Besonderheiten der
jeweiligen Anwendungsdomäne berücksichtigt und mit dem Entwicklungsprozess für
mobile Anwendungen in Einklang gebracht werden. Drittens sollten domänenspezifis-
che Anforderungen für Benutzeroberflächen und -schnittstellen berücksichtigt werden.
Insgesamt ist die Entwicklung und Wartung mobiler Anwendungen zur kontrollierten
Datenerhebung daher kostspielig, aufwändig und fehleranfällig.
Die vorliegende Dissertation stellt ein umfassendes Rahmenwerk vor, welches es ermöglicht,
papierbasierte Fragebögen in mobile Anwendungen zur digitalen Datenerhebung zu trans-
formieren. Die resultierenden Anwendungen können dann auf unterschiedlichen mobilen
Betriebssystemen und Gerätetypen ausgeführt werden. Das entwickelte Rahmenwerk
soll insbesondere Fachanwender aus verschiedenen Domänen (beispielsweise Medizin
oder Psychologie) in die Lage versetzen, solche mobilen Anwendungen eigenständig zu
entwickeln und zu nutzen, d.h. ohne Einbinden von IT-Experten oder Programmierer.
Das Rahmenwerk erlaubt es einerseits, komplexe Anwendungen zur Datenerhebung
wesentlich schneller als bisher zu entwickeln, andererseits werden manuelle Tätigkeiten,
wie das Übertragen der erhobenen Daten in digitale Formate und deren Analyse, erheblich
vereinfacht.
ix
List of Publications
This cumulative dissertation is a consolidated report of the research results obtained
during the author’s Ph. D. project. The detailed results have been published in the
following refereed papers:
SPPSSR18
J. Schobel, R. Pryss, T. Probst, W. Schlee, M. Schickler, and M. Reichert.
Learnability of a Configurator Empowering End Users to Create Mobile
Data Collection Instruments: Usability Study. JMIR mHealth and uHealth,
6(6):e148, 2018
SPSR17a
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Patterns for
Defining and Changing Data Collection Instruments in Mobile Healthcare
Scenarios. In 30th IEEE Int’l Symp on Computer-Based Medical Systems
(CBMS), June 2017
SPSPGSR17
J. Schobel, R. Pryss, W. Schlee, T. Probst, D. Gebhardt, M. Schickler,
and M. Reichert. Development of Mobile Data Collection Applications
by Domain Experts: Experimental Results from a Usability Study. In
29th Int’l Conf on Advanced Information Systems Engineering (CAiSE),
number 10253 in LNCS, pages 60–75. Springer, June 2017
SPSR16
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Lightweight
Process Engine for Enabling Advanced Mobile Applications. In 24th Int’l
Conf on Cooperative Information Systems (CoopIS), number 10033 in
LNCS, pages 552–569. Springer, October 2016
SPSR16a
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Configurator
Component for End-User Defined Mobile Data Collection Processes.
In Demo Track of the 14th Int’l Conf on Service Oriented Computing
(ICSOC), October 2016
SPSR16b
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Flexible
Mobile Data Collection in Healthcare. In 29th IEEE Int’l Symp on
Computer-Based Medical Systems (CBMS), pages 181–182, June 2016
xi
SPWSR16
J. Schobel, R. Pryss, W. Wipp, M. Schickler, and M. Reichert. A Mobile
Service Engine Enabling Complex Data Collection Applications. In 14th
Int’l Conf on Service Oriented Computing (ICSOC), number 9936 in
LNCS, pages 626–633, October 2016
SPSRER16
J. Schobel, R. Pryss, M. Schickler, M. Ruf-Leuschner, T. Elbert, and
M. Reichert. End-User Programming of Mobile Services: Empowering
Domain Experts to Implement Mobile Data Collection Applications. In
5th IEEE Int’l Conf on Mobile Services (MS), pages 1–8. IEEE Computer
Society Press, May 2016
SPR15
J. Schobel, R. Pryss, and M. Reichert. Using Smart Mobile Devices
for Collecting Structured Data in Clinical Trials: Results From a Large-
Scale Case Study. In 28th IEEE Int’l Symp on Computer-Based Medical
Systems (CBMS), pages 13–18. IEEE Computer Society Press, June 2015
SSPR15
J. Schobel, M. Schickler, R. Pryss, and M. Reichert. Process-Driven
Data Collection with Smart Mobile Devices. In 10th Int’l Conf on Web
Information Systems and Technologies (Revised Selected Papers), number
226 in LNBIP, pages 347–362. Springer, 2015
Additionally, parts of this thesis have been published in the following publications.
•
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Process-Driven Mobile Data
Collection (Extended Abstract). In 8th Int’l Workshop on Enterprise Modeling and
Information Systems Architectures (EMISA), June 2017
•
M. Schickler, M. Reichert, R. Pryss, J. Schobel, W. Schlee, and B. Langguth.
Entwicklung mobiler Apps: Konzepte, Anwendungsbausteine und Werkzeuge im
Business und E-Health. eXamen.press. Springer Vieweg, October 2015
•
J. Schobel, M. Schickler, R. Pryss, M. Reichert, and T. Elbert. A Domain-Specific
Framework for Collecting Data in Trials with Smart Mobile Devices. In XIV
Congress of European Society for Traumatic Stress Studies (ESTSS) Conf, June
2015
•
D. Isele, M. Ruf-Leuschner, R. Pryss, M. Schauer, M. Reichert, J. Schobel, A. Schindler,
and T. Elbert. Detecting Adverse Childhood Experiences with a Little Help from
Tablet Computers. In XIII Congress of European Society of Traumatic Stress
Studies (ESTSS) Conf, pages 69–70, June 2013
•
M. Ruf-Leuschner, R. Pryss, M. Liebrecht, J. Schobel, A. Spyridou, M. Reichert,
and M. Schauer. Preventing Further Trauma: KINDEX Mum Screen - Assessing
and Reacting Towards Psychosocial Risk Factors in Pregnant Women with the Help
xii
of Smartphone Technologies. In XIII Congress of European Society of Traumatic
Stress Studies (ESTSS) Conf, pages 70–70, June 2013
•
J. Schobel, M. Ruf-Leuschner, R. Pryss, M. Reichert, M. Schickler, M. Schauer,
R. Weierstall, D. Isele, C. Nandi, and T. Elbert. A Generic Questionnaire Framework
Supporting Psychological Studies with Smartphone Technologies. In XIII Congress
of European Society of Traumatic Stress Studies (ESTSS) Conf, pages 69–69, June
2013
•
J. Schobel, M. Schickler, R. Pryss, H. Nienhaus, and M. Reichert. Using Vital Sensors
in Mobile Healthcare Business Applications: Challenges, Examples, Lessons Learned.
In 9th Int’l Conf on Web Information Systems and Technologies (WEBIST), Special
Session on Business Apps, pages 509–518, May 2013
xiii
Contents
I Problem Description and Backgrounds 1
1 Introduction 3
1.1 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Research Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Fundamentals 9
2.1 Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Business Process Management . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Mobile Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4 End-User Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3 Development Strategies 17
II The QuestionSys Framework 19
4 Concept 23
4.1 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2 Mobile Data Collection Lifecycle . . . . . . . . . . . . . . . . . . . . . . . 25
4.3 Model-Driven Development of Instruments . . . . . . . . . . . . . . . . . . 26
5 Architecture 29
5.1 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.2 Process-Aware Instrument Configurator . . . . . . . . . . . . . . . . . . . 31
5.3 Flexible Mobile Data Collection Client . . . . . . . . . . . . . . . . . . . . 32
6 Discussion 35
7 Related Approaches 41
7.1 Research Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.2 Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
xv
Contents
III Validation 45
8 Studies 49
8.1 Pilot Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
8.2 Usability Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
8.2.1 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
8.2.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
9 Related Studies 57
IV Conclusion 61
10 Summary and Outlook 63
10.1 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
10.2 Additional Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
10.3 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Bibliography 69
Index 83
Acronyms 85
V Appendix 87
A Appendix Files 89
A.1 Example Modeling Task Description . . . . . . . . . . . . . . . . . . . . . 89
B List of Publications 93
B.1
Using Smart Mobile Devices for Collecting Structured Data in Clinical
Trials: Results From a Large-Scale Case Study . . . . . . . . . . . . . . . 95
B.2 Process-Driven Data Collection with Smart Mobile Devices . . . . . . . . 95
B.3 A Lightweight Process Engine for Enabling Advanced Mobile Applications 96
B.4
A Configurator Component for End-User Defined Mobile Data Collection
Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
B.5 Towards Flexible Mobile Data Collection in Healthcare . . . . . . . . . . . 97
B.6 A Mobile Service Engine Enabling Complex Data Collection Applications 97
B.7
End-User Programming of Mobile Services: Empowering Domain Experts
to Implement Mobile Data Collection Applications . . . . . . . . . . . . . 98
B.8
Towards Patterns for Defining and Changing Data Collection Instruments
in Mobile Healthcare Scenarios . . . . . . . . . . . . . . . . . . . . . . . . 98
B.9
Development of Mobile Data Collection Applications by Domain Experts:
Experimental Results from a Usability Study . . . . . . . . . . . . . . . . 99
xvi
Contents
B.10
Learnability of a Configurator Empowering End Users to Create Mobile
Data Collection Instruments: Usability Study . . . . . . . . . . . . . . . . 99
C Complete List of Publications 101
D Discussion of Personal Contribution 107
D.1
Using Smart Mobile Devices for Collecting Structured Data in Clinical
Trials: Results From a Large-Scale Case Study . . . . . . . . . . . . . . . 109
D.2 Process-Driven Data Collection with Smart Mobile Devices . . . . . . . . 109
D.3 A Lightweight Process Engine for Enabling Advanced Mobile Applications 109
D.4
A Configurator Component for End-User Defined Mobile Data Collection
Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
D.5 Towards Flexible Mobile Data Collection in Healthcare . . . . . . . . . . . 110
D.6 A Mobile Service Engine Enabling Complex Data Collection Applications 110
D.7
End-User Programming of Mobile Services: Empowering Domain Experts
to Implement Mobile Data Collection Applications . . . . . . . . . . . . . 110
D.8
Towards Patterns for Defining and Changing Data Collection Instruments
in Mobile Healthcare Scenarios . . . . . . . . . . . . . . . . . . . . . . . . 111
D.9
Development of Mobile Data Collection Applications by Domain Experts:
Experimental Results from a Usability Study . . . . . . . . . . . . . . . . 111
D.10
Learnability of a Configurator Empowering End Users to Create Mobile
Data Collection Instruments: Usability Study . . . . . . . . . . . . . . . . 111
E Curriculum Vitae 113
xvii
Part I
Problem Description and
Backgrounds
1
1
Introduction
In a variety of application domains, the controlled collection of large datasets with a high
quality and validity is of paramount importance. Domains like healthcare, psychology
and social sciences, for example, rely on well designed and established instruments
(e.g., self-report questionnaires) to collect data in large-scale scenarios (e.g., clinical
or psychological trials [
20
]). In current practice, datasets are predominantly collected
with paper-based questionnaires, which are disadvantageous in several respects. Before
processing and analyzing the collected data, for example, the latter has to be transferred
to digital spreadsheets – a process that is time-consuming as well as error-prone, especially
in the context of large-scale trials.
According to [
63
], approximately 50
−
60% of the costs related to the collection, transfer
and processing of data could be saved when relying on digital instruments instead of
paper-based ones. This especially applies to long-running data collection procedures.
Studies have proven that the use of digital instruments does not affect psychometric
properties of subjects [
14
], but rather contributes to more complete datasets compared
to the ones collected in a paper-based way [
52
]. This, in turn, significantly increases data
quality [
62
], while decreasing the time required to collect the data [
43
]. Finally, studies
have revealed that the use of smart mobile devices for collecting data might pave the way
for new findings [
19
]. Moreover, digitally collected data may be enriched with contextual
information (e.g., time and location of an interview [
70
]), vital parameters collected with
sensors (e.g., pulse measurement during an interview [
49
]), or environmental data (e.g.,
weather [
102
] or noise). Altogether, digital data collection is increasingly demanded by
domain experts in a multitude of application domains.
3
1 Introduction
1.1 Problem Statement
Although there exists research works demonstrating the applicability of smart mobile
devices in data collection scenarios [
24
,
64
,
76
], current approaches are rarely used in
large-scale scenarios (e.g., clinical or psychological trials). Note that in such scenarios
thousands instances of an instrument need to be processed. Other works investigated the
use of smart mobile technologies in limited scenarios [7,13,55].
Regarding the development process of mobile data collection applications several chal-
lenges need to be tackled.
Mobile operating systems:
The application to be developed may have to be provided
for a broader audience. Amongst others, the application needs to be provided for a
variety of mobile operating systems (e.g., Android vs. iOS). However, each mobile
operating system relies on specific programming languages (e.g.,
Java
or
Kotlin
for Android and
ObjectiveC
or
Swift
for iOS) and proposes specific user interface
guidelines, adding complexity the to development process of the mobile application.
Cross-platform development frameworks [
31
] may be used to bridge this gap and to
deploy the developed application to various platforms. Corresponding approaches,
however, are usually limited to features provided across all platforms.
Lifecycle and release management:
Developers need to cope with short release
cycles of mobile platforms, resulting in costly and time-consuming adaptations to
be able to continuously support new releases. In this regard, multiple versions of
a mobile application may co-exist at the same time, which increases complexity
significantly.
Sensors:
Internal and external sensors have to be integrated properly in order to meet
advanced requirements set out by domain experts.
Practical use:
Challenges related to the deployment and practical use of mobile
application may emerge (e.g., security concerns in a hospital environments) [18].
Assist users:
Transferring complex navigation logic of a paper-based instrument
to its digital counterpart to guide (untrained) users during the process of data
collection (e.g., to skip questions based on already given answers or to validate data)
causes considerable communication efforts between domain experts and application
developers [37].
Fig. 1.1 illustrates these challenges. To the best of the authors knowledge, no generic
approach exists that supports the transformation of paper-based instruments to smart
mobile applications in the context of data collection scenarios.
In a pre-study of this Ph. D. thesis, various mobile data collection applications were
realized (cf. Table 1.1). As these mobile applications were specifically tailored to their
application scenario, the aforementioned drawbacks re-emerged for each scenario. Domain
4
1.1 Problem Statement
Traditional Paper-Based
Instruments
Structured Data Collection
using Smart Mobile Devices
Generic Approach to Address Specific Issues
Transfer
Error-Prone
Cost-Intensive
Huge
Communication
Effort
Constant
Adaptations
Time-Consuming
Expensive
Development Fragile
Inflexible
Figure 1.1: Contribution of the Thesis
experts working with these mobile applications, however, craved for more advanced
features, e.g., to enable audio recordings during interviews or to provide on-demand
evaluations based on pre-specified rules. Maintaining such complex applications for
various mobile platforms, in multiple versions and languages, and over a long period of
time constitutes a challenging endeavor.
Data Collection Scenario Country CN Duration
(Years)
Application
Versions
Collected
Datasets
Study on Tinnitus Research [68] World-Wide #5 + 5 ≥45,000
Risk Factors during Pregnancy [84] Germany #5 + 5 ≥1,500
Risk Factors after Pregnancy Germany #2 + 1 ≥500
PTSD in War Regions [108] Burundi 4+5≥2,200
PTSD in War Regions [129] Uganda #1+1≥200
Adverse Childhood Experiences
[35]
Germany 2+3≥150
Learning Deficits among Medical
Students
Germany 1+3≥200
Supporting Parents after Accidents
of Children
EU #3 + 6 ≥5,000
Overall 29 ≥54,750
CN = Complex Navigation; PTSD = Post-Traumatic Stress Disorder
Table 1.1: Mobile Data Collection Applications Developed
In order to cope with these drawbacks as well as emerging requirements from various
application domains, a generic approach is required. On one hand, such an approach,
needs to cope with the process of developing mobile data collection applications in general,
with the goal to reduce the time and costs required for realizing a data collection scenario.
On the other, the approach shall reduce the communication efforts for application
5
1 Introduction
developers and domain experts. Most importantly, the approach shall empower domain
experts to develop specific mobile data collection applications serving their needs.
1.2 Research Contribution
When writing this thesis, only little research was available targeting at the support of
domain experts in developing domain-specific data collection applications, which then can
be used in large-scale scenarios (e.g., clinical trials). In consequence, this thesis aims at
developing fundamental concepts, techniques and prototypes for developing mobile data
collection applications. The research contributions are summarized in the following:
1.
Requirements from a variety of application scenarios are elicited and collected based
on structured interviews with experts from the respective domains. Furthermore,
additional insights are gathered when realizing several mobile data collection appli-
cations that shall support domain experts in their daily data collection procedures.
2.
The thesis proposes a lifecycle that covers different phases of data collection scenarios
in general. These phases, in turn, as well as their characteristics are observed for
several mobile data collection applications. Further, the lifecycle may act as a
blueprint for collecting data in large-scale scenarios in general.
3.
A well-formed mapping is described based on which paper-based instruments can
be transformed to digital data collection instruments that can then be deployed to
and executed on smart mobile devices. This mapping, in turn, is based on the idea
of describing instruments in a process-centric way.
4.
Adomain-specific modeling language is proposed that enables domain experts to
develop data collection instruments themselves. Moreover, the proposed (graphical)
modeling notation builds upon
BPMN
2.0, but omits language elements not needed
in the given application context. Finally, this language shall simplify the modeling
process in general, by not overloading domain experts with unnecessary information.
5.
A conceptual architecture is presented, which enables domain experts to develop
specific data collection applications: The components of this architecture are related
to the lifecycle, and aim at properly supporting domain experts in collecting data in
large-scale scenarios. Finally, it is illustrated how process management technology
may serve as a fundamental pillar of the architecture, allowing for a high degree of
flexibility.
6.
The conceptual architecture is implemented in the QuestionSys framework. In this
context, proof-of-concept prototypes are developed for all major components to
demonstrate the applicability of the proposed approach.
6
1.3 Outline
7.
A set of common patterns that may be used to create and adapt data collection
instruments are discovered from real-world data collection scenarios. These patterns
are implemented by an advanced configurator component. Their semantics, however,
is independent from a specific modeling language, i.e., the patterns may be applied
in the context of other configurators as well.
8.
The configurator as well as its underlying graphical modeling language are validated
in various usability studies to demonstrate its applicability. In this context, recruited
participants from various fields worked with the configurator to develop specific
mobile data collection instruments.
Altogether, the developed framework aims at supporting domain experts to develop
domain-specific mobile data collection applications on their own. A comprehensive set
of prototype applications have been developed proving the practical feasibility of the
QuestionSys approach and overall applicability.
1.3 Outline
This cumulative Ph. D. thesis is structured into four parts:
Part I motivates the need for utilizing smart mobile devices for collecting data in large-
scale scenarios (cf. Chapter 1). Furthermore, background information is presented in
Chapter 2, whereas Chapter 3 discusses design approaches for realizing mobile data
collection applications.
Part II presents the developed QuestionSys framework. The publications SPSR16 [
112
],
SPSR16a [
113
], SPWSR16 [
115
], and SPSRER16 [
114
] report on major contributions
achieved in this context. Additional insights are provided in [
107
,
109
,
111
,
116
]. Chapter
4 presents the overall concept of modeling data collection instruments. Further, it
describes how (Business) Process Management technologies can be used to drive the
developed approach. Chapter 5 considers core components of the QuestionSys framework,
e.g., the configurator and mobile data collection applications. In Chapter 6, key aspects
of the framework are summarized, whereas Chapter 7 discusses related approaches from
research and industry.
Part III validates the QuestionSys framework and its components. The publications
SPSPGSR17 [118] and SPPSSR18 [119] discuss major findings gathered in this context.
Chapter 8 presents usability studies. More specifically, the complex study design used to
evaluate the developed configurator is sketched, and major findings of the studies are
presented. Chapter 9 reports on related work and compares the approaches with the
conducted studies.
Part IV concludes the thesis. Chapter 10 highlights major scientific contributions and
provides an outlook on future work.
7
2
Fundamentals
This chapter introduces fundamental concepts needed for the understanding of this
thesis.
2.1 Instruments
The term psychological instrument origins from the German Psychologischer Apparat
Psychological
Instrument
[
26
], as it originally referred to machines for properly executing scientific experiments.
Nowadays, psychology utilizes tests performed in standardized situations and environ-
ments to draw conclusions on specific human behavior. Corresponding tests can be
subdivided into questionnaires, rating scales, and standardized interviews [
26
]. The
author of [
26
] defines psychological instruments as “an association of some material
object and a process-generating rule, or a somehow materialized procedural rule, which
for psychological research, teaching or practice, represents or adapts a part of the rational
knowledge of a particular society at a particular time, that knowledge possibly but not
necessarily being psychological”. This thesis denotes questionnaires as instruments in the
following.
During this Ph. D. thesis, a multitude of instruments from various domains (e.g., health-
care, psychology, and logistics) were analyzed. More precisely, interviews with experts
from various domains were conducted to gain insights into the use of instruments in
practice as well as to reveal recurring structural elements. In this context, a set of
Instrument
Structure
basic elements frequently used across a variety of instruments were discovered (cf. Table
2.1). These elements include, for example, descriptive elements, like headlines or texts
9
2 Fundamentals
Structuring Elements
1. Blocks Group thematically related elements for better understanding.
2. Embedded Instruments Embed an existing instrument into another one.
Descriptive Elements
3. Headline Introduces the following elements.
4. Text Provides additional information to assist participants.
5. Media Provides additional media information (e.g., images) to assist participants.
Data Collection Elements
6. Question Types
6.1. DropDown Only one item may be selected.
6.2. Single Choice Only one item may be selected.
6.3. Yes No Switch Only one item may be selected.
6.4. Range Multiple items may be selected.
6.5. Multiple Choice Multiple items may be selected.
6.6. Ranking Items may be ordered according own preferences.
6.7. Distribution Points may be spent among available items.
6.8. Slider One value from a predefined range may be selected.
6.9. Freetext Answer using regular text input (text, number, date).
7. Sensor Types
7.1. Camera Take a picture during data collection.
7.2. Microphone Record audio during data collection
7.3. Pulse Sensor Measure the pulse rate during data collection.
Table 2.1: Frequently Used Basic Elements Within Instruments
used to guide participants through the data collection procedure. To collect data from
participants, common question and answer types are used. Moreover, blocks that allow
thematically structuring instruments were found in manifold situations. Furthermore,
blocks may comprise the previously mentioned elements (e.g., headlines, questions).
Especially in medicine and psychology, several considered instruments were hierarchically
structured in order to allow comparing collected data amongst other datasets. Finally,
specific control structures could be discovered that had been used to represent the logic,
like
if-then-else
statements (e.g., “If you answered this question with ’yes’, please
continue with Question 15, otherwise continue below.”) and
loops
(e.g., “Please indicate
the diseases that were diagnosed for all your siblings.”).
Fig. 2.1 illustrates a validated paper-based instrument for detecting risk factors during
pregnancy [
84
], which has been successfully applied in various scenarios. In particular,
data is collected following a standardized procedure, i.e., pregnant women fill in the
questionnaires, while waiting for their consultation with the physician. Fig. 2.1 provides
annotations for the aforementioned descriptive and data collection elements.
Depending on the application scenario an instrument is used, various modes for collecting
Instrument
Modes
data may be applied. On one hand, data is often collected with self-report questionnaires
in medical trials [
20
]. The participant, thereby, processes the instrument autonomously,
with no further assistance from staff members. In corresponding scenarios, additional
information on how to properly process respective instrument must be provided to guide
users through the process of data collection. This may include instructions on how to
navigate within the instrument as well as texts on how to answer specific questions
(e.g., to indicate specific body parts the participant was hurt). On the other hand,
10
2.2 Business Process Management
Media Elements (e.g., Logos)
Headline Elements
Text Elements
Block (e.g., thematically group
Elements)
Question Element
Question Element with
Decision Statement and Loop
Question Element with
Decision Statement
Hierarchically Structured
Instrument (e.g., embedding an
Instrument in another one)
Figure 2.1: Example of a Validated Paper-Based Psychological Instrument [84]
psychological studies often collect data based on interviews, i.e., domain experts interact
with participants to collect required data. Moreover, interviews may have to be flexibly
adapted to the respective participants or their situation (e.g., a participant may want to
talk about drug abuse first, before answering questions on his childhood experiences), or
based on already given answers (e.g., the block dealing with drug abuse may be skipped
if the participant does not take drugs).
Table 2.2 summarizes the findings obtained from the analysis of well-established instru-
ments from various application domains.
2.2 Business Process Management
Contrary to traditional information systems, Process-Aware Information Systems (PAIS)
Process-Aware
Information
System
[
78
] separate process logic from application code. The latter is accomplished by relying
on a process schema, which provides a common interface for executing a large number
Process
Schema
of processes [
79
]. Such a process schema may be defined with different notations (e.g.,
graphs, constraints), each having their own peculiarities and application domains. As
11
2 Fundamentals
Domain
Analyzed Instruments
Mode
Multiple Languages
Multiple Versions
Complex Navigation
Decisions
Loops
Embedded Instruments
Require Sensors
Psychology 11 B H# # H# #
Healthcare 9 B H# H# H#
Logistics 4 S # # H# H# # # H#
Automotive 5 S H# # H#
Finance & Taxes 4 S H# H# # H# #
Education 6 I # H# H# H# # # #
Tourism 5 S H# # # # # #
Retail 4 S # # H# H# # # #
= almost every instrument; H# = some instruments; #= almost no instrument;
Mode: Different modes for answering an instrument (S = mostly Self-Rated; I = mostly
Interviews; B = both).
Multiple Versions: There exist multiple versions of the same instrument at the same time.
Complex Navigation: The instrument comprises complex branching logic based on al-
ready given answers or pre-defined rules.
Embedded Instruments: Instruments are hierarchically structured (e.g., an existing in-
strument is used within another one).
Table 2.2: Findings when analyzing Instruments from various Domains
graph-based modeling notations (e.g.,
BPMN
2.0 or
EPC
) are common in both industry
and science, the figures presented in this thesis are consequently represented with the
Business Process Modeling and Notation (i.e., BPMN 2.0) [60].
A process schema is specified through a process model. In this thesis, a process model
PProcess Model
is represented as a directed, structured graph that consists of a set of nodes
N
and a
set of directed edges
E
connecting them. A node either may represent an activity, or
agateway (e.g.,
AND
,
XOR
) to allow expressing a more complex behavior. Each process
model has exactly one start node and exactly one end node. Furthermore, the process
model needs to be connected; i.e., each node
n
can be reached from the start node;
likewise, the end node can be reached from each node
n
.Data elements
D
correspond to
variables connected to nodes, which may then read or write corresponding values during
process execution. Process models are usually created at design time.Design Time
As a prerequisite, a process model
P
must be block-structured (i.e., well-formed), i.e.,
Block-
Structure
each block spanned by a gateway has a single entry and a single exit point of the same
type (e.g.,
AND
). In general, blocks may be arbitrarily nested, but must not overlap each
other (cf. Fig. 2.2). Note that this structure is similar to the one of XML documents.
Extensible
Markup
Language
12
2.3 Mobile Data Collection
Process
Loop Block AND Block XOR Block
Sequence Block
A
B
C
D E
XOR GatewayAND GatewayStart Node End Node Node / Activity
Control Flow Default Flow
Figure 2.2: Block-Structured Process Model (using the BPMN 2.0 Notation)
During run time, a
PAIS
creates a process instance
I
for each (business) process to be
Run Time
Process
Instance
executed. Such a process instance is then executed according to its predefined process
model
P
as well as a generic set of execution rules [
128
]. The current state of one
particular process instance is expressed through the markings of its nodes (e.g.,
STARTED
,
EXECUTED
) and edges. Furthermore, data element values are stored in log files to properly
reflect the execution history of one particular instance. In general, a
PAIS
is able to
execute multiple instances of different models concurrently.
2.3 Mobile Data Collection
Over the last decades, in many application domains large amounts of data were collected
with paper and pencil. Due to the described drawbacks of this traditional approach,
domain experts crave for more convenient approaches supporting their data collection
procedures in daily life. More specifically, the digitization of an instrument should cover
emerging demands from domain experts, like the support of an expressive navigation
logic or the integration of sensor data (e.g., pulse sensor, GPS location, microphone, or
camera) enhancing the expressiveness of the data collected.
In line with this trend, a multitude of web-based questionnaire applications have been
developed. In particular, some of them specifically aim to support sophisticated data
collection scenarios (e.g., in medical and psychological trials). Although many of these web-
Web
Application
13
2 Fundamentals
based applications have proven their applicability in a variety of application scenarios, they
have been unable to cover all relevant use cases. For example, web-based questionnaires
require a stable Internet connection (i.e., they do not work in offline mode), or might be
unable to interact with external sensors (depending on the features of respective web
browser).
In the early 1990s, [
22
] evaluated the applicability of a handheld computer for data
collection purposes in healthcare scenarios. In particular, a significant decrease in time
needed for collecting data as well a significant increase of data quality could be observed.
In 2004, [
1
] investigated whether or not mobile phones can be used for properly collecting
data from patients on a daily basis. On one hand, SMS messages were used to remind
Short Message
Service
patients about their medication or to ask them about their current status. On the other,
the patients were asked to reply to the questions via SMS as well. As a result, the
applicability of corresponding devices for mobile data collection scenarios on a daily basis
was successfully demonstrated.
The aforementioned approaches have been adopted to smart mobile devices (e.g., smart-
phones or tablets) by enterprises and research projects in the large scale (cf. Chapter 7).
Often, the term Mobile Data Collection describes this behavior. In 2009, [
75
] showed
Mobile Data
Collection that smartphones act as a valuable device for leveraging data collection capabilities. In
recent years, numerous mobile data collection applications were developed that allow
domain experts (e.g., researchers) to collect data in a more convenient fashion.
2.4 End-User Programming
In the era of digitization, more and more software applications will be developed, cus-
tomized, and maintained by non-professional programmers. According to [
91
], approxi-
mately 90 million US citizens perform basic programming tasks in their job. In turn, [
41
]
distinguishes between professional and non-professional programmers depending on their
intention and motivation. For example, code written by non-professional programmers
might not meet high quality standards, but is of a rather opportunistic nature. Such
individuals are called end-users in the following.End-Users
The term End-User Programming (
EUP
)summarizes sophisticated techniques that
End-User
Programming
shall empower end-users with little (or no) programming knowledge to develop their
own software applications. Respective approaches allow end-users to perform complex
operations on their data, to run jobs in an easy-to-understand manner, or to design
sophisticated user interfaces without need to learn any complex programming language.
Examples of such end-user developed applications range from simple Wiki applications,
which enable end-users to document tasks, to simplified database query languages
(e.g., Query by Example applications), to sophisticated 3D modeling applications [
67
].
14
2.4 End-User Programming
According to [
90
], the most popular end-user programming approach are spreadsheet-
like applications, which are frequently used in enterprises to automate tasks or raise
productivity (e.g., [121]).
Over time, a variety of end-user programming approaches, each providing their own
(graphical) language, have emerged. Graphical notations rely on visual elements that
may be interlinked by end-users to create a software application. Such elements represent
specific code constructs (e.g.,
loops
or
if-then-else
statements) or basic functions
(e.g.,
sum()
,
concatenate()
, or
append()
) that may be flexibly composed. The textual
representation of the source code is hidden from users. Sophisticated wizards guide
untrained users to reduce complexity and, hence, errors. Experiments conducted with
pupils compared graphical programming approaches with common textual ones. Teachers
reported that the graphical representation significantly improved the understanding of
program code [
39
]. Domain-Specific Languages (
DSL
s), in turn, try to abstract from
Domain-
Specific
Language
complex programming languages and provide a rather limited, but easy-to-use syntax
for end-users. Common examples are query languages for databases that are mapped to
SQL
. Experts from specific domains may use their domain-specific terminology instead
Structured
Query
Language
of a generic terminology as known from common programming languages.
In a broader scope, End-User Development (
EUD
)covers additional phases of the software
End-User
Development
development lifecycle (e.g., requirements engineering, testing or documentation).
Although end-user programming approaches shall enable non-professional programmers
to develop applications themselves to a certain extent or to adapt existing ones to
their specific needs, there also exists criticism for respective approaches [
29
]. This
includes arguments like outsourcing efforts to end-users instead of properly paying
skilled application developers or security concerns regarding applications developed by
non-programmers.
From a business perspective, end-user programming shows significant benefits. On one
hand, end-user programming approaches contribute to reduce the Business-IT alignment
gap (i.e., domain experts are unable to describe what a developer shall realize) as domain
experts are empowered to actively participate in the process of developing software
applications [
12
]. On the other, minor changes on an application can be accomplished by
domain experts themselves (e.g., by adapting configuration files), relieving application
developers from costly and time-consuming code adaptations.
15
3
Development Strategies
This chapter summarizes existing approaches for developing mobile data collection
applications. An overview is provided by Fig. 3.1.
When developing mobile applications that support domain experts in collecting large
amounts of data in a convenient fashion, various strategies can be applied. The latter, in
turn, need to meet the requirements raised in the various application domains (e.g., in
healthcare “the data needs to be stored in a secure way and must not be accessible for
unauthorized users”.)
a
. Furthermore, the requirements specific to a given setting must
be met (e.g., the application must not rely on a stable Internet connection)
b
. Finally,
requirements related to technical aspects (e.g., the mobile platform or the integration of
external sensors to collect additional data) c
need to be covered.
Three core development strategies exist: First, one may develop a specifically tailored
mobile application
1
that meets the given requirements. Accordingly, the structure and
logic of an instrument is directly translated to user interface elements of the respective
mobile platform. To relieve application developers from implementing the same applica-
tion for multiple platforms (e.g., Android and iOS may need to be covered properly),
cross-platform development frameworks may be used [
106
]. Corresponding frameworks
usually rely on web technologies (e.g.,
HTML
,
CSS
and JavaScript) that are executed
within a web container. This container provides native APIs to interact with the smart
mobile device (e.g., access the camera of the device). Changes of an instrument (e.g., add
or remove elements) result in code adaptations and, consequently, the new application
needs to be deployed to and installed on the respective smart mobile devices. Finally,
a sophisticated release management is required if several versions of an application are
available concurrently.
17
3 Development Strategies
Domain-Specific
Requirements
Scenario-Specific
Requirements Technical Requirements
Mobile Data Collection and Sensing
Application
Implementation Technique Responsive Web Page Instrument Representation
Native Web
Application
Hybrid
Application HTML Client REST Client Individual
Realization
Interpreted
Model
UML Model Process
Model
Proprietary
Model
e.g., Android, iOS, ... e.g., Cordova, Ionic, ...
e.g., Psychology, Healthcare,
Logistics, ...
e.g., Sensors, Offline
Functionality, Rural Areas, ...
e.g., Mobile Platform,
Sensors, Target Device
a b c
123
Figure 3.1: Strategies for Developing Mobile Data Collection Applications
As second strategy, one can develop a web page based on responsive web design practices
2
to implement the instrument. Such techniques, in turn, allow quickly develop applications
that may adapt the user interface of web pages according to the device. More specifically,
CSS
media queries may be used to switch styles accordingly [
11
] and present content
adequately on smart mobile devices. As responsive web pages are accessible on a desktop
computer, the instruments are available on all platforms being able to run modern web
browsers. Instead of
HTML
, one may return a structured representation of an instrument.
For web applications,
JSON
is commonly used as a lightweight data exchange format,
JavaScript
Object
Notation
which is easy to understand and use in the context of JavaScript applications [
59
]. Data
provided by a RESTful server may be consumed and processed by a modern JavaScript
Representational
State Transfer
application. However, the participant interacting with the instrument still uses a web
application with all its limitations.
The third strategy represents an instrument in terms of specific models
3
. The Ques-
tionSys framework developed in this thesis applies techniques from this category. More
specifically, process models are used to describe the structure and logic of a data collection
Process Model
instrument. Note that other models may be used for this purpose as well. For example,
UML
provides various (graphical) notations to describe the structure of an application as
Unified
Modeling
Language
well as its control flow within specific components. Generators, in turn, may automatically
transform the latter to code fragments, which can then be executed on corresponding
devices. Finally, other (proprietary) models may be used in this context as well.
18
Part II
The QuestionSys Framework
19
Core contributions presented in this part were mainly published in the following articles:
SPSR16
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Lightweight
Process Engine for Enabling Advanced Mobile Applications. In 24th Int’l
Conf on Cooperative Information Systems (CoopIS), number 10033 in
LNCS, pages 552–569. Springer, October 2016
SPSR16a
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Configurator
Component for End-User Defined Mobile Data Collection Processes.
In Demo Track of the 14th Int’l Conf on Service Oriented Computing
(ICSOC), October 2016
SPWSR16
J. Schobel, R. Pryss, W. Wipp, M. Schickler, and M. Reichert. A Mobile
Service Engine Enabling Complex Data Collection Applications. In 14th
Int’l Conf on Service Oriented Computing (ICSOC), number 9936 in
LNCS, pages 626–633, October 2016
SPSRER16
J. Schobel, R. Pryss, M. Schickler, M. Ruf-Leuschner, T. Elbert, and
M. Reichert. End-User Programming of Mobile Services: Empowering
Domain Experts to Implement Mobile Data Collection Applications. In
5th IEEE Int’l Conf on Mobile Services (MS), pages 1–8. IEEE Computer
Society Press, May 2016
The original articles are added to the Appendix of this thesis.
21
4
Concept
The main goal of the QuestionSys framework is to enable domain experts (e.g., medical
doctors or psychologists), having no programming skills, to develop sophisticated mobile
applications on their own. These applications, in turn, may then be used to collect data
in large-scale scenarios like clinical trials or psychological studies [
114
]. Moreover, the
deployment and execution of these applications on smart mobile devices shall be possible
without help of any IT expert. On one hand, this shall reduce the costs and time required
to develop and deploy data collection applications. On the other, the quality of collected
data shall be increased.
Fig. 4.1 sketches the overall approach the QuestionSys framework realizes. A sophisticated
configurator is provided that relies on end-user programming techniques to guide untrained
End-User
Programming
users through the process of creating mobile data collection applications
1
. To properly
support untrained experts, a Domain-Specific Language for modeling instruments is
Domain-
Specific
Language
introduced, hiding most of the complexity emerging during the development process and
allowing for an easy-to-use approach.
Each instrument modeled with the configurator, in turn, is mapped to a process model.
Process Model
As described earlier, process management technology acts as a solid technical foundation
for the QuestionSys framework in general
2
. On one hand, a process model specifies
the control flow of the instrument (e.g., the order and constraints for processing the
questions of an instrument). On the other, the process model enables a platform-
independent representation of the instrument, e.g., no information on the user interface
or the formatting of the instrument are included. Note that this approach is similar to
the ones proposed in the field of model-driven development.Model-Driven
Development
23
4 Concept
Alcohol
Consumption
Cigarette
Consumption
StartFlow
Activity XORjoin
DataElement
WriteAccess
ReadAccess
EndFlow
ET_ControlFlow_Default
ET_DataFlow
AlcoholCigarettes
(Cigarettes = yes)
AND (Alcohol = yes)
XORsplit
else
(Cigarettes = yes) AND
(Alcohol = no)
ET_ControlFlow
Cigarettes
& Alcohol
Page
Intro
Page
General EndCigarettes
Process-Centric
Instrument Logic
1) Configurator 3) Process-Driven Mobile
Data Collection Application
2) Process Model
Lightweight Process Engine for
Process Execution and Monitoring
Sensor Framework To
Integrate Hardware
UI Generator with
Custom Control Elements
Create Mobile Data Collection
Instrument through End-User
Programming
Figure 4.1: Model-Driven Development of Mobile Data Collection Applications
The derived process model may then be deployed to smart mobile devices. The latter
comprise a run time environment (i.e., a process engine) that allows for the robust
Process Engine
enactment of data collection instruments
3
. The run time environment may further
provide application-specific features, like the integration of sensors or the on-demand
evaluation of collected datasets.
4.1 Requirements
In the context of this thesis, more than 50 established instruments from different domains
were analyzed. This includes, for example, instruments from psychology (e.g., mental
issues, drug abuse), healthcare (e.g., patient information), automotive industry (e.g.,
TÜV
1
vehicle inspection), and finance (e.g., tax declaration). Their analysis revealed
requirements that are fundamental for the digitization of any instrument [
107
] and thus
need to be properly covered by mobile data collection applications.
Depending on the scenario in which a mobile data collection application shall be applied,
it might be crucial to enable its offline use as well. For example, when supporting
researchers in collecting data related to
PTSD
in rural areas in Africa no stable Internet
Post-
Traumatic
Stress Disorder
connection was available [
108
]. The psychological instruments used in this scenario as
well as the data collected, had to be stored in a secure (e.g., encrypted) way on the smart
mobile device. Furthermore, instruments may be processed using different presentation
modes (e.g., interview vs. self-rating), having an impact on how questions are presented
to participants, answers are selected, or the instrument itself is executed (e.g., if the
participants are allowed to navigate within the instrument). Moreover, participants
1
Technischer Überwachungsverein (Technical Inspection Association), a company that provides inspection
and product certification services.
24
4.2 Mobile Data Collection Lifecycle
should be enabled to select the language the instrument is presented, to one that suits
them best. Collected data, in turn, shall be automatically evaluated based on predefined
rules. To only grant authorized users access to the collected data as well as the rules
applied to the latter, a complex user management needs to be provided by the mobile
application. Moreover, domain experts without programming skills shall be able to
flexibly adapt the structure of an instrument (e.g., by adding questions or changing
labels). In this context, it is crucial to always ensure the validity of the instrument at
all time. Finally, the developed mobile data collection instruments shall be available on
both prevailing mobile platforms (i.e., Android and iOS).
Obviously, the presented requirements adhere to different phases of the data collection
procedure.
4.2 Mobile Data Collection Lifecycle
Insights from realizing long-running, real-world mobile data collection applications re-
vealed a generic lifecycle. In order to properly assist the experts in collecting data in
their specific application scenarios, the QuestionSys framework covers the entire Mobile
Lifecycle
Data Collection Lifecycle. The latter consists of 5 phases as illustrated in Fig. 4.2.
Archiving &
Versioning
Monitoring
& Analysis
Enactment &
Execution
Deployment
Design &
Modeling
Mobile Data
Collection Lifecycle
Domain Specific Requirements
Execution & Monitoring
End-User Programming
Figure 4.2: Mobile Data Collection Lifecycle
In the Design & Modeling phase, mobile data collection instruments with complex
navigation logic are created by domain experts. The Deployment phase, in turn, allows
25
4 Concept
for a secure and robust deployment of the created instruments to smart mobile devices.
During the Enactment & Execution phase, multiple instances of the previously deployed
data collection instrument may be created and executed on the smart mobile devices
to collect data in a convenient fashion. The Monitoring & Analysis phase deals with
the real-time analysis of the data collected on the smart mobile device. Finally, the
Archiving & Versioning phase provides sophisticated techniques for managing different
releases of a modeled data collection instrument as well as for archiving the collected
data. In order to adequately support domain experts in modeling sophisticated mobile
data collection instruments as well as to meet domain-specific requirements, end-user
programming techniques are applied in certain phases of the lifecycle.
4.3 Model-Driven Development of Instruments
To address the discussed requirements and to provide an approach supporting the entire
lifecycle of an instrument, the QuestionSys framework follows a model-driven approach,
Model-Driven
Development
which consists of four models (cf. Fig. 4.3). Thereby, a model is continuously transformed
into another one by enriching the former with additional information. Furthermore,
platform-specific information may be added to tailor models to specific operating systems
or execution environments. Finally, code is automatically derived from this information
[10]. Often, UML is used in this context as a vendor-neutral standard [86].Unified
Modeling
Language
Computation
Independent Model
Platform Independent
Model
Platform Specific
Model Code Model
MDD
Instrument Process Model Process Model &
Executable Components
Process Engine, Process
Model & Executable
Components
QuestionSys
transform transform transform
mapped to
Figure 4.3: Aligning the QuestionSys Approach to Model-Driven Development
As discussed, the QuestionSys framework describes the logic of a data collection instrument
in terms of a process model (cf. Fig. 4.4), which is enacted by a lightweight process
engine running on heterogeneous smart mobile devices [112,115].
Further, QuestionSys allows mapping instruments to executable process models. More
specifically, the content and logic of paper-based instruments can be mapped to a process
model. In more detail, pages of an instrument correspond to activities within the process,
whereas the flow between activities matches the navigation logic of the instrument.
Questions are mapped to data elements, which, in turn, are connected via
WRITE
data
edges to respective pages. A data element stores an answer given during the execution of
the model on the smart mobile device. Sophisticated navigation logic can be specified
using gateways (i.e.,
READ
specific data elements). Fig. 4.5 illustrates the described
26
4.3 Model-Driven Development of Instruments
Figure 4.4: A Psychological Instrument Represented as BPMN 2.0 Model
mapping from a paper-based instrument to a process model, and the subsequent execution
of this model by a lightweight process engine running on a smart mobile device. Currently,
the QuestionSys framework relies on
ADEPT
2 [
42
,
77
], but can be easily adapted to
other meta-models (e.g., WS-BPEL [127]) as well. Business
Process
Execution
Language
Applying an established process modeling notation (e.g.,
BPMN
2.0 or
EPC
) for specifying
the logic of an instrument has proven to be useful. In general, however, aspects other
than the control flow need to be considered as well [
89
], e.g., the data flow [
87
], resource
[
88
], and time [
44
,
46
] perspectives. When applying graphical notations for modeling
data collection procedures, additional issues emerged. For example, domain experts
were overwhelmed by the multitude of graphical elements (as well as their semantical
meaning) to properly represent their data collection instrument [
136
], e.g., dealing with
data elements and corresponding data flow was especially challenging for non-modeling
experts.
To address these revealed issues, a novel graphical domain-specific language is proposed.
Domain-
Specific
Language
The latter relies on concepts known from
BPMN
2.0, but omits elements not needed in data
collection scenarios (e.g., temporal constraints [
45
] or business events [
17
]). Furthermore,
the modeling of data elements has been simplified; i.e., data flow is implicitly modeled.
The QuestionSys framework provides a sophisticated configurator relying on this graphical
notation to enable domain experts to model instruments in a convenient approach. This
graphical notation, in turn, needs to be evaluated regarding its usability, especially when
being used by domain experts having no experience with process modeling notations.
Chapter 8 presents results from extensive studies addressing such usability aspects.
27
4 Concept
Process Model Alcohol
Consumption
Cigarette
Consumption
StartFlow
Activity
XORjoin
DataElement
WriteAccess
ReadAccess
EndFlow
ET_ControlFlow_Default
ET_DataFlow
AlcoholCigarettes
(Cigarettes = yes)
AND (Alcohol = yes)
XORsplit
else
(Cigarettes = yes)
AND (Alcohol = no)
ET_ControlFlow
Cigarettes
& Alcohol
Page
Intro
Page
General EndCigarettes
Mapping an Instrument to a Process Model
Questionnaire
Model Page Question
Process
Model
Process
Activity
Process
Data Element
Questionnaire
Instance
Process
Instance
maps to
n 1 1 n n n
n n1 nn 1
maps to
maps to
maps to
Navigation Operation Based
on Already Given Answers
Mobile Data Collection Application
Figure 4.5:
Mapping an Instrument to a Process Model and its Execution on a Smart
Mobile Device
28
5
Architecture
This chapter presents the architecture of the QuestionSys framework and describes
selected components (e.g., the configurator and mobile client application). Further, it
presents the procedure of modeling, deploying and executing an instrument with the
QuestionSys framework.
5.1 Architecture
The architecture of the QuestionSys framework applies a process-driven approach, i.e., it
is driven by process management technology (cf. Fig. 5.1). Business
Process
Management
Creating data collection instruments based on process management techno-
logy:
A domain expert can create a data collection instruments using a process-
aware configurator
a
[
113
]. The latter, in turn, provides a domain-specific modeling
language for graphically creating instruments. Furthermore, the configurator allows
defining rules for automatically evaluating the collected data
b
(e.g., to calculate
the body mass index of participants).
Generating mobile applications based on process models:
The process model
(i.e., the modeled data collection instrument) acts as a schema for the subsequent
execution of corresponding instances by a lightweight mobile process engine that
runs on various mobile operation systems [
115
]. By interpreting such models directly
on the smart mobile device, changes to an instrument can be realized in a cost- and
time-efficient manner. A sophisticated rendering algorithm not only takes different
29
5 Architecture
Process-Aware Instrument Configurator Flexible Mobile Data Collection Clients
Cigarettes Consumption
How many Cigarettes do you smoke each
day?
Do you smoke in your flat?
yes
no
XML
Web Service & Database
Execution Log
Files (XML)
alc = yes
age = 16
cigarettes
= no
v = 6
w = yes
x = no
y = 10
z = 4
alc = yes
age = 16
cigarettes
= no
v = 6
w = yes
x = no
y = 10
z = 4
alc = yes
age = 16
cig. = no
v = 6
w = yes
x = no
y = 10
z = 4
Domain Expert
e.g., Analyst
Domain Expert
e.g., Interviewer
Participant
e.g., Study Subject
Process-Aware Data Evaluation
Domain Expert
e.g., Study Director
Underage Alcohol Usage:
(age < 18) && (alc. = true)
Underage Alcohol Usage
< =
age 18 alc. true
Anonymized
Execution Log
Files (XML)
alc = yes
age = 16
cigarettes
= no
v = 6
w = yes
x = no
y = 10
z = 4
alc = yes
age = 16
cigarettes
= no
v = 6
w = yes
x = no
y = 10
z = 4
alc = yes
age = 16
cig. = no
v = 6
w = yes
x = no
y = 10
z = 4
1
2
3
4
5
Integrate Domain
Experts
Create Collection Instruments Using
Process Management Technology
Relieve IT Experts Through
Automatic Process Management Generate Mobile Applications Based On Process Models
PROCESS-DRIVEN
a
b
Cigarettes
Consumption
How many Cigarettes
do you smoke each
day?
Do you smoke in your
flat?
yes
no
Cigarettes Consumption
How many Cigarettes do you smoke
each day?
Do you smoke in your flat?
yes
no
Figure 5.1: The QuestionSys Architecture
mobile operating systems into account, but also device properties (e.g., screen size)
and the language to be displayed.
Relieving IT experts through automatic process management:
The created
instrument (i.e., the process model along with its rules for evaluating the data
collected) is mapped to an XML document, which can then be deployed to available
Extensible
Markup
Language
smart mobile devices. During the execution of instrument instances, collected
data is stored directly on the smart mobile device. The QuestionSys architecture
relies on RESTful Web Services [
21
] for realizing the communication between the
Representational
State Transfer components 1
–5
.
The architecture in general and the procedure of deploying instruments to smart mobile
devices, help mitigating several issues known from data collection projects [
37
]. Releasing
a new version of an already existing data collection instrument, for example, does not
require the involvement of mobile application developers anymore. Instead, domain
experts themselves may apply the desired changes to the model and deploy its new
version on a server component. Mobile data collection clients, in turn, can then download
the new version and enact further instrument instances based on the new model.
In the following, core components of the QuestionSys framework are discussed in detail.
Specifically, the configurator used for creating data collection instruments as well as the
mobile client used for executing the latter are presented.
30
5.2 Process-Aware Instrument Configurator
5.2 Process-Aware Instrument Configurator
The developed QuestionSys configurator allows domain experts with little or no program-
ming skills or knowledge in process modeling to flexibly create data collection instruments
on their own. More specifically, it applies process management concepts and technologies
Business
Process
Management
in a broader scope [
85
]. In this thesis, only the most relevant aspects of the configurator
are presented (see [113] for an in-depth description of details).
Show all available
Questionnaires Select question element
Select different
types of elements
Combine elements to pages
using drag and drop operations
Provide an interactive live
preview of elements
Manage details of
selected elements
Select page element
Figure 5.2: The QuestionSys Configurator
Figs. 5.2 and 5.3 show the user interface of the configurator. Fig. 5.2 depicts the
Element and Page Repository, which enables domain experts to create basic elements of
an instrument (e.g., headlines or questions). The latter may be further customized using
an editor being able to handle multiple languages as well as to track different revisions
of elements. Using drag and drop operations, elements may be combined to pages (i.e.,
one screen in the resulting mobile application). An interactive live preview, which allows
simulating different mobile devices, provides immediate feedback to domain experts.
Fig. 5.3 illustrates the Modeling Area where the data collection instrument under
development is visualized. More precisely, previously created pages may be dragged
to the graph in the middle of the screen to define the structure and navigation flow of
an instrument. In order to properly support untrained domain experts, a correctness-
by-construction principle [
16
] is applied. The latter ensures that only currently valid
operations can be applied to the model. Moreover, the model can be executed by the
lightweight process engine at any time. Finally, a specifically developed domain-specific
modeling notation is applied in order to simplify the modeling process.
31
5 Architecture
Export model in different
formats (e.g., process
models, PDF documents, …)
Select page elements
Provide detailed
error messages if
model is not correct
Provide an interactive live
preview of a page
Specify a device, orientation
and language for live preview
Design complex navigation
operations with multiple branches
references
references
Figure 5.3: The QuestionSys Configurator
5.3 Flexible Mobile Data Collection Client
The mobile data collection application is a client that runs on multiple operating sys-
tems (i.e., Android or iOS). This application consists of basic interfaces for managing
instruments, e.g., to download them from a server. Furthermore, it provides features for
executing an instrument and, hence, collecting data based on the created model. In this
thesis, a lightweight process engine running on smart mobile devices was developed. In
particular, this engine enables a robust execution of data collection instruments, while
providing the flexibility to enhance and adapt the provided functionality.
The lightweight process engine is designed as a service that may be embedded in other
mobile applications as well. It provides a high-level communication interface to create
new instances based on a given process model and to execute them according the specified
logic of the instrument. The communication and data flow between the mobile data
collection application and the process engine is illustrated in Fig. 5.4.
It is noteworthy that the process engine only handles the execution of the model itself.
For example, it evaluates the nodes (i.e., pages) to be processed next or reads data from
respective data elements and passes it to the node. However, the lightweight process
engine is not responsible for processing the actual content of the page (i.e., for displaying
the user interface to collect data). The QuestionSys framework defines and provides
Executable Components (
EC
s) for the latter purpose. An
EC
can be seen as a micro
Executable
Component
32
5.3 Flexible Mobile Data Collection Client
Smart Mobile Application
User Interface
Controller
Model
Libraries
Mobile Process Engine
Node
...
Activity
Node
Activity
Runtime Manager
Instance Manager
Execution Manager
EC User Interface
Resources
LayoutsColors Languages
Executable
Component 2 (EC)
User Interface
Controller
Model
Interface
1
2
3
4
5
Executable
Component 1 (EC)
User Interface
Controller
Model
Interface
Resources / Libraries
Figure 5.4: The QuestionSys Mobile Data Collection Client
service [
58
] which is self-contained, providing a limited functionality (e.g., rendering the
specified user interface in order to collect data).
All available
ECs
are coordinated by the lightweight process engine. As depicted in
Fig. 5.4,
ECs
are not part of the mobile data collection client itself, but are installed
as separate applications on the smart mobile device. For example, the Page
EC
, which
provides basic presentation logic, may be adapted independently from the mobile data
collection application or process engine. Furthermore, an
EC
may be replaced by another
one providing more functionality (e.g., to collect vital parameters via external sensors)
or by changing the overall style of the user interface. Note that such a feature may
be crucial depending on the context and application scenario in which the mobile data
collection application is used. Altogether, this approach fosters the separation of duties
and provides an easy-to-extend approach for mobile application development.
Note that the
ECs
and process engine communicate through well-defined interfaces. For
example, the process engine may query the current status of an
EC
(e.g., to check whether
all mandatory fields are filled in). Likewise, the
EC
may notify the engine if errors occur
that need to be handled.
Fig. 5.5 shows the user interface of the developed mobile data collection client. Fur-
thermore, it illustrates how specific parts of the user interface are rendered by different
parts of the mobile data collection application (i.e., the main application and the
ECs
33
5 Architecture
UI Fragment Created by Mobile
Data Collection Application
UI Fragment Depending on
State of Executable Component
UI Fragment Created by Executable
Component and Mobile Context
Figure 5.5: User Interface of the Mobile Data Collection Client
respectively). The floating action button to proceed to the next page of the instrument
(cf. Fig. 5.5; rightmost part), for example, is only rendered if a corresponding
EC
notifies
the process engine upon its completion. For the end-user interacting with the mobile
data collection application, however, everything is combined into one interface, enabling
a consistent user experience.
34
6
Discussion
This chapter discusses key aspects of the QuestionSys framework and relates them to
requirements gathered in case studies from various domains.
The fundamental goal of the QuestionSys framework is to empower domain experts having
no programming expertise to develop mobile applications for data collection purposes
themselves. For this purpose, a sophisticated instrument configurator (cf. Fig. 6.1,
a
)
was developed, which applies techniques known from end-user programming to guide
domain experts through the process of modeling instruments.
When deploying an instrument, it is automatically transformed into a process model
(cf. Fig. 6.1,
b
). This process model expresses the complex navigation logic of the
instrument (e.g., via gateways), depending on the needs of the considered application.
Furthermore, the process model specifies the data flow within the instrument (e.g., the
data collected on a specific page of an instrument). To reduce the number of available
graphical elements for modeling as well as the overall complexity during the modeling
procedure, QuestionSys uses its own domain-specific modeling language.
After downloading an instrument to a smart mobile device (cf. Fig. 6.1,
c
) it can
be executed with the lightweight process engine. This engine, in turn, is capable of
dynamically executing instances of the instrument based on its process model in order
to collect data in a convenient fashion. The concept of
ECs
allows flexibly adapting
the provided functionality of the developed mobile data collection application. More
specifically,
EC
s may be customized depending on the considered application scenario.
Thereby, not only the visual appearance (e.g., colors) of the application may be changed,
but also new control elements be introduced. Fig. 6.2 shows a specifically developed
35
6 Discussion
Process Model
Drives
Mobile
Application
Logic
ExportExport
Is Mapped
to a
Modeling Area View Page Repository View
Element View
Preview Mode
Use OS Independent
Export Format
Select Various
Question Types
Provide
Multilingualism
Get On Demand
Preview of Elements
Provide UI
Generator
Custom UI Elements for
Easy & Intuitive Interaction
Combining Process Technology with End-User Programming
Changes to the Model are Directly
Propagated to the Smart Mobile Device
→ No Programming Skills Needed
Executing Process Model
I) Manually Created II) Automatically Generated, Manually Executed
a
b
c
Drag & Drop Pages for Modeling
the Data Collection Instrument
Specify Branch
Parameters
Select Different
Versions of Elements
Model Complex Navigation
Logic using Decision Elements
Show Page Containing
Elements of Different Types
Figure 6.1:
Empowering Domain Experts Developing their own Mobile Data Collection
Applications
control element Button Bar that allows selecting appropriate values. This selection, in
turn, is then displayed directly on the bar (i.e., the dark gray overlays).
Figure 6.2: Implementing Custom UI Controls for Collecting Data
Note that changes applied to the instrument solely affect the model. QuestionSys relies
on high-level change patterns describing frequently required operations, like inserting a
new page to the instrument or moving elements within a page (cf. Table 6.1). These
patterns support domain experts on a more abstract level to perform required adaptations
to instruments. Further, they assure that the model is kept in a correct state, e.g., if a
SPLIT
gateway is added, the corresponding
JOIN
gateway is simultaneously added. After
36
downloading the new version of the instrument, the adaptations become immediately
available on respective smart mobile devices. Based on this approach, changes to data
collection instruments no longer require the costly involvement of IT experts in the
process of developing sophisticated data collection instruments.
Name Insert Block. Add a new block to an existing instrument. Available types are IF,ALL, and
REPEAT.
Signature insertBlock(type, before, after)
Example Depending on the type of the block, various scenarios are possible:
•IF blocks solely select one path based on already given answers during run time.
•ALL blocks select all paths to be executed, however, the person interacting with the
smart mobile device may choose its order of execution.
•REPEAT blocks allow for repeating the content of the block multiple times. The
amount of repetitions may be determined at run time (e.g., based on given answers)
or are pre-defined by the domain expert (e.g., ntimes).
A B A B
Pre-
Condition
The position to insert the block must be exactly specified; i.e., after must directly follow
before.
Post-
Condition
An empty block comprising a split and join gateway that are directly connected is
inserted; For IF and REPEAT blocks, data elements for evaluating the conditions need to be
connected using READ data edges.
Table 6.1: Example of a Change Pattern
The change patterns were evaluated in a usability study with 111 participants from various
domains. During this evaluation, participants had to work with an early prototype
application to apply change patterns to an existing data collection instrument (e.g.,
change the order of pages or insert new elements at specific positions).
Figs. 6.3, 6.4, and 6.5 provide insights of this study. More specifically, Fig. 6.3 evaluates
the overall complexity of modeling data collection instruments. Fig. 6.4, in turn,
illustrates the perceived complexity when applying specific change requests. Interestingly,
most of the participants rated the modeling concept as easy or even better. When it
comes to the modeling of complex navigation operations, however, participants reported
problems (cf. Fig. 6.5). As illustrated, the perceived mental effort was considerable
high for the majority of the participants. Consequently, the QuestionSys approach
for modeling the navigation logic of an instrument (i.e., decisions) was refined in later
versions of the framework. Finally, the evaluation compared the submitted models with
a reference model designed by experts. Approximately 81% of the provided models were
sound, whereas 64% of the Psychologists submitted correct models.
The presented QuestionSys approach guarantees flexibility along the various phases of
the mobile data collection lifecycle [
111
]. In particular, through the combined use of
37
6 Discussion
0%
10%
20%
30%
40%
50%
60%
very hard hard normal easy very easy
Overall
Computer Scientists
Psychologists
n = 111
Figure 6.3: Perceived Mental Effort during Modeling
0%
10%
20%
30%
40%
50%
very bad bad neutral good very good
Inserting an Element
Inserting a Page
Moving an Element
Moving a Page
n = 111
Figure 6.4: Perceived Mental Effort when using Basic Operations
0%
5%
10%
15%
20%
25%
30%
35%
very bad bad neutral good very good
Overall
Computer Scientists
Psychologists
n = 111
Figure 6.5: Perceived Mental Effort when using Complex Navigation Operations
well-established and standardized technologies, domain experts are empowered to create
mobile data collection instruments without the involvement of any IT experts (cf. Table
6.2). Finally, requirements that were elaborated by
1
analyzing instruments from various
domains,
2
conducting multiple interviews with domain experts, and
3
implementing
mobile data collection applications for various scenarios are adequately addressed with
the presented approach and corresponding architecture.
38
Custom Modeling Language
End-User Programming
Process Model
Process Engine
Executable Components
Sensor Framework
Process Mining
Model-Based Approach • • • • •
Complex Navigation • • • • •
Different Releases • •
Flexible & Robust Execution • • •
Monitoring & Analysis • • • •
Sensors • • • •
Multilingualism • • •
UI Generator •
Domain-Specific Requirements • • •
Evolution of Instruments • • • • • •
Table 6.2:
Combining Technologies Enabling Flexibility in Mobile Data Collection Sce-
narios
39
7
Related Approaches
This section discusses approaches from research and industry that are related to the
QuestionSys framework. As there exists a plethora of such applications, only the most
relevant approaches are discussed.
7.1 Research Approaches
In [
32
], a rather generic approach for developing applications that run on mobile operating
systems is discussed. This approach allows mobile application developers to describe
their application scenario, entities, and mobile device features with a meta-programming
language. The models are then translated into native application code for iOS and
Android. Furthermore, the approach automatically generates
REST
ful code for a server
backend allowing mobile applications to store respective data online. As opposed to
QuestionSys, most configuration is done in a textual way, neglecting the advantages
of graphical notations. Based on these insights, the approach specifically targets at
individuals being experienced with mobile application development.
Based on prior work, [
81
] describes a process-oriented approach for creating mobile
business applications. In particular, the authors illustrate how their models can be
transformed into application code running on different mobile operating systems. However,
the graphically specified application logic is completely transformed to native application
code, resulting in a specifically tailored mobile application. As a consequence, many
promising key features known from process management technology research are not
available. As opposed to the QuestionSys approach, these applications are limited with
41
7 Related Approaches
respect to the provided instrument features. In particular, advanced features like the
process-driven navigation logic (e.g., influencing the further course of the instrument
based on already given answers) are not provided.
A more sophisticated
WYSIWYG
editor for developing smart mobile applications in
What You See
Is What You
Get
general is presented in [
4
]. Similar to QuestionSys, it relies on a model-driven approach
that uses its own domain-specific language. These models are then transformed into
native application code that runs on smart mobile devices. This approach, however,
specifically targets at mobile application developers who shall be relieved from complex
programming tasks. As a drawback, this approach does not involve domain experts.
An approach involving end-users is proposed in [
38
]: medical staff is empowered to model
care plans for chronically ill patients. These plans are then automatically transformed
into
DHTML
applications and deployed to smart mobile devices. This way, specifically
Dynamic
HTML
tailored mobile applications for individuals can be created. As opposed to the QuestionSys
framework, however, the approach presented by the authors is domain-specific, i.e., its
application scenario is limited to care plans.
A noteworthy approach is illustrated in [
131
,
132
]. WordPress, a well-known blogging
software, is combined with iBuildApp, a Web-based application builder, in order to
create a platform supporting students from clinical psychiatry. The platform focuses
on information retrieval for users (e.g., provide psychiatric guidelines). However, it also
provides limited support for developing digital questionnaires to allow students to check
their knowledge.
The research projects Manage My Pain [
76
], TrackYourTinnitus [
64
], and PsychLog [
24
]
apply crowdsensing techniques for collecting vital healthcare data on a regular basis. These
projects developed smart mobile applications for convenient data collection purposes.
Compared to the QuestionSys approach, they only provide rudimentary configuration
possibilities for instruments. In this context, PsychLog offers more sophisticated features
for configuring data collection studies. This includes, for example, time-based triggers or
instruments that may affect (e.g., exclude) each other.
7.2 Products
Along the trend of no-code (or low-code) approaches, which enable individuals to develop
specific (mobile) applications, a number of software products emerged. For example,
WebRatio [
8
], Mendix [
33
], or OutSystems provide model-driven platforms supporting
non-developers in creating mobile applications. These models are then deployed to web
platforms and smart mobile devices respectively. The applications, in turn, are created
with common web technologies (e.g.,
HTML
5, JavaScript,
CSS
3), which are rendered in
Hypertext
Markup
Language
Cascading
Style Sheets
a web browser. In consequence, only functionality available in modern web browsers can
be used.
42
7.2 Products
BuildFire
1
positions itself as a rapid development framework, which provides a sophis-
ticated
WYSIWYG
editor to visually define mobile applications. Although the focus
What You See
Is What You
Get
of this framework is not set on data collection in general, it still may be used for this
purpose as well. Depending on the considered application scenarios the available user
interface elements may have restricted functionality. More complex navigation logic for
electronic forms requires adaptations on code level. By contrast, Bubble
2
uses end-user
programming techniques to enable non-programmers to develop mobile applications. In
particular, Bubble relies on a workflow-based programming language, which may be used
by non-programmers. However, generating mobile applications based on this approach is
in a premature stadium.
In the field of online questionnaire applications, a plethora of applications exist. Examples
include LimeSurvey
3
,SurveyMonkey
4
,Qualtrics
5
, and SmartSurvey
6
. These applications
provide configurators that allow designing online surveys. However, the vast majority
of applications rely on form-based editors, which are limited with respect to the design
of navigation logic. There is no graphical modeling approach as provided by this thesis.
Furthermore, questionnaires are usually filled in with a web browser. These web pages,
in turn, are developed with responsive web technologies [
23
]; i.e., the user interface is
properly adjusted depending on the respective device. Therefore, solely features provided
by the web browser are supported.
MovisensXS, an online application targeting at ambulatory assessments, allows researchers
to create instruments with a form-based editor. These instruments may then be executed
on smart mobile devices. Compared to QuestionSys, the application does not allow for
multilingualism. Instead a researcher would need to copy the instrument and change
its labels. A noteworthy feature of movisensXS is the sampling graph, which allows
researchers to properly define the way the study shall be presented on the respective
devices. Specific triggers (e.g., time- or value-based) may be used to start instruments.
Participants, in turn, get notified when the respective instrument needs to be processed.
Such approach may be of particular interest in the field of experience sampling. Like
Experience
Sampling
Method
QuestionSys, movisensXS applies a graphical notation for configuring the sampling
graph.
1https://buildfire.com ; accessed 2018-03-12
2https://bubble.is ; accessed: 2018-03-12
3https://www.limesurvey.org ; accessed: 2018-03-12
4https://www.surveymonkey.com ; accessed: 2018-03-12
5https://www.qualtrics.com ; accessed: 2018-03-12
6https://www.smartsurvey.co.uk ; accessed: 2018-03-12
43
Part III
Validation
45
Core contributions presented in this part were mainly published in the following articles:
SPSPGSR17
J. Schobel, R. Pryss, W. Schlee, T. Probst, D. Gebhardt, M. Schickler,
and M. Reichert. Development of Mobile Data Collection Applications
by Domain Experts: Experimental Results from a Usability Study. In
29th Int’l Conf on Advanced Information Systems Engineering (CAiSE),
number 10253 in LNCS, pages 60–75. Springer, June 2017
SPPSSR18
J. Schobel, R. Pryss, T. Probst, W. Schlee, M. Schickler, and M. Reichert.
Learnability of a Configurator Empowering End Users to Create Mobile
Data Collection Instruments: Usability Study. JMIR mHealth and uHealth,
6(6):e148, 2018
The original articles are added to the Appendix of this thesis.
47
8
Studies
To evaluate the applicability of the QuestionSys framework, various studies were con-
ducted. On one hand, these studies evaluated how efficient participants work with the
QuestionSys configurator. On the other, their mental effort for modeling tasks and the
complexity perceived in this context were assessed. Finally, insights into the participants’
process of modeling instruments were gained.
44 Participants
From 2 different Study Fields
1 Modeling Session
With 2 Modeling Tasks
Pilot Study
80 Participants
From 6 different Study Fields
2 Modeling Sessions
With 10 Modeling Tasks in total
Usability Study
Refine Study Design
Learning Effect could
be observed between
the Modeling Tasks
Figure 8.1: Conducted Usability Studies
Fig. 8.1 illustrates the course of the two studies. First, a pilot study was conducted (cf.
Section 8.1) in order to assess specific performance measures when working with the
developed configurator. When analyzing the results, a learning effect could be observed.
49
8 Studies
In order to focus on investigating this effect, a large-scale study with a more elaborated
study design was conducted (cf. Section 8.2).
8.1 Pilot Study
A pilot study assessing the usability of the developed QuestionSys configurator was
conducted with 44 participants [
118
]. The following research question, including several
hypotheses, was defined in concordance with the Goal Question Metric [5]:Goal Question
Metric
Research Question
Do end-users understand the modeling concept of the QuestionSys configurator with respect to the
complexity of the provided application?
Participants were recruited from various departments at Ulm University and were classified
into novices and experts, depending on their prior knowledge in process modeling, which
is a fundamental pillar of the QuestionSys approach. During the study, participants had
to model two data collection instruments by only using the provided configurator. The
time and operations needed to complete the tasks were assessed automatically, whereas
errors were assessed manually.
Although the obtained results were quite promising with respect to the assessed per-
formance measures, only one hypothesis could be statistically confirmed. However, the
hypotheses stating that experts are faster and make less errors than novices could not
be statistically proven. Furthermore, the conducted pilot study showed limitations with
respect to the validity of the results. For example, most of the participants already worked
with process models. This may act as confounder when evaluating the mental efforts
required for applying the change patterns. It may further affect the categorization of
participants into notices and experts. Furthermore, the process of recruiting participants
itself might be subject for discussion.
Besides these limitations interesting findings could be obtained. For example, a learning
effect was observed when analyzing the results from the novices sample. In particular,
the number of errors decreased from 4 in Task 1 to 1 in Task 2 (cf. Fig. 8.2), whereas
the errors remained stable for the experts sample (cf. Fig. 8.3).
8.2 Usability Study
In order to specifically focus on the discovered learning effect as well as to cope with the
Learnability
limitations of the pilot study (cf. Section 8.1), another usability study was conducted. On
one hand, the already promising results indicated by the pilot study should be replicated,
whereas additional research questions should be addressed on the other. Based on the
50
8.2 Usability Study
●
●
●
●
●
●
0
2
4
6
8
10
12
14
16
Task 1 Task 2
Number of Errors
Figure 8.2: Number of Errors (Novices)
0
2
4
6
8
10
12
14
16
Task 1 Task 2
Number of Errors
Figure 8.3: Number of Errors (Experts)
gained insights, a larger study with an improved and a more sophisticated study design
was conducted. More precisely, participants had to model 10 data collection instruments
with the QuestionSys configurator across 2 modeling sessions. In particular, the following
research questions were addressed by conducting this large-scale study [119]:
Research Questions
RQ 1: How are the performances of novices and the performances of experts changing from the first to
the last task (data collection instrument) of Session 1?
RQ 2: How are the performances of novices and the performances of experts changing from the last task
(data collection instrument) of Session 1 to the first task (data collection instrument) of Session
2?
RQ 3: How are the performances of novices and the performances of experts changing from the first to
the last task (data collection instrument) of Session 2?
RQ 4: How are the performances of novices and the performances of experts changing from the first task
(data collection instrument) of Session 1 to the last task (data collection instrument) of Session
2?
RQ 5: How many tasks (data collection instruments) are necessary until the performance metrics of
novices are as good as the performance metrics of experts at the first task (data collection instru-
ment)?
RQ 6: How does the self-reported mental effort change when modeling several data collection instru-
ments?
RQ 7: How are the performance measures of novices and experts compared to the self-reported mental
effort at each data collection instrument?
RQ 8: How are the performance measures of novices and experts compared to the self-reported mental
effort across all data collection instruments?
RQ 9: How are performance measures of novices and experts compared to the perceived complexity of
each data collection instrument?
51
8 Studies
RQ 10: How are performance measures of novices compared to the perceived complexity across all data
collection instruments?
RQ 11: How are performance measures of experts compared to the perceived complexity across all data
collection instruments?
8.2.1 Methods
For this study, a more complex procedure, compared to the pilot study, was designed.
Recruited participants had to model 10 data collection instruments over the course of two
consecutive sessions at Ulm University, using the provided QuestionSys configurator.
approx. 10 minapprox. 10 min
Demographic
Questionnaire
approx. 35 min
Video Tutorial
using a Beamer
(approx. 5 min)
approx. 5 min
Hand out Reward
Describe procedure
of experiment
Tutorial
Quality of Model
Questionnaire
Introduction &
Consent Form
Subjects got a
chocolate bar for
participating
Cognitive Load
Test 1
Cognitive test to assign
symbols to numbers
Exactly 2 min
Cognitive Load
Test 2
Cognitive Test to
detect if symbols occur
in set of symbols
Exactly 2 min
Modeling
Task 5
Mental Effort for
Task 5
Modeling
Task 4
Mental Effort for
Task 4
Modeling
Task 3
Mental Effort for
Task 3
Modeling
Task 2
Mental Effort for
Task 2
Modeling
Task A1
Mental Effort for
Task A1
5 Tasks in Session A
Each task comprises
the same amount of
operations and were
comparable in
complexity
approx. 35 min approx. 5 min
Hand out Reward
Quality of Model
Questionnaire
Subjects got a
chocolate bar for
participating
Modeling
Task 5
Mental Effort for
Task 5
Modeling
Task 4
Mental Effort for
Task 4
Modeling
Task 3
Mental Effort for
Task 3
Modeling
Task 2
Mental Effort for
Task 2
Modeling
Task B1
Mental Effort for
Task B1
5 Tasks in Session B
Each task comprises
the same amount of
operations and were
comparable in
complexity
Wait exactly for one week before starting the next session
Figure 8.4: Study Design
Fig. 8.4 illustrates the overall study design: participants were informed about the study
and had to process two tests measuring their cognitive load when working under stress.
Before collecting demographic data, a screencast introducing the QuestionSys configurator
was presented. For the first study session, participants were asked to solve five tasks
(i.e., model data collection instruments). After modeling each instrument, they had to go
through a short questionnaire collecting data on mental effort and perceived complexity
when working on the respective task. Finally, participants had to go through one last
52
8.2 Usability Study
questionnaire asking details on the overall quality of the modeled instruments. Altogether,
the first session took about 50 to 60 minutes in total, depending on the participants
speed.
Session 2 started exactly one week later. Demographic questions and the screencast
were skipped, resulting in a shorter duration of the session (i.e., approximately 30 to 40
minutes in total) compared to the first one. Participants had to process five additional
tasks and fill in related questionnaires capturing their mental efforts. Finally, feedback
on the quality of their modeled instruments was requested.
Participants
For the study, 80 participants – mainly students and research associates – from different
departments (e.g., Computer Science, Economics, Chemistry, Psychology, Medicine) at
Ulm University were recruited. It was ensured that the number of female and male
participants were almost equal. Then, the participants were
1
instructed to adhere to
the study design and
2
informed about the need to pass two consecutive sessions in
order to successfully complete the study. All materials (e.g., task descriptions, consent
form, questionnaires) were provided in German [
130
]. According to the study design,
participants who answered the question “Do you have experience in process modeling?”
with yes were classified as experts. On the other, participants who answered this question
with no were classified as novices for the subsequent analysis. Altogether, this resulted in
45 novices and 35 experts (80 in total). Note that only 3 out of the 80 participants did
not show up for Session 2 (one novice and two experts). Research questions that require
data from Session 2, therefore, were investigated with 77 participants (44 novices and 33
experts) instead of 80 (45 novices and 35 experts).
Performance Measures
For the study, the QuestionSys configurator was enhanced with a Study Mode enabling
specific features. Performance measures were automatically assessed. When participants
started or completed modeling a data collection instrument, the current time (i.e.,
timestamp) was logged to an Excel file. Furthermore, when editing the instrument (e.g.,
adding a page) the currently applied operation and timestamp were logged as well. Finally,
after each modeling step (i.e., operation) an image of the current state of the model was
generated and stored, which enabled the (manual) assessment of errors of the modeled
instruments.
To detect differences in the cognitive abilities of both groups (i.e., experts and novices)
two established tests measuring processing speed were performed [
125
]. Participants were
given 2 minutes for each test to assign symbols to numbers (“Digital-Symbol-Coding”)
and to detect symbols in a set of symbols (“Symbol Search”). Noticeable differences in
53
8 Studies
their cognitive abilities may be a confounder for the conducted study as a higher cognitive
ability could result in better / faster learnability of the QuestionSys configurator.
All data that was collected in this study is available in [119].
Tasks
All 10 tasks that needed to be processed were comparable regarding their complexity.
Note that a difference in the complexity of the models might limit the validity of the
obtained results. A change of assessed performance measures, in turn, might be attributed
to a diverse model complexity or a measurable learning effect. All tasks were designed in
the same way. This includes the textual representation handed out to participants as
well as the amount of operations needed in the best case.
The data collection instruments to be modeled were selected from various domains in
order to evaluate the feasibility and practical applicability in these settings. Appendix A.1
exemplarily shows one task description that was translated from German to English.
Questionnaires
Throughout the study, additional data was collected using traditional paper-based ques-
tionnaires. For example, a demographic questionnaire collecting personal information
(e.g., gender or education) was handed out to the participants. More specifically, infor-
mation regarding prior knowledge on process modeling was assessed, as this information
was used to classify participants into novices and experts. After modeling an instrument,
participants had to answer 5 questions assessing their mental effort. Thereby, a 7 point
Likert-scale with respective answers ranging from “I strongly agree” (1) to “I strongly
disagree” (7) with an additional “neutral” element (4) was presented. Finally, participants
had to answer questions regarding the quality of the modeled instruments.
8.2.2 Discussion
The overall goal of the study was to evaluate whether end-users are able to develop data
collection instruments when using the QuestionSys configurator. In order to measure
their performance, various measures were assessed. The time and operations needed to
complete the tasks were automatically tracked by the configurator. Furthermore, all
created models were manually assessed to evaluate potential errors. Finally, participants
assessed their performance and mental effort required to solve the given tasks by filling
in self-reporting questionnaires. During each session, a learning effect was observed; the
time and operations needed for modeling were decreasing from task to task (
RQ1
&
RQ3
).
Across the two sessions the participants increased their overall performance (
RQ4
), i.e.,
the errors in the modeled data collection instruments were decreasing from the 1
st
to the
54
8.2 Usability Study
10
th
task. This learning effect with respect to errors, however, could not be observed for
experts, as their models contained few errors already in the first task.
After a break of one week (i.e., participants were not using the QuestionSys configurator
during this period), the performance measures for novices decreased again, whereas the
ones of experts remained stable (
RQ2
). This may be explained due to the fact that experts
are working with corresponding applications on a day-to-day basis. Novices, in turn, need
to get reacquainted with the application. Note that novices performed significantly better,
regarding the time and operations needed, from the third task on, compared to experts
in the first task (
RQ5
). Unfortunately, novices were unable to catch up regarding the
errors in their modeled data collection instruments. To enable untrained domain experts
to properly create more error-free data collection instruments, it might be necessary to
increase the number of training sessions.
When analyzing the questionnaires, the participants had to fill in right after modeling a
data collection instrument, further insights could be obtained.
RQ6
, for example, revealed
that the mental effort for modeling instruments is decreasing. Furthermore,
RQ7
showed a
strong correlation between the self-assessed mental effort and the performance measures
assessed by the QuestionSys configurator for each task. In detail, novices showed 19
(out of 30) and experts 11 (out of 30) significant correlations. These correlations may
be explained due to the fact that experts initially rated the mental effort for modeling
instruments lower than novices did. Furthermore, the experts’ performance was better
than the one of novices, i.e., experts were faster, required less operations to complete a
task, and made less errors. Finally, a lower mental effort significantly correlated with
an overall better performance (
RQ8
). Above all, this could be shown for both groups,
whereas the effect is stronger for novices. Again, this may be explained due to the fact
that experts are more likely working on a day-to-day basis with similar applications, and,
therefore, retain basic expertise.
In addition to the self-rated mental effort, participants had to give feedback on the
perceived complexity of the data collection instruments to be modeled. Again, the results
showed significant correlations between the perceived complexity and the performance
measures. More specifically, novices had 14 (out of 30) and experts 12 (out of 30)
correlations (
RQ9
). In other words, the more time or operations the participants needed,
or the more errors they made, the higher the perceived complexity was rated by them.
Regarding
RQ10
and
RQ11
the overall performance of both groups were compared to their
perceived complexity. Results revealed that an increase in the perceived complexity is
also associated with a decrease in performance for both groups.
This study shows limitations that need to be discussed properly. First, the process of
recruiting participants might affect generalizability as the study mainly involved students
and research associates. However, [
34
] showed that students can act as proper substitutes
in empirical research. Second, categorizing the participants only into two groups may be
subject for discussion. The categorization solely based on a single “yes / no” question
may be subject for further investigations. Both aspects could be addressed in another
55
8 Studies
study with a more sophisticated categorization. One could, for example, categorize the
participants by directly observing their modeling behavior. However, tests measuring the
processing speed of both groups were performed before working with the QuestionSys
configurator, indicating similar cognitive abilities. Third, a baseline comparison between
the groups show differences regarding gender, education, and their field of study. Some
differences (e.g., the field of study) are intended as participants were recruited specifically
from various domains to enable comparison. As stated throughout this thesis, the
QuestionSys framework targets at domain experts having no knowledge regarding process
modeling and mobile application development respectively. In this study, end-users from
medicine, psychology or social sciences are involved. Again, tests measuring processing
speed indicate similar abilities regarding their cognitive behavior. Fourth, the experts
sample was smaller than the one with novices (35 experts vs. 45 novices), resulting in a
weaker statistical power. Fifth, all tasks to be modeled origin from various domains (e.g.,
healthcare, travel expense, food delivery) to illustrate the applicability of the developed
approach in a multitude of application scenarios. It may be subject to discussion, whether
some of the instruments to be modeled were more familiar to participants than others.
Fortunately, the
RQs
comparing the perceived complexity as well as the mental effort
with the automatically assessed performance measures show a strong correlation between
those metrics.
Altogether, the study replicates valuable findings from the pilot study [
118
]. In particular,
results confirm that even novices were able to properly model data collection instruments.
In detail, the recruited participants got significantly better (i.e., needed less time and
operations and made less errors), the longer they worked with the QuestionSys configu-
rator. More precisely, a learning effect could be noticed within each session and across
both sessions. Altogether, the QuestionSys configurator constitutes a feasible approach
for enabling domain experts having little or no prior knowledge in process modeling or
mobile application development to create data collection applications themselves.
56
9
Related Studies
Several studies measuring mental efforts during process modeling are described in litera-
ture. Common to them is their focus on the process model. In this context, [
53
] analyzes
the process of process modeling, whereas the approach described in [
15
] visualizes different
Process
Modeling
steps a process modeler undertakes when modeling (business) processes. Moreover, [
126
]
applies eye tracking software to gain a better understanding of factors that may influence
the way process models are specified by individuals. Furthermore, [
135
] presents insights
into and lessons learned in studies on process model comprehension that rely on eye
tracking studies. In the studies described in Chapter 8, data collection instruments are
technically represented by process models. Additional aspects have to be modeled by
domain experts, which are irrelevant in the context of process modeling, like the ability to
support different languages, element versions, or modes of an instrument (i.e., self-rating
vs. interview mode). These aspects might increase the overall mental effort for untrained
domain experts when working with a configurator like QuestionSys. Consequently, the
studies described in Chapter 8 differ from the above ones.
Psychological studies revealed manifold insights into the measurement of mental efforts.
Mental Effort
For example, [
122
] introduces the Cognitive Load Theory that provides guidelines to
Cognitive Load
Theory
assist learners to actively process available information as working memory capacity is
limited. Closely related, [
56
] presents concepts on how to effectively measure mental
effort when working on specific tasks. Additional ideas on how to derive conclusions
with respect to individuals are proposed. The approach described in [
61
] focuses on
educational perspectives and discusses the process of learning more generally. Finally,
[
124
] summarizes related challenges and discusses potential research directions. Related
to the studies presented in Chapter 8, [
134
] describes an eye tracking study measuring
57
9 Related Studies
the mental effort of participants when modeling (business) processes. [
134
] showed that
the mental effort for modeling tasks quickly reaches cognitive limitations thwarting the
performance of experts and novices in modeling.
End-User Programming approaches have proven their feasibility in a multitude of studies.
End-User
Programming
In particular, they shall support non-programmers in developing software applications.
For example, [
36
] provides an environment allowing system administrators to visually
model script-based applications. An experiment investigated the practical applicability
of the proposed approach. In turn, [
6
] introduced a graphical programming language,
representing each function of a computer program as a block. Blocks, in turn, may be
built upon each other to (graphically) develop a software application.
In the field of domain-specific configurators for developing, configuring, and maintaining
Domain-
Specific
Configurator
software applications, only few evaluations have been reported in literature. For example,
[
3
] compared a web-based configurator for ambulatory assessments against movisensXS,
which is a commercial solution for Ecological Momentary Assessments. The authors
evaluated their configurator with two experts on one hand. On the other, 10 participants
evaluated the respective client application for enacting the configured assessment. Both
parts of the study rely on standardized user-experience questionnaires (e.g.,
SUS
[
9
]) to
System
Usability Scale
collect feedback from individuals working with the application. Compared to the studies
presented in this thesis, the results are limited due to the low number of participants.
Furthermore, only self-rated user perception has been considered as a metric for the
usability of the application, completely ignoring automatically collected performance
measures.
A web-based configurator to create and coordinate experiments in the context of infor-
mation retrieval is presented in [
80
]. In particular, the authors evaluated the application
they have developed in two ways: The backend management system was evaluated by
one researcher focusing on human-computer interaction and by one regular student. Both
participants confirmed a good usability. The frontend, however, was evaluated with a
study comprising 48 participants. Comparable to the studies presented in this thesis, the
application tracked the time to complete respective tasks. Furthermore, participants were
asked to provide feedback with respect to their performance. Compared to these studies,
the studies presented in Chapter 8 pursued different approaches. More specifically, the
QuestionSys configurator was evaluated along observable correctness properties of the
modeled instruments. Following this approach, performance measures (e.g., the time to
complete a model or the operations needed) were automatically assessed and evaluated
over a certain period of time. Finally, the performance measures were compared with the
self-reported mental effort of participants when modeling respective instruments.
In conclusion, the conducted studies specifically focused on measuring the learnability
Learnability
of the QuestionSys configurator. This approach may be considered as a promising way
to evaluate the usability of applications in general. Measuring learnability, however, is
a time-consuming endeavor. In more detail, learning is often considered as a process
over time, taking practical experience into account as well. When measuring learnability,
58
this means that a consecutive series of tasks need to be executed over a certain period
of time and corresponding performance measures need to be assessed. Due to the
time-consuming procedure of measuring learnability, however, such studies are often
neglected [
133
]. Learnability, however, might have an impact on the success or failure of
an application in real-world scenarios [
28
]. In order to ensure usability, more effort is
put into the evaluation of best practices for creating a user-friendly software application.
Mostly, the design of the user interface [
47
,
54
,
120
] or overall user experience [
30
,
48
]
are considered. Although standardized self-report questionnaires (e.g., common usability
scales, like
SUS
) may assess respective properties fairly easy, they might be misleading
when evaluating sophisticated applications like the QuestionSys configurator.
59
Part IV
Conclusion
61
10
Summary and Outlook
To mitigate the limitations of paper-based instruments, digital solutions based on common
Web technologies were established. However, most solutions are unable to cope with
the demanding requirements of many large-scale data collection scenarios like the use
of sensors to collect vital parameters during interviews or the offline processing of
instruments (i.e., if no stable Internet connection is available). Smart mobile devices (i.e.,
smartphones or tablets), have the potential to close this gap and to meet these complex
requirements.
Developing mobile applications for collecting data in large-scale scenarios is a challenging
task. For example, the short release cycles from vendors as well as platform-specific
peculiarities (e.g., diverging user interfaces) need to be taken into account. Maintaining
mobile data collection applications, therefore, is complex, time-consuming, and costly as
both domain experts and application developers need to be involved.
The QuestionSys framework describes methods that empower domain experts to develop
sophisticated mobile data collection applications themselves, i.e., without need to involve
any IT experts. The QuestionSys framework presents a model-driven configurator
that applies techniques known from end-user programming to properly support domain
experts. Usability studies conducted with untrained participants have shown the practical
feasibility of the approach. The instruments modeled with the configurator, in turn, can
be deployed to and flexibly executed on smart mobile devices to collect data.
The QuestionSys framework with its model-driven approach will significantly increase
the speed for developing mobile data collection applications. Furthermore, it will reduce
costs and relieve application developers from manual tasks, like migrating existing
63
10 Summary and Outlook
applications to new operating system versions. Moreover, the communication overhead
between domain experts and application developers will be significantly reduced as
domain experts are empowered to digitize instruments themselves. Finally, the developed
approach demonstrates the applicability of process management techniques in a broader
scope compared to business process support.
10.1 Contribution
The thesis presents the QuestionSys framework that empowers domain experts to create
QuestionSys
Framework
data collection instruments. In a multitude of interviews conducted with experts from
various domains, fundamental requirements for realizing data collection scenarios were
elaborated. In this thesis, a generic concept was developed serving a multitude of
application scenarios from various domains. Fig. 10.1 summarizes core contributions of
the thesis.
For developing complex data collection instruments, QuestionSys provides a model-driven
configurator. The latter applies process management technologies in a broader scope
Configurator
by mapping instruments to executable process models. The mapping allows specifying
the flow of an instrument on an abstract, i.e., platform-independent, level. Note that
QuestionSys applies techniques known from end-user programming.
Usability studies showed that participants were able to properly use the configurator.
Usability Study
First, a pilot study revealed a learning effect from modeling task to modeling task.
Second, to investigate this effect, a more complex study was designed, which reproduced
valuable insights from the pilot study and enabled additional insights into the process of
modeling instruments. In particular, it was shown that participants were able to properly
use the configurator, i.e., the mental effort for modeling data collection instruments with
the QuestionSys configurator continuously decreased from task to task. In this context,
not only self-reported mental effort was evaluated, but also performance measures like the
time and operations needed to complete specific tasks. Overall, the values were decreasing
over time, indicating a promising approach for domain experts without knowledge in
process modeling or application development in general.
The QuestionSys framework enables the deployment of modeled instruments to smart
mobile devices (e.g., smartphones or tablets) to collect data with them. For this purpose,
alightweight mobile process engine was developed that is capable of interpreting and
Lightweight
Process Engine
executing instruments in a robust and efficient manner. To enable the later extension of
the mobile data collection application, Executable Components were introduced. These
Executable
Component
components, in turn, provide the logic for presenting the user interface or for collecting
entered data. Further, they are not part of the lightweight mobile process engine itself,
but rather extend its functionality. More specifically, the engine communicates with
ECs
like with external services and allows exchanging these components on demand. This
allows flexibly adapting the mobile data collection client to new emerging requirements.
64
10.1 Contribution
Archiving &
Versioning
Monitoring
& Analysis
Enactment &
Execution
Deployment
Design &
Modeling
Mobile Data
Collection Lifecycle
Domain Specific Requirements
Execution & Monitoring
End-User Programming
Process Model Mapping
(Graphical) Modeling Language
Mobile Process Engine
Model Data Collection
Instrument
Lightweight Process Engine
for Execution and Monitoring
Questionnaire
Model
Page
Question
Process
Model
Process
Activity
Process
Data Element
Questionnaire
Instance
Process
Instance
n
1
1
n
1
n
n
1
1
n
1
n
maps to
maps to
maps to
maps to
Alcohol
Consumption
Cigarette
Consumption
StartFlow Activity
XORjoin
DataElement
WriteAccess
ReadAccess
EndFlow
ET_ControlFlow_Default
ET_DataFlow
AlcoholCigarettes
(Cigarettes = yes)
AND (Alcohol = yes)
XORsplit
else
(Cigarettes = yes)
AND (Alcohol = no)
ET_ControlFlow
Cigarettes
& Alcohol
Page
Intro
Page
General EndCigarettes
Process Technology
(e.g., Process Model)
Navigation Operation Based
on Already Given Answers
QuestionSys Configurator
UI Generator with
Custom Control Elements
Customized Executable
Components
Design Complex
Navigation Logic
Large-Scale Usability
Study
End-User Programming
Techniques
Figure 10.1: QuestionSys Framework
Altogether, the QuestionSys framework has the potential to significantly influence the
way mobile data collection applications will be developed in future. First, large-scale
data collection scenarios, like clinical trials or psychological studies, will benefit from the
short development cycles for mobile instruments. Second, the modeling language used
by the configurator might contribute towards a common (graphical) notation and, thus,
foster the communication between domain experts and application developers. Finally,
QuestionSys can serve as a valuable benchmark for mobile data collection in general.
This includes the graphical modeling of instruments as well as their flexible execution on
smart mobile devices.
65
10 Summary and Outlook
10.2 Additional Publications
In addition to the core publications of this thesis, its author was involved in a number of
additional publications related to data collection applications.
[
35
,
83
,
105
] describe early works dealing with data collection applications in the psy-
chological domain. In detail, psychologists were supported in various scenarios (e.g.,
collecting data on adverse childhood experience or detecting risky pregnancies) by imple-
menting mobile applications. Note that the requirements gathered from interviews with
domain experts as well as the experiences gained during the process of developing and
maintaining these applications, triggered the research on the QuestionSys framework.
Insights into the related development process were published in [110].
Mobile data collection in the context of mobile crowdsensing applications is introduced
in [
94
]. More specifically, the work presented in [
68
,
71
,
74
] focuses on technical details of
the TrackYourTinnitus platform, whereas [
65
,
66
,
101
] present scientific results obtained
from the analysis of the data collected with this data collection application.
In [
106
], a sophisticated sensor framework is presented that allows connecting sensors to
smart mobile devices. The framework provides features to retrieve data from internal
sensors (e.g., camera or microphone) as well as external ones (e.g., pulse sensor connected
via
Bluetooth
). Corresponding data can be both analyzed and visualized in order to
provide additional information during data collection in clinical trials.
Insights into the development and maintenance of mobile applications were reported in
[93].
10.3 Outlook
Properly supporting end-users in collecting data in large-scale application scenarios, like
clinical trials, is a complex endeavor, which can be only partially covered by a thesis.
The latter revealed several aspects that are not part of this manuscript, but may be
addressed in future research:
•
The integration of QuestionSys with concepts, methods and technologies known
from the field of (business) process management offers promising perspectives and,
Process
Management
hence, should be investigated in future work as well. For example, the management
of instruments may adopt concepts like process configuration [
2
,
27
], process
compliance checking and monitoring [40,50], context-aware process injection [57],
and process schema evolution [82].
•
Process mining algorithms [
123
] may be applied to discover additional insights into
Process Mining
the execution of mobile processes, i.e., data collection instruments running on smart
mobile devices. The benefits of corresponding approaches are described in [
51
],
66
10.3 Outlook
which applied existing process mining algorithms to evaluate selected hospital pro-
cesses. More specifically, the control flow perspective (i.e., “Are processes executed
exactly as they were specified or do they deviate from the specified behavior?”), the
organizational perspective (i.e., “Which participants are involved in this process
and with whom do they work together?”), and the performance perspective (i.e.,
“What is the execution time for specific cases and which participants did work on
the latter?”) were evaluated and visualized.
•
In the current version of the QuestionSys configurator, the integration of sensors
Sensors
into data collection instruments is limited. Although smart mobile devices offer
plenty of internal sensors, the mobile data collection application currently focuses
on the most common ones (e.g., microphone or camera). However, external sensors,
connected via
Bluetooth
or
WLAN
, need to be explicitly integrated by implementing
respective code. Note that this is aggravated due to the fact that in most cases the
vendors of such sensors do not offer public
API
s to communicate with corresponding
devices. For this purpose, the QuestionSys configurator provides a generic element
that allows specifying sensor configurations. A corresponding mobile data collection
application, or rather the
EC
being responsible for handling the logic to interact
with the sensor, need to properly interpret and handle this configuration. Though
the used JSON configuration objects constitute a rather pragmatic approach, it is
not suitable for domain experts. In consequence, end-user programming techniques
should be applied as well.
•
When realizing mobile data collection applications novel control elements for enter-
ing data were introduced and evaluated in usability studies. More specifically, these
elements were compared to common ones known from other mobile applications.
Results indicate that some of the new control elements were well understood by
the users interacting with a created mobile application, i.e., the elements were
rated positively in respect to usability aspects. Similar to the study measuring
Usability Study
learnability of the configurator, another study should measure the performance of
participants when processing an instrument, i.e., entering data. Especially, the
newly introduced control elements may enable a faster processing of instruments.
Likewise, an additional study on modeling sensors should evaluate whether or
not domain experts are able to properly use this feature in their data collection
instruments.
67
Bibliography
[1]
J. Anhøj and C. Møldrup. Feasibility of Collecting Diary Data from Asthma
Patients Through Mobile Phones and SMS (Short Message Service): Response
Rate Analysis and Focus Group Evaluation from a Pilot Study. Journal of Medical
Internet Research, 6(4):e42, 2004.
[2]
C. Ayora, V. Torres, B. Weber, M. Reichert, and V. Pelechano. VIVACE: A
Framework for the Systematic Evaluation of Variability Support in Process-Aware
Information Systems. Information and Software Technology, 57:248–276, 2015.
[3]
A. Bachmann, R. Zetzsche, A. Schankin, T. Riedel, M. Beigl, M. Reichert, P. San-
tangelo, and U. Ebner-Priemer. ESMAC: A Web-Based Configurator for Context-
Aware Experience Sampling Apps in Ambulatory Assessment. In 5th Int’l Conf on
Wireless Mobile Communication and Healthcare, pages 15–18, 2015.
[4]
F. Balagtas-Fernandez, M. Tafelmayer, and H. Hussmann. Mobia Modeler: Easing
the Creation Process of Mobile Applications for Non-Technical Users. In 15th Int’l
Conf on Intelligent User Interfaces, pages 269–272. ACM, 2010.
[5]
V. R. Basili. Software Modeling and Measurement: The Goal/Question/Metric
Paradigm. Technical report, University of Maryland, MD, USA, 1992.
[6]
A. Begel and E. Klopfer. Starlogo TNG: An Introduction to Game Development.
Journal of E-Learning, 53:146, 2007.
[7]
T. W. Boonstra, A. Werner-Seidler, B. O’Dea, M. E. Larsen, and H. Christensen.
Smartphone App to Investigate the Relationship between Social Connectivity and
Mental Health. arXiv preprint arXiv:1702.02644, 2017.
[8]
M. Brambilla and P. Fraternali. Large-Scale Model-Driven Engineering of Web
User Interaction: The WebML and WebRatio Experience. Science of Computer
Programming, 89:71–87, 2014.
[9]
J. Brooke. SUS: a ’Quick and Dirty’ Usability Scale. In P. W. Jordan, B. Thomas,
B. A. Weerdmeester, and I. L. McClelland, editors, Usability Evaluation in Industry,
volume 189, pages 4–7. Taylor and Francis, London, 1996.
[10]
A. W. Brown. Model Driven Architecture: Principles and Practice. Software and
Systems Modeling, 3(4):314–327, 2004.
69
Bibliography
[11]
J. Bryant and M. Jones. Responsive Web Design. In Pro HTML5 Performance,
pages 37–49. Springer, 2012.
[12]
S. Buchwald, T. Bauer, and M. Reichert. Bridging the Gap Between Business
Process Models and Service Composition Specifications. In Service Life Cycle
Tools and Technologies: Methods, Trends and Advances, pages 124–153. Idea Group
Referenc, November 2011.
[13]
J. A. Cafazzo, M. Casselman, N. Hamming, D. K. Katzman, and M. R. Palmert.
Design of an mHealth App for the Self-Management of Adolescent Type 1 Diabetes:
A Pilot Study. Journal of Medical Internet Research, 14(3):e70, 2012.
[14]
P. Carlbring, S. Brunt, S. Bohman, D. Austin, J. Richards, L.-G. Öst, and G. An-
dersson. Internet vs. Paper and Pencil Administration of Questionnaires Commonly
Used in Panic/Agoraphobia Research. Computers in Human Behavior, 23(3):
1421–1434, 2007.
[15]
J. Claes, I. Vanderfeesten, J. Pinggera, H. A. Reijers, B. Weber, and G. Poels.
A Visual Analysis of the Process of Process Modeling. Information Systems and
e-Business Management, 13(1):147–190, 2015.
[16]
P. Dadam and M. Reichert. The ADEPT Project: A Decade of Research and Devel-
opment for Robust and Flexible Process Support – Challenges and Achievements.
Computer Science - Research and Development, 23(2):81–97, 2009.
[17]
G. Decker, A. Grosskopf, and A. Barros. A Graphical Notation for Modeling
Complex Events in Business Processes. In 11th IEEE Int’l Enterprise Distributed
Object Computing Conference (EDOC), pages 27–27. IEEE, Oct 2007.
[18]
F. Ehrler, R. Wipfli, D. Teodoro, E. Sarrey, M. Walesa, and C. Lovis. Challenges
in the Implementation of a Mobile Application in Clinical Practice: Case Study
in the Context of an Application that Manages the Daily Interventions of Nurses.
JMIR mHealth and uHealth, 1(1):e7, 2013.
[19]
J. A. Ellis. Leveraging Mobile Phones for Monitoring Risks for Noncommunicable
Diseases in the Future. Journal of Medical Internet Research, 19(5):e137, 2017.
[20]
R. Fernandez-Ballesteros. Self-Report Questionnaires. In M. Hersen, S. N. Haynes,
and E. M. Heiby, editors, Comprehensive Handbook of Psychological Assessment,
volume 3, pages 194–221. John Wiley & Sons Hoboken, NJ, 2004.
[21]
R. Fielding. Architectural Styles and the Design of Network-Based Software Archi-
tectures. PhD thesis, University of California, Irvine, 2000.
[22]
D. Forster, R. Behrens, H. Campbell, and P. Byass. Evaluation of a Computerized
Field Data Collection System for Health Surveys. Bulletin of the World Health
Organization, 69(1):107, 1991.
70
Bibliography
[23]
B. Frain. Responsive Web Design with HTML5 and CSS3. Packt Publishing Ltd,
2012. ISBN 1849693188.
[24]
A. Gaggioli, G. Pioggia, G. Tartarisco, G. Baldus, D. Corda, P. Cipresso, and
G. Riva. A Mobile Data Collection Platform for Mental Health Research. Personal
and Ubiquitous Computing, 17(2):241–251, 2013.
[25]
P. Geiger, M. Schickler, R. Pryss, J. Schobel, and M. Reichert. Location-based
Mobile Augmented Reality Applications: Challenges, Examples, Lessons Learned.
In 10th Int’l Conf on Web Information Systems and Technologies (WEBIST),
Special Session on Business Apps, pages 383–394, April 2014.
[26]
H. Gundlach. What is a Psychological Instrument? In M. G. Ash and T. Sturm,
editors, Psychology’s Territories, chapter 9, pages 195–224. Lawrence Erlbaum
Associates, Mahwah, New Jersey, 2007.
[27]
A. Hallerbach, T. Bauer, and M. Reichert. Capturing Variability in Business
Process Models: The Provop Approach. Journal of Software Maintenance and
Evolution: Research and Practice, 22(6-7):519–546, November 2010.
[28]
R. Harrison, D. Flood, and D. Duce. Usability of Mobile Applications: Literature
Review and Rationale for a New Usability Model. Journal of Interaction Science, 1
(1):1, 2013.
[29]
W. Harrison. The Dangers of End-User Programming. IEEE Software, 21(4):5–7,
2004.
[30]
M. Hassenzahl and N. Tractinsky. User Experience – A Research Agenda. Behaviour
& Information Technology, 25(2):91–97, 2006.
[31]
H. Heitkötter, S. Hanschke, and T. A. Majchrzak. Evaluating Cross-Platform
Development Approaches for Mobile Applications. In Int’l Conf on Web Information
Systems and Technologies (WEBIST), pages 120–138. Springer, 2012.
[32]
H. Heitkötter, T. A. Majchrzak, and H. Kuchen. Cross-Platform Model-Driven
Development of Mobile Applications with
md2
. In 28th Annual ACM Symp on
Applied Computing, pages 526–533. ACM, 2013.
[33]
M. Henkel and J. Stirna. Pondering on the Key Functionality of Model-Driven
Development Tools: The Case of Mendix. In Int’l Conf on Business Informatics
Research, pages 146–160. Springer, 2010.
[34]
M. Höst, B. Regnell, and C. Wohlin. Using Students as Subjects —- A Comparative
Study of Students and Professionals in Lead-Time Impact Assessment. Empirical
Software Engineering, 5(3):201–214, 2000.
71
Bibliography
[35]
D. Isele, M. Ruf-Leuschner, R. Pryss, M. Schauer, M. Reichert, J. Schobel,
A. Schindler, and T. Elbert. Detecting Adverse Childhood Experiences with
a Little Help from Tablet Computers. In XIII Congress of European Society of
Traumatic Stress Studies (ESTSS) Conf, pages 69–70, June 2013.
[36]
E. Kandogan, E. Haber, R. Barrett, A. Cypher, P. Maglio, and H. Zhao. A1:
End-User Programming for Web-based System Administration. In 18th ACM
Symposium on User Interface Software and Technology. ACM, 2005.
[37]
H. Keedle, V. Schmied, E. Burns, and H. Dahlen. The Design, Development,
and Evaluation of a Qualitative Data Collection Application for Pregnant Women.
Journal of Nursing Scholarship, 50(1):47–55, 2018.
[38]
A. Khambati, J. Grundy, J. Warren, and J. Hosking. Model-Driven Development
of Mobile Personal Health Care Applications. In 23rd IEEE/ACM Int’l Conf on
Automated Software Engineering, pages 467–470. IEEE Computer Society, 2008.
[39]
E. Klopfer, S. Yoon, and T. Um. Teaching Complex Dynamic Systems to Young
Students with StarLogo. The Journal of Computers in Mathematics and Science
Teaching, 24(2):157, 2005.
[40]
D. Knuplesch, M. Reichert, and A. Kumar. A Framework for Visually Monitoring
Business Process Compliance (Extended Abstract). In 8th Int’l Workshop on
Enterprise Modeling and Information Systems Architectures (EMISA), June 2017.
[41]
A. J. Ko, R. Abraham, L. Beckwith, A. Blackwell, M. Burnett, M. Erwig, C. Scaffidi,
J. Lawrance, H. Lieberman, B. Myers, et al. The state of the art in end-user software
engineering. ACM Computing Surveys, 43(3):21, 2011.
[42]
U. Kreher. Konzepte, Architektur und Implementierung adaptiver Prozessmanage-
mentsysteme. PhD thesis, Ulm University, 2014.
[43]
S. J. Lane, N. M. Heddle, E. Arnold, and I. Walker. A Review of Randomized
Controlled Trials Comparing the Effectiveness of Hand Held Computers with Paper
Methods for Data Collection. BMC Medical Informatics and Decision Making, 6
(1):1, 2006.
[44]
A. Lanz. Adaptive Time-and Process-Aware Information Systems. PhD thesis, Ulm
University, 2017.
[45]
A. Lanz, B. Weber, and M. Reichert. Time Patterns for Process-Aware Information
Systems. Requirements Engineering, 19(2):113–141, May 2014.
[46]
A. Lanz, M. Reichert, and B. Weber. Process Time Patterns: A Formal Foundation.
Information Systems, 57:38–68, April 2016.
[47]
B. Laurel and S. J. Mountford. The Art of Human-Computer Interface Design.
Addison-Wesley Longman Publishing Co., Inc., 1990.
72
Bibliography
[48]
E. L.-C. Law, V. Roto, M. Hassenzahl, A. P. Vermeeren, and J. Kort. Understanding,
Scoping and Defining User Experience: A Survey Approach. In SIGCHI Conf on
Human Factors in Computing Systems, pages 719–728. ACM, 2009.
[49]
D. D. Luxton, R. A. McCann, N. E. Bush, M. C. Mishkind, and G. M. Reger.
mHealth for Mental Health: Integrating Smartphone Technology in Behavioral
Healthcare. Professional Psychology: Research and Practice, 42(6):505, 2011.
[50]
L. T. Ly, D. Knuplesch, S. Rinderle-Ma, K. Goeser, M. Reichert, and P. Dadam.
SeaFlows Toolset – Compliance Verification Made Easy. In 22th Int’l Conf on
Advanced Information Systems Engineering (CAiSE), Demos, June 2010.
[51]
R. S. Mans, M. Schonenberg, M. Song, W. M. P. van der Aalst, and P. J. Bakker.
Application of Process Mining in Healthcare – A Case Study in a Dutch Hospital.
In Int’l Conf on Biomedical Engineering Systems and Technologies, pages 425–438.
Springer, 2008.
[52]
J. S. Marcano-Belisario, K. Huckvale, A. Saje, A. Porcnik, C. Morrison, and J. Car.
Comparison of Self Administered Survey Questionnaire Responses Collected Using
Mobile Apps versus other Methods. Cochrane Database of Systematic Reviews, 4,
2014.
[53]
M. Martini, J. Pinggera, M. Neurauter, P. Sachse, M. R. Furtner, and B. Weber.
The Impact of Working Memory and the "Process of Process Modelling" on Model
Quality: Investigating Experienced Versus Inexperienced Modellers. Scientific
Reports, 6, 2016.
[54]
D. J. Mayhew. The Usability Engineering Lifecycle. In CHI’99 Extended Abstracts
on Human Factors in Computing Systems, pages 147–148. ACM, 1999.
[55]
J. Mirkovic, D. R. Kaufman, and C. M. Ruland. Supporting Cancer Patients in
Illness Management: Usability Evaluation of a Mobile App. JMIR mHealth and
uHealth, 2(3):e33, 2014.
[56]
G. Mulder. The Concept and Measurement of Mental Effort. In Energetics and
Human Information Processing, pages 175–198. Springer, 1986.
[57]
N. Mundbrod, G. Grambow, J. Kolb, and M. Reichert. Context-Aware Process
Injection: Enhancing Process Flexibility by Late Extension of Process Instances.
In 23rd Int’l Conf on Cooperative Information Systems (CoopIS), number 9415 in
LNCS, pages 127–145. Springer, October 2015.
[58]
S. Newman. Building Microservices: Designing Fine-Grained Systems. O’Reilly
Media, Inc., 2015.
[59]
N. Nurseitov, M. Paulson, R. Reynolds, and C. Izurieta. Comparison of JSON and
XML Data Interchange Formats: A Case Study. Caine, 9:157–162, 2009.
73
Bibliography
[60]
Object Management Group. Business Process Model and Notation (BPMN) Version
2.0. https://www.omg.org/spec/BPMN/2.0, 2011. last accessed: 26.02.2018.
[61]
F. Paas, J. E. Tuovinen, J. J. Van Merrienboer, and A. A. Darabi. A Motivational
Perspective on the Relation Between Mental Effort and Performance: Optimiz-
ing Learner Involvement in Instruction. Educational Technology Research and
Development, 53(3):25–34, 2005.
[62]
T. M. Palermo, D. Valenzuela, and P. P. Stork. A Randomized Trial of Electronic
versus Paper Pain Diaries in Children: Impact on Compliance, Accuracy, and
Acceptability. Pain, 107(3):213–219, 2004.
[63]
I. Pavlović, T. Kern, and D. Miklavčič. Comparison of Paper-Based and Electronic
Data Collection Process in Clinical Trials: Costs Simulation Study. Contemporary
Clinical Trials, 30(4):300–316, 2009.
[64]
T. Probst, R. Pryss, B. Langguth, and W. Schlee. Emotion dynamics and tinnitus:
daily life data from the “TrackYourTinnitus” application. Scientific Reports, 6:
31166, 2016.
[65]
T. Probst, R. Pryss, B. Langguth, J. Rauschecker, J. Schobel, M. Reichert,
M. Spiliopoulou, W. Schlee, and J. Zimmermann. Does tinnitus depend on time-
of-day? An ecological momentary assessment study with the ”TrackYourTinnitus”
application. Frontiers in Aging Neuroscience, 9:253–253, 2017.
[66]
T. Probst, R. Pryss, B. Langguth, M. Spiliopoulou, M. Landgrebe, M. Vesala,
S. Harrison, J. Schobel, M. Reichert, M. Stach, and W. Schlee. Outpatient Tinnitus
Clinic, Self-Help Web Platform, or Mobile Application to Recruit Tinnitus Study
Samples? Frontiers in Aging Neuroscience, 9:113–113, April 2017.
[67]
R. Pryss, M. Reichert, A. Bachmeier, and J. Albach. BPM to Go: Supporting
Business Processes in a Mobile and Sensing World. In BPM Everywhere, pages
167–182. Future Strategies Inc, 2015.
[68]
R. Pryss, M. Reichert, B. Langguth, and W. Schlee. Mobile Crowd Sensing Services
for Tinnitus Assessment, Therapy, and Research. In IEEE Int’l Conf on Mobile
Services (MS), pages 352–359. IEEE, 2015.
[69]
R. Pryss, P. Geiger, M. Schickler, J. Schobel, and M. Reichert. Advanced Algorithms
for Location-Based Smart Mobile Augmented Reality Applications. Procedia
Computer Science, 94:97–104, 2016.
[70]
R. Pryss, P. Geiger, M. Schickler, J. Schobel, and M. Reichert. The AREA
Framework for Location-Based Smart Mobile Augmented Reality Applications.
Int’l Journal of Ubiquitous Systems and Pervasive Networks, 9(1):13–21, 2017.
74
Bibliography
[71]
R. Pryss, T. Probst, W. Schlee, J. Schobel, B. Langguth, P. Neff, M. Spiliopoulou,
and M. Reichert. Mobile Crowdsensing for the Juxtaposition of Realtime Assess-
ments and Retrospective Reporting for Neuropsychiatric Symptoms. In 30th IEEE
Int’l Symp on Computer-Based Medical Systems (CBMS). IEEE Computer Society
Press, June 2017.
[72]
R. Pryss, M. Schickler, J. Schobel, M. Weilbach, P. Geiger, and M. Reichert.
Enabling Tracks in Location-Based Smart Mobile Augmented Reality Applications.
Procedia Computer Science, 110:207–214, 2017.
[73]
R. Pryss, T. Probst, W. Schlee, J. Schobel, B. Langguth, P. Neff, M. Spiliopoulou,
and M. Reichert. Prospective crowdsensing versus retrospective ratings of tinnitus
variability and tinnitus – stress associations based on the TrackYourTinnitus mobile
platform. Int’l Journal of Data Science and Analytics, March 2018.
[74]
R. Pryss, J. Schobel, and M. Reichert. Requirements for a Flexible and Generic API
Enabling Mobile Crowdsensing mHealth Applications. In 4th IEEE Int’l Workshop
on Requirements Engineering for Self-Adaptive, Collaborative, and Cyber Physical
Systems (RESACS). IEEE Computer Society Press, August 2018.
[75]
M. Raento, A. Oulasvirta, and N. Eagle. Smartphones: An Emerging Tool for
Social Scientists. Sociological Methods & Research, 37(3):426–454, 2009.
[76]
Q. A. Rahman, T. Janmohamed, M. Pirbaglou, P. Ritvo, J. M. Heffernan, H. Clarke,
and J. Katz. Patterns of User Engagement With the Mobile App, Manage My
Pain: Results of a Data Mining Investigation. JMIR mHealth and uHealth, 5(7):
e96, 2017.
[77]
M. Reichert and P. Dadam. Enabling Adaptive Process-Aware Information Systems
with ADEPT2. Handbook of Research on Business Process Modeling, pages 173–203,
2009.
[78]
M. Reichert and R. Pryss. Flexible Support of Healthcare Processes. In C. Combi,
G. Pozzi, and P. Veltri, editors, Process Modeling and Management for Healthcare,
pages 35–66. Taylor & Francis Group, November 2017.
[79]
M. Reichert and B. Weber. Enabling Flexibility in Process-Aware Information
Systems: Challenges, Methods, Technologies. Springer Science & Business Media,
2012.
[80]
G. Renaud and L. Azzopardi. SCAMP: A Tool for Conducting Interactive Infor-
mation Retrieval Experiments. In IIiX, pages 286–289, 2012.
[81]
C. Rieger and H. Kuchen. A Process-Oriented Modeling Approach for Graphical
Development of Mobile Business Apps. Computer Languages, Systems & Structures,
53:43–58, 2018.
75
Bibliography
[82]
S. Rinderle, M. Reichert, and P. Dadam. Evaluation of Correctness Criteria for
Dynamic Workflow Changes. In 1st Int’l Conf. on Business Process Management
(BPM), number 2678 in LNCS, pages 41–57. Springer, June 2003.
[83]
M. Ruf-Leuschner, R. Pryss, M. Liebrecht, J. Schobel, A. Spyridou, M. Reichert,
and M. Schauer. Preventing Further Trauma: KINDEX Mum Screen - Assessing
and Reacting Towards Psychosocial Risk Factors in Pregnant Women with the Help
of Smartphone Technologies. In XIII Congress of European Society of Traumatic
Stress Studies (ESTSS) Conf, pages 70–70, June 2013.
[84]
M. Ruf-Leuschner, N. Brunnemann, M. Schauer, R. Pryss, E. Barnewitz,
M. Liebrecht, W. Kratzer, M. Reichert, and T. Elbert. The KINDEX-App –
An Instrument for Assessment and Immediate Analysis of Psychosocial Risk Fac-
tors in Pregnant Women in Daily Practice by Gynecologists, Midwives and in
Gynecological Hospitals. Verhaltenstherapie, 26(3):171–181, 2016.
[85]
D. Ruiz-Fernández, D. Marcos-Jorquera, V. Gilart-Iglesias, V. Vives-Boix, and
J. Ramírez-Navarro. Empowerment of Patients with Hypertension through BPM,
IoT and Remote Sensing. Sensors, 17(10):2273, 2017.
[86]
J. Rumbaugh, I. Jacobson, and G. Booch. The Unified Modeling Language Reference
Manual. Pearson Higher Education, 2004.
[87]
N. Russell, A. H. Ter Hofstede, D. Edmond, and W. M. P. van der Aalst. Workflow
data patterns: Identification, representation and tool support. In Int’l Conf on
Conceptual Modeling, pages 353–368. Springer, 2005.
[88]
N. Russell, W. M. P. Van der Aalst, A. H. Ter Hofstede, and D. Edmond. Workflow
resource patterns: Identification, representation and tool support. In 17th Int’l Conf
on Advanced Information Systems Engineering (CAiSE), pages 216–232. Springer,
2005.
[89]
N. Russell, A. H. M. Ter Hofstede, W. M. P. Van Der Aalst, and N. Mulyar.
Workflow control-flow patterns: A revised view. BPM Center Report BPM-06-22,
BPMcenter. org, pages 06–22, 2006.
[90]
C. Scaffidi. The Impact of Human-Centric Design on the Adoption of Information
Systems: A Case Study of the Spreadsheet. In 11th Iberian Conf on Information
Systems and Technologies (CISTI), pages 1–7. IEEE, 2016.
[91]
C. Scaffidi, M. Shaw, and B. Myers. Estimating the Numbers of End Users and
End User Programmers. In IEEE Symp on Visual Languages and Human-Centric
Computing, pages 207–214. IEEE, 2005.
[92]
M. Schickler, R. Pryss, J. Schobel, and M. Reichert. An Engine Enabling Location-
based Mobile Augmented Reality Applications. In 10th Int’l Conf on Web Infor-
mation Systems and Technologies (Revised Selected Papers), number 226 in LNBIP,
pages 363–378. Springer, 2015.
76
Bibliography
[93]
M. Schickler, M. Reichert, R. Pryss, J. Schobel, W. Schlee, and B. Langguth.
Entwicklung mobiler Apps: Konzepte, Anwendungsbausteine und Werkzeuge im
Business und E-Health. eXamen.press. Springer Vieweg, October 2015.
[94]
M. Schickler, J. Schobel, R. Pryss, and M. Reichert. Mobile Crowd Sensing: A
New Way of Collecting Data from Trauma Samples? In XIV Congress of European
Society for Traumatic Stress Studies (ESTSS) Conf, page 244, June 2015.
[95]
M. Schickler, R. Pryss, M. Reichert, M. Heinzelmann, J. Schobel, B. Langguth,
T. Probst, and W. Schlee. Using Wearables in the Context of Chronic Disorders
- Results of a Pre-Study. In 29th IEEE Int’l Symp on Computer-Based Medical
Systems, pages 68–69, June 2016.
[96]
M. Schickler, R. Pryss, M. Reichert, J. Schobel, B. Langguth, and W. Schlee.
Using Mobile Serious Games in the Context of Chronic Disorders - A Mobile Game
Concept for the Treatment of Tinnitus. In 29th IEEE Int’l Symp on Computer-Based
Medical Systems (CBMS), pages 343–348, June 2016.
[97]
M. Schickler, R. Pryss, J. Schobel, and M. Reichert. Supporting Remote Therapeutic
Interventions with Mobile Processes. In 6th IEEE Int’l Conf on AI & Mobile Services
(AIMS). IEEE Computer Society Press, June 2017.
[98]
M. Schickler, R. Pryss, J. Schobel, W. Schlee, T. Probst, and M. Reichert. Towards
Flexible Remote Therapeutic Interventions. In 30th IEEE Int’l Symp on Computer-
Based Medical Systems (CBMS), pages 260–261. IEEE Computer Society Press,
June 2017.
[99]
M. Schickler, R. Pryss, M. Stach, J. Schobel, W. Schlee, T. Probst, B. Langguth,
and M. Reichert. An IT Platform Enabling Remote Therapeutic Interventions.
In 30th IEEE Int’l Symp on Computer-Based Medical Systems (CBMS), pages
111–116. IEEE Computer Society Press, June 2017.
[100]
M. Schickler, R. Pryss, W. Schlee, T. Probst, B. Langguth, J. Schobel, and
M. Reichert. Usability Study on Mobile Processes Enabling Remote Therapeutic
Interventions. In 31th IEEE Int’l Symp on Computer-Based Medical Systems
(CBMS). IEEE Computer Society Press, June 2018.
[101]
W. Schlee, R. Pryss, T. Probst, J. Schobel, A. Bachmeier, M. Reichert, and
B. Langguth. Measuring the Moment-to-Moment Variability of Tinnitus: The
TrackYourTinnitus Smart Phone App. Frontiers in Aging Neuroscience, 8:294–294,
December 2016.
[102]
W. Schmidt, C. Sarran, N. Ronan, G. Barrett, D. J. Whinney, L. E. Fleming, N. J.
Osborne, and J. Tyrrell. The Weather and Meniere’s Disease: A Longitudinal
Analysis in the UK. Otology & Neurotology, 38(2):225, 2017.
77
Bibliography
[103]
J. Schobel and M. Reichert. Business Process Intelligence Tools. In G. Grambow,
R. Oberhauser, and M. Reichert, editors, Advances in Intelligent Process-Aware In-
formation Systems: Concepts, Methods, and Technologies, volume 123 of Intelligent
Systems Reference Library, pages 225–249. Springer, May 2017.
[104]
J. Schobel and M. Reichert. A Predictive Approach Enabling Process Execution
Recommendations. In G. Grambow, R. Oberhauser, and M. Reichert, editors,
Advances in Intelligent Process-Aware Information Systems: Concepts, Methods,
and Technologies, volume 123 of Intelligent Systems Reference Library, pages
155–170. Springer, May 2017.
[105]
J. Schobel, M. Ruf-Leuschner, R. Pryss, M. Reichert, M. Schickler, M. Schauer,
R. Weierstall, D. Isele, C. Nandi, and T. Elbert. A Generic Questionnaire Framework
Supporting Psychological Studies with Smartphone Technologies. In XIII Congress
of European Society of Traumatic Stress Studies (ESTSS) Conf, pages 69–69, June
2013.
[106]
J. Schobel, M. Schickler, R. Pryss, H. Nienhaus, and M. Reichert. Using Vi-
tal Sensors in Mobile Healthcare Business Applications: Challenges, Examples,
Lessons Learned. In 9th Int’l Conf on Web Information Systems and Technologies
(WEBIST), Special Session on Business Apps, pages 509–518, May 2013.
[107]
J. Schobel, M. Schickler, R. Pryss, F. Maier, and M. Reichert. Towards Process-
Driven Mobile Data Collection Applications: Requirements, Challenges, Lessons
Learned. In 10th Int’l Conf on Web Information Systems and Technologies (WE-
BIST), Special Session on Business Apps, pages 371–382, April 2014.
[108]
J. Schobel, R. Pryss, and M. Reichert. Using Smart Mobile Devices for Collecting
Structured Data in Clinical Trials: Results From a Large-Scale Case Study. In
28th IEEE Int’l Symp on Computer-Based Medical Systems (CBMS), pages 13–18.
IEEE Computer Society Press, June 2015.
[109]
J. Schobel, M. Schickler, R. Pryss, and M. Reichert. Process-Driven Data Collection
with Smart Mobile Devices. In 10th Int’l Conf on Web Information Systems and
Technologies (Revised Selected Papers), number 226 in LNBIP, pages 347–362.
Springer, 2015.
[110]
J. Schobel, M. Schickler, R. Pryss, M. Reichert, and T. Elbert. A Domain-Specific
Framework for Collecting Data in Trials with Smart Mobile Devices. In XIV
Congress of European Society for Traumatic Stress Studies (ESTSS) Conf, June
2015.
[111]
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Flexible Mobile Data
Collection in Healthcare. In 29th IEEE Int’l Symp on Computer-Based Medical
Systems (CBMS), pages 181–182, June 2016.
78
Bibliography
[112]
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Lightweight Process Engine
for Enabling Advanced Mobile Applications. In 24th Int’l Conf on Cooperative
Information Systems (CoopIS), number 10033 in LNCS, pages 552–569. Springer,
October 2016.
[113]
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Configurator Component
for End-User Defined Mobile Data Collection Processes. In Demo Track of the 14th
Int’l Conf on Service Oriented Computing (ICSOC), October 2016.
[114]
J. Schobel, R. Pryss, M. Schickler, M. Ruf-Leuschner, T. Elbert, and M. Reichert.
End-User Programming of Mobile Services: Empowering Domain Experts to
Implement Mobile Data Collection Applications. In 5th IEEE Int’l Conf on Mobile
Services (MS), pages 1–8. IEEE Computer Society Press, May 2016.
[115]
J. Schobel, R. Pryss, W. Wipp, M. Schickler, and M. Reichert. A Mobile Service
Engine Enabling Complex Data Collection Applications. In 14th Int’l Conf on
Service Oriented Computing (ICSOC), number 9936 in LNCS, pages 626–633,
October 2016.
[116]
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Patterns for Defining
and Changing Data Collection Instruments in Mobile Healthcare Scenarios. In 30th
IEEE Int’l Symp on Computer-Based Medical Systems (CBMS), June 2017.
[117]
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Process-Driven Mobile Data
Collection (Extended Abstract). In 8th Int’l Workshop on Enterprise Modeling and
Information Systems Architectures (EMISA), June 2017.
[118]
J. Schobel, R. Pryss, W. Schlee, T. Probst, D. Gebhardt, M. Schickler, and
M. Reichert. Development of Mobile Data Collection Applications by Domain
Experts: Experimental Results from a Usability Study. In 29th Int’l Conf on
Advanced Information Systems Engineering (CAiSE), number 10253 in LNCS,
pages 60–75. Springer, June 2017.
[119]
J. Schobel, R. Pryss, T. Probst, W. Schlee, M. Schickler, and M. Reichert. Learn-
ability of a Configurator Empowering End Users to Create Mobile Data Collection
Instruments: Usability Study. JMIR mHealth and uHealth, 6(6):e148, 2018.
[120]
B. Shneiderman. Designing the User Interface: Strategies for Effective Human-
Computer Interaction. Pearson Education India, 2010.
[121]
M. Stach, R. Pryss, M. Schnitzlein, T. Mohring, M. Jurisch, and M. Reichert.
Lightweight Process Support with Spreadsheet-Driven Processes: A Case Study in
the Finance Domain. In 10th Workshop on Social and Human Aspects of Business
Process Management (BPMS), Workshops, pages 323–334, 2017.
[122]
J. Sweller. Cognitive Load During Problem Solving: Effects on Learning. Cognitive
Science, 12(2):257–285, 1988.
79
Bibliography
[123]
W. M. P. Van der Aalst. Process Discovery: An Introduction. In Process Mining,
pages 163–194. Springer, 2016.
[124]
J. J. Van Merrienboer and J. Sweller. Cognitive Load Theory and Complex Learning:
Recent Developments and Future Directions. Educational Psychology Review, 17
(2):147–177, 2005.
[125]
M. von Aster, A. Neubauer, and R. v. Horn. Wechsler Intelligenztest für Erwachsene:
WIE; Übersetzung und Adaption der WAIS-III. Harcourt Test Services, 2006.
[126]
B. Weber, J. Pinggera, M. Neurauter, S. Zugal, M. Martini, M. Furtner, P. Sachse,
and D. Schnitzer. Fixation Patterns during Process Model Creation: Initial Steps
Toward Neuro-Adaptive Process Modeling Environments. In 49th Hawaii Int’l
Conf on System Sciences (HICSS), pages 600–609. IEEE, 2016.
[127]
S. Weerawarana, F. Curbera, F. Leymann, T. Storey, and D. F. Ferguson. Web
Services Platform Architecture: SOAP, WSDL, WS-Policy, WS-Addressing, WS-
BPEL, WS-Reliable Messaging and More. Prentice Hall PTR, 2005.
[128]
M. Weske. Business Process Management: Concepts, Languages, Architectures.
Springer, 2012.
[129]
S. Wilker, A. Pfeiffer, S. Kolassa, T. Elbert, B. Lingenfelder, E. Ovuga, A. Papas-
sotiropoulos, D. de Quervain, and I.-T. Kolassa. The role of FKBP5 genotype in
moderating long-term effectiveness of exposure-based psychotherapy for posttrau-
matic stress disorder. Translational psychiatry, 4(6):e403, 2014.
[130]
C. Wohlin, P. Runeson, M. Höst, M. C. Ohlsson, B. Regnell, and A. Wesslén.
Experimentation in Software Engineering. Springer Science & Business Media,
2012.
[131]
M. Zhang, E. Cheow, C. S. Ho, B. Y. Ng, R. Ho, and C. C. S. Cheok. Application
of Low-Cost Methodologies for Mobile Phone App Development. JMIR mHealth
and uHealth, 2(4):e55, 2014.
[132]
M. W. Zhang, T. Tsang, E. Cheow, C. S. Ho, N. B. Yeong, and R. C. Ho. Enabling
Psychiatrists to be Mobile Phone App Developers: Insights into App Development
Methodologies. JMIR mHealth and uHealth, 2(4):e53, 2014.
[133]
L. Zhou, J. Bao, and B. Parmanto. Systematic Review Protocol to Assess the
Effectiveness of Usability Questionnaires in mHealth App Studies. JMIR Research
Protocols, 6(8):e151, 2017.
[134]
M. Zimoch, R. Pryss, T. Probst, W. Schlee, and M. Reichert. Cognitive Insights
into Business Process Model Comprehension: Preliminary Results for Experienced
and Inexperienced Individuals. In Enterprise, Business-Process and Information
Systems Modeling, pages 137–152. Springer, 2017.
80
Bibliography
[135]
M. Zimoch, R. Pryss, J. Schobel, and M. Reichert. Eye Tracking Experiments on
Process Model Comprehension: Lessons Learned. In 18th Int’l Conf on Business
Process Modeling, Development, and Support (BPMDS), number 287 in LNBIP,
pages 153–168. Springer, June 2017.
[136]
M. zur Muehlen and J. Recker. How Much Language Is Enough? Theoretical and
Practical Use of the Business Process Modeling Notation. In Z. Bellahsène and
M. Léonard, editors, 20th Int’l Conf on Advanced Information Systems Engineering
(CAiSE), number 5074 in LNCS, pages 465–479. Springer, June 2008.
81
Index
Business Process
Execution Language, 27
Management, 7, 29, 31
Modeling and Notation, 27
Cascading Style Sheets, 42
Client, 29, 32
Cognitive Load Theory, 57
Configurator, 29, 31, 58, 64
Design Time, 12
Domain-Specific Language, 15, 23, 27, 31,
35
End-User, 14, 54
Development, 15
Programming, 14, 23, 35, 58, 67
Executable Component, 32, 35, 64, 67
Experience Sampling Method, 43
Goal Question Metric, 50
HTML, 42
Instrument, 9
Mode, 10
Structure, 9
JSON, 18
Learnability, 49, 50, 54, 56–58, 67
Mental Effort, 57
Mobile Data Collection, 14
Lifecycle, 25
Model-Driven Development, 23, 26
Post-Traumatic Stress Disorder, 24
Process
Engine, 24, 29, 32, 64
Instance, 13
Management, 66
Mining, 66
Model, 12, 18, 23, 26, 29, 35, 54, 57
Schema, 11
Process-Aware Information System, 11
QuestionSys Framework, 6, 29, 64
Representational State Transfer, 18, 30
Run Time, 13, 24
Sensor, 13, 67
Short Message Service, 14
Structured Query Language, 15
UML, 18, 26
Usability, 50
Study, 7, 37, 64, 67
System Usability Scale, 58
Web Application, 14
WYSIWYG, 42, 43
XML, 12, 30
83
Acronyms
ADEPT Application Development Based on Pre-Modeled Process Templates
API Application Programming Interface
BPEL Business Process Execution Language
BPM Business Process Management
BPMN Business Process Modeling and Notation
CLT Cognitive Load Theory
CSS Cascading Style Sheets
DHTML Dynamic HTML
DSL Domain-Specific Language
EC Executable Component
EMA Ecological Momentary Assessment
EPC Event-driven Process Chain
ESM Experience Sampling Method
EUD End-User Development
EUP End-User Programming
GPS Global Positioning System
GQM Goal Question Metric
HTML Hypertext Markup Language
JSON JavaScript Object Notation
MDC Mobile Data Collection
MDD Model-Driven Development
PAIS Process-Aware Information System
PTSD Post-Traumatic Stress Disorder
85
Index
REST Representational State Transfer
SMS Short Message Service
SQL Structured Query Language
SUS System Usability Scale
UML Unified Modeling Language
WS Web Service
WYSIWYG What You See Is What You Get
XML Extensible Markup Language
86
Part V
Appendix
87
A
Appendix Files
A.1 Example Modeling Task Description
89
Task: Patient Information Questionnaire
In this task you should model a questionnaire asking for information needed in the context of a medical
intervention (e.g., a gastroscopy) and educate the patient about possible risks.
Carefully read the text before starting modeling. If you have read the task to be modeled, please start the
QuestionSys configurator application via the provided shortcut on your desktop.
Select the following workspace:
Workspace: Study
Questionnaire: Patient Information
Open the “Editor” view and start modeling the respective questionnaire. Save the final model to the desk-
top of your computer.
1. The first page of the questionnaire contains a headline and text element with general information
regarding the upcoming medical intervention (e.g., gastroscopy). Furthermore, demographic infor-
mation of the patient (e.g., name, age, gender, …) shall be collected. Additionally, the patient should
answer, whether a family member shall be contacted after the intervention. If “yes”, the patient
shall continue with page 2, otherwise with page 3.
2. The second page shall only be displayed if the patient wants a family member to be informed re-
garding the course of the intervention. This page shall ask about details of the person to be informed
(e.g., name, phone number, …).
3. Regardless of whether one should be informed, the 3rd page will be displayed next. This page con-
tains a text if the patient wishes to be anesthetized or not. Thereby, the following options may be
available: none, local, full.
4. This page, in turn, shall only be displayed if the intervention takes place under local anesthesia.
Thereby, an additional form shall be displayed to provide information regarding possible risks. The
patient has to sign this form in order to continue.
5. This page, in turn, shall only be displayed if the intervention takes place under full anesthesia. It
shall display similar information as described before, however, texts shall be adapted in order to
reflect the given circumstances.
This task description was originally provided in German and was translated to English for the publication J. Schobel, R. Pryss,
T. Probst, W. Schlee, M. Schickler, and M. Reichert. Learnability of a Configurator Empowering End Users to Create Mobile
Data Collection Instruments: Usability Study. JMIR mHealth and uHealth, 6(6):e148, 2018
Page 2 / 2
Mental Effort: Patient Information Questionnaire
Answer the following questions:
1. The mental effort for creating the model was considerably high.
strongly
agree
agree
rather
agree
neutral
rather
disagree
disagree
strongly
disagree
○
○
○
○
○
○
○
2. I was able to properly solve the given task.
strongly
agree
agree
rather
agree
neutral
rather
disagree
disagree
strongly
disagree
○
○
○
○
○
○
○
3. The task was rather difficult.
strongly
agree
agree
rather
agree
neutral
rather
disagree
disagree
strongly
disagree
○
○
○
○
○
○
○
4. I had to concentrate myself when creating the model.
strongly
agree
agree
rather
agree
neutral
rather
disagree
disagree
strongly
disagree
○
○
○
○
○
○
○
5. Creating the model was exhausting.
strongly
agree
agree
rather
agree
neutral
rather
disagree
disagree
strongly
disagree
○
○
○
○
○
○
○
This task description was originally provided in German and was translated to English for the publication J. Schobel, R. Pryss,
T. Probst, W. Schlee, M. Schickler, and M. Reichert. Learnability of a Configurator Empowering End Users to Create Mobile
Data Collection Instruments: Usability Study. JMIR mHealth and uHealth, 6(6):e148, 2018
B
List of Publications
This appendix contains all publications that are part of this Ph. D. thesis.
SPR15
J. Schobel, R. Pryss, and M. Reichert. Using Smart Mobile Devices
for Collecting Structured Data in Clinical Trials: Results From a
Large-Scale Case Study. In 28th IEEE Int’l Symp on Computer-
Based Medical Systems (CBMS), pages 13–18. IEEE Computer
Society Press, June 2015 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 95
SSPR15
J. Schobel, M. Schickler, R. Pryss, and M. Reichert. Process-Driven
Data Collection with Smart Mobile Devices. In 10th Int’l Conf
on Web Information Systems and Technologies (Revised Selected
Papers), number 226 in LNBIP, pages 347–362. Springer, 2015 95
SPSR16
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Lightweight
Process Engine for Enabling Advanced Mobile Applications. In
24th Int’l Conf on Cooperative Information Systems (CoopIS),
number 10033 in LNCS, pages 552–569. Springer, October 2016
96
SPSR16a
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Configurator
Component for End-User Defined Mobile Data Collection Pro-
cesses. In Demo Track of the 14th Int’l Conf on Service Oriented
Computing (ICSOC), October 2016 . . . . . . . . . . . . . . . . . . . . . . . . . .. 96
SPSR16b
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards
Flexible Mobile Data Collection in Healthcare. In 29th IEEE
Int’l Symp on Computer-Based Medical Systems (CBMS), pages
181–182, June 2016 . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . .. 97
93
B List of Publications
SPWSR16
J. Schobel, R. Pryss, W. Wipp, M. Schickler, and M. Reichert. A
Mobile Service Engine Enabling Complex Data Collection Applica-
tions. In 14th Int’l Conf on Service Oriented Computing (ICSOC),
number 9936 in LNCS, pages 626–633, October 2016 . . . . . . . . .. 97
SPSRER16
J. Schobel, R. Pryss, M. Schickler, M. Ruf-Leuschner, T. Elbert,
and M. Reichert. End-User Programming of Mobile Services: Em-
powering Domain Experts to Implement Mobile Data Collection
Applications. In 5th IEEE Int’l Conf on Mobile Services (MS),
pages 1–8. IEEE Computer Society Press, May 2016 .. . . . . .. . . 98
SPSR17a
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards
Patterns for Defining and Changing Data Collection Instruments
in Mobile Healthcare Scenarios. In 30th IEEE Int’l Symp on
Computer-Based Medical Systems (CBMS), June 2017 . . . . . . . . . 98
SPSPGSR17 J. Schobel, R. Pryss, W. Schlee, T. Probst, D. Gebhardt, M. Schick-
ler, and M. Reichert. Development of Mobile Data Collection
Applications by Domain Experts: Experimental Results from a
Usability Study. In 29th Int’l Conf on Advanced Information Sys-
tems Engineering (CAiSE), number 10253 in LNCS, pages 60–75.
Springer, June 2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 99
SPPSSR18
J. Schobel, R. Pryss, T. Probst, W. Schlee, M. Schickler, and
M. Reichert. Learnability of a Configurator Empowering End
Users to Create Mobile Data Collection Instruments: Usability
Study. JMIR mHealth and uHealth, 6(6):e148, 2018 . . . . . . . . . . . 99
94
B.1 Using Smart Mobile Devices for Collecting Structured Data in Clinical Trials: Results From
a Large-Scale Case Study
B.1 Using Smart Mobile Devices for Collecting Structured
Data in Clinical Trials: Results From a Large-Scale Case
Study
The following article was published as follows:
J. Schobel, R. Pryss, and M. Reichert. Using Smart Mobile Devices for Collecting
Structured Data in Clinical Trials: Results From a Large-Scale Case Study. In 28th IEEE
Int’l Symp on Computer-Based Medical Systems (CBMS), pages 13–18. IEEE Computer
Society Press, June 2015
The original article is available at:
https://ieeexplore.ieee.org/document/7167446/
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://ieeexplore.ieee.org/document/7167446/
B.2 Process-Driven Data Collection with Smart Mobile Devices
The following article was published as follows:
J. Schobel, M. Schickler, R. Pryss, and M. Reichert. Process-Driven Data Collection with
Smart Mobile Devices. In 10th Int’l Conf on Web Information Systems and Technologies
(Revised Selected Papers), number 226 in LNBIP, pages 347–362. Springer, 2015
The original article is available at:
https://link.springer.com/chapter/10.1007/978-3-319-27030-2_22
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://link.springer.com/chapter/10.1007/978-3-319-27030-2_22
95
B List of Publications
B.3 A Lightweight Process Engine for Enabling Advanced
Mobile Applications
The following article was published as follows:
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Lightweight Process Engine for
Enabling Advanced Mobile Applications. In 24th Int’l Conf on Cooperative Information
Systems (CoopIS), number 10033 in LNCS, pages 552–569. Springer, October 2016
The original article is available at:
https://link.springer.com/chapter/10.1007/978-3-319-48472-3_33
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://link.springer.com/chapter/10.1007/978-3-319-48472-3_33
B.4 A Configurator Component for End-User Defined Mobile
Data Collection Processes
The following article was published as follows:
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Configurator Component for
End-User Defined Mobile Data Collection Processes. In Demo Track of the 14th Int’l
Conf on Service Oriented Computing (ICSOC), October 2016
The original article is available at:
https://link.springer.com/chapter/10.1007/978-3-319-68136-8_28
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://link.springer.com/chapter/10.1007/978-3-319-68136-8_28
96
B.5 Towards Flexible Mobile Data Collection in Healthcare
B.5 Towards Flexible Mobile Data Collection in Healthcare
The following article was published as follows:
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Flexible Mobile Data
Collection in Healthcare. In 29th IEEE Int’l Symp on Computer-Based Medical Systems
(CBMS), pages 181–182, June 2016
The original article is available at:
https://ieeexplore.ieee.org/document/7545980/
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://ieeexplore.ieee.org/document/7545980/
B.6 A Mobile Service Engine Enabling Complex Data
Collection Applications
The following article was published as follows:
J. Schobel, R. Pryss, W. Wipp, M. Schickler, and M. Reichert. A Mobile Service
Engine Enabling Complex Data Collection Applications. In 14th Int’l Conf on Service
Oriented Computing (ICSOC), number 9936 in LNCS, pages 626–633, October 2016
The original article is available at:
https://link.springer.com/chapter/10.1007/978-3-319-46295-0_42
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://link.springer.com/chapter/10.1007/978-3-319-46295-0_42
97
B List of Publications
B.7 End-User Programming of Mobile Services: Empowering
Domain Experts to Implement Mobile Data Collection
Applications
The following article was published as follows:
J. Schobel, R. Pryss, M. Schickler, M. Ruf-Leuschner, T. Elbert, and M. Reichert.
End-User Programming of Mobile Services: Empowering Domain Experts to Implement
Mobile Data Collection Applications. In 5th IEEE Int’l Conf on Mobile Services (MS),
pages 1–8. IEEE Computer Society Press, May 2016
The original article is available at:
https://ieeexplore.ieee.org/document/7787028/
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://ieeexplore.ieee.org/document/7787028/
B.8 Towards Patterns for Defining and Changing Data
Collection Instruments in Mobile Healthcare Scenarios
The following article was published as follows:
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Patterns for Defining and
Changing Data Collection Instruments in Mobile Healthcare Scenarios. In 30th IEEE
Int’l Symp on Computer-Based Medical Systems (CBMS), June 2017
The original article is available at:
https://ieeexplore.ieee.org/document/8104165/
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://ieeexplore.ieee.org/document/8104165/
98
B.9 Development of Mobile Data Collection Applications by Domain Experts: Experimental
Results from a Usability Study
B.9 Development of Mobile Data Collection Applications by
Domain Experts: Experimental Results from a Usability
Study
The following article was published as follows:
J. Schobel, R. Pryss, W. Schlee, T. Probst, D. Gebhardt, M. Schickler, and M. Reichert.
Development of Mobile Data Collection Applications by Domain Experts: Experimental
Results from a Usability Study. In 29th Int’l Conf on Advanced Information Systems
Engineering (CAiSE), number 10253 in LNCS, pages 60–75. Springer, June 2017
The original article is available at:
https://link.springer.com/chapter/10.1007/978-3-319-59536-8_5
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://link.springer.com/chapter/10.1007/978-3-319-59536-8_5
B.10 Learnability of a Configurator Empowering End Users to
Create Mobile Data Collection Instruments: Usability
Study
The following article was published as follows:
J. Schobel, R. Pryss, T. Probst, W. Schlee, M. Schickler, and M. Reichert. Learn-
ability of a Configurator Empowering End Users to Create Mobile Data Collection
Instruments: Usability Study. JMIR mHealth and uHealth, 6(6):e148, 2018
The original article is available at:
https://mhealth.jmir.org/2018/6/e148
99
B List of Publications
Due to copyright restrictions the paper has been removed from this version.
Please refer to the following address for the publishers version of the article:
https://mhealth.jmir.org/2018/6/e148
100
C
Complete List of Publications
The following is the complete list of publications the author of this Ph. D. thesis has been
involved in.
Peer-Reviewed Journals
2016
•
R. Pryss, P. Geiger, M. Schickler, J. Schobel, and M. Reichert. Advanced Algo-
rithms for Location-Based Smart Mobile Augmented Reality Applications. Procedia
Computer Science, 94:97–104, 2016
•
W. Schlee, R. Pryss, T. Probst, J. Schobel, A. Bachmeier, M. Reichert, and
B. Langguth. Measuring the Moment-to-Moment Variability of Tinnitus: The
TrackYourTinnitus Smart Phone App. Frontiers in Aging Neuroscience, 8:294–294,
December 2016
2017
•
T. Probst, R. Pryss, B. Langguth, J. Rauschecker, J. Schobel, M. Reichert,
M. Spiliopoulou, W. Schlee, and J. Zimmermann. Does tinnitus depend on time-
of-day? An ecological momentary assessment study with the ”TrackYourTinnitus”
application. Frontiers in Aging Neuroscience, 9:253–253, 2017
101
C Complete List of Publications
•
T. Probst, R. Pryss, B. Langguth, M. Spiliopoulou, M. Landgrebe, M. Vesala,
S. Harrison, J. Schobel, M. Reichert, M. Stach, and W. Schlee. Outpatient Tinnitus
Clinic, Self-Help Web Platform, or Mobile Application to Recruit Tinnitus Study
Samples? Frontiers in Aging Neuroscience, 9:113–113, April 2017
•
R. Pryss, M. Schickler, J. Schobel, M. Weilbach, P. Geiger, and M. Reichert.
Enabling Tracks in Location-Based Smart Mobile Augmented Reality Applications.
Procedia Computer Science, 110:207–214, 2017
•
R. Pryss, P. Geiger, M. Schickler, J. Schobel, and M. Reichert. The AREA
Framework for Location-Based Smart Mobile Augmented Reality Applications.
Int’l Journal of Ubiquitous Systems and Pervasive Networks, 9(1):13–21, 2017
2018
•
J. Schobel, R. Pryss, T. Probst, W. Schlee, M. Schickler, and M. Reichert. Learn-
ability of a Configurator Empowering End Users to Create Mobile Data Collection
Instruments: Usability Study. JMIR mHealth and uHealth, 6(6):e148, 2018
•
R. Pryss, T. Probst, W. Schlee, J. Schobel, B. Langguth, P. Neff, M. Spiliopoulou,
and M. Reichert. Prospective crowdsensing versus retrospective ratings of tinnitus
variability and tinnitus – stress associations based on the TrackYourTinnitus mobile
platform. Int’l Journal of Data Science and Analytics, March 2018
Peer-Reviewed Conferences
2013
•
J. Schobel, M. Schickler, R. Pryss, H. Nienhaus, and M. Reichert. Using Vital Sensors
in Mobile Healthcare Business Applications: Challenges, Examples, Lessons Learned.
In 9th Int’l Conf on Web Information Systems and Technologies (WEBIST), Special
Session on Business Apps, pages 509–518, May 2013
•
J. Schobel, M. Ruf-Leuschner, R. Pryss, M. Reichert, M. Schickler, M. Schauer,
R. Weierstall, D. Isele, C. Nandi, and T. Elbert. A Generic Questionnaire Framework
Supporting Psychological Studies with Smartphone Technologies. In XIII Congress
of European Society of Traumatic Stress Studies (ESTSS) Conf, pages 69–69, June
2013
•
M. Ruf-Leuschner, R. Pryss, M. Liebrecht, J. Schobel, A. Spyridou, M. Reichert,
and M. Schauer. Preventing Further Trauma: KINDEX Mum Screen - Assessing
and Reacting Towards Psychosocial Risk Factors in Pregnant Women with the Help
of Smartphone Technologies. In XIII Congress of European Society of Traumatic
Stress Studies (ESTSS) Conf, pages 70–70, June 2013
102
•
D. Isele, M. Ruf-Leuschner, R. Pryss, M. Schauer, M. Reichert, J. Schobel, A. Schindler,
and T. Elbert. Detecting Adverse Childhood Experiences with a Little Help from
Tablet Computers. In XIII Congress of European Society of Traumatic Stress
Studies (ESTSS) Conf, pages 69–70, June 2013
2014
•
J. Schobel, M. Schickler, R. Pryss, F. Maier, and M. Reichert. Towards Process-
Driven Mobile Data Collection Applications: Requirements, Challenges, Lessons
Learned. In 10th Int’l Conf on Web Information Systems and Technologies (WE-
BIST), Special Session on Business Apps, pages 371–382, April 2014
•
P. Geiger, M. Schickler, R. Pryss, J. Schobel, and M. Reichert. Location-based
Mobile Augmented Reality Applications: Challenges, Examples, Lessons Learned. In
10th Int’l Conf on Web Information Systems and Technologies (WEBIST), Special
Session on Business Apps, pages 383–394, April 2014
2015
•
J. Schobel, M. Schickler, R. Pryss, M. Reichert, and T. Elbert. A Domain-Specific
Framework for Collecting Data in Trials with Smart Mobile Devices. In XIV
Congress of European Society for Traumatic Stress Studies (ESTSS) Conf, June
2015
•
M. Schickler, J. Schobel, R. Pryss, and M. Reichert. Mobile Crowd Sensing: A
New Way of Collecting Data from Trauma Samples? In XIV Congress of European
Society for Traumatic Stress Studies (ESTSS) Conf, page 244, June 2015
•
J. Schobel, R. Pryss, and M. Reichert. Using Smart Mobile Devices for Collecting
Structured Data in Clinical Trials: Results From a Large-Scale Case Study. In 28th
IEEE Int’l Symp on Computer-Based Medical Systems (CBMS), pages 13–18. IEEE
Computer Society Press, June 2015
•
J. Schobel, M. Schickler, R. Pryss, and M. Reichert. Process-Driven Data Collection
with Smart Mobile Devices. In 10th Int’l Conf on Web Information Systems and
Technologies (Revised Selected Papers), number 226 in LNBIP, pages 347–362.
Springer, 2015
•
M. Schickler, R. Pryss, J. Schobel, and M. Reichert. An Engine Enabling Location-
based Mobile Augmented Reality Applications. In 10th Int’l Conf on Web Infor-
mation Systems and Technologies (Revised Selected Papers), number 226 in LNBIP,
pages 363–378. Springer, 2015
103
C Complete List of Publications
2016
•
M. Schickler, R. Pryss, M. Reichert, J. Schobel, B. Langguth, and W. Schlee.
Using Mobile Serious Games in the Context of Chronic Disorders - A Mobile Game
Concept for the Treatment of Tinnitus. In 29th IEEE Int’l Symp on Computer-Based
Medical Systems (CBMS), pages 343–348, June 2016
•
M. Schickler, R. Pryss, M. Reichert, M. Heinzelmann, J. Schobel, B. Langguth,
T. Probst, and W. Schlee. Using Wearables in the Context of Chronic Disorders
- Results of a Pre-Study. In 29th IEEE Int’l Symp on Computer-Based Medical
Systems, pages 68–69, June 2016
•
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Flexible Mobile Data
Collection in Healthcare. In 29th IEEE Int’l Symp on Computer-Based Medical
Systems (CBMS), pages 181–182, June 2016
•
J. Schobel, R. Pryss, M. Schickler, M. Ruf-Leuschner, T. Elbert, and M. Reichert.
End-User Programming of Mobile Services: Empowering Domain Experts to Im-
plement Mobile Data Collection Applications. In 5th IEEE Int’l Conf on Mobile
Services (MS), pages 1–8. IEEE Computer Society Press, May 2016
•
J. Schobel, R. Pryss, W. Wipp, M. Schickler, and M. Reichert. A Mobile Service
Engine Enabling Complex Data Collection Applications. In 14th Int’l Conf on
Service Oriented Computing (ICSOC), number 9936 in LNCS, pages 626–633,
October 2016
•
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Lightweight Process Engine
for Enabling Advanced Mobile Applications. In 24th Int’l Conf on Cooperative
Information Systems (CoopIS), number 10033 in LNCS, pages 552–569. Springer,
October 2016
2017
•
M. Zimoch, R. Pryss, J. Schobel, and M. Reichert. Eye Tracking Experiments on
Process Model Comprehension: Lessons Learned. In 18th Int’l Conf on Business
Process Modeling, Development, and Support (BPMDS), number 287 in LNBIP,
pages 153–168. Springer, June 2017
•
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Process-Driven Mobile Data
Collection (Extended Abstract). In 8th Int’l Workshop on Enterprise Modeling and
Information Systems Architectures (EMISA), June 2017
•
J. Schobel, R. Pryss, W. Schlee, T. Probst, D. Gebhardt, M. Schickler, and M. Re-
ichert. Development of Mobile Data Collection Applications by Domain Experts:
Experimental Results from a Usability Study. In 29th Int’l Conf on Advanced
104
Information Systems Engineering (CAiSE), number 10253 in LNCS, pages 60–75.
Springer, June 2017
•
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. Towards Patterns for Defining
and Changing Data Collection Instruments in Mobile Healthcare Scenarios. In 30th
IEEE Int’l Symp on Computer-Based Medical Systems (CBMS), June 2017
•
M. Schickler, R. Pryss, J. Schobel, and M. Reichert. Supporting Remote Therapeutic
Interventions with Mobile Processes. In 6th IEEE Int’l Conf on AI & Mobile Services
(AIMS). IEEE Computer Society Press, June 2017
•
M. Schickler, R. Pryss, J. Schobel, W. Schlee, T. Probst, and M. Reichert. Towards
Flexible Remote Therapeutic Interventions. In 30th IEEE Int’l Symp on Computer-
Based Medical Systems (CBMS), pages 260–261. IEEE Computer Society Press,
June 2017
•
M. Schickler, R. Pryss, M. Stach, J. Schobel, W. Schlee, T. Probst, B. Langguth,
and M. Reichert. An IT Platform Enabling Remote Therapeutic Interventions.
In 30th IEEE Int’l Symp on Computer-Based Medical Systems (CBMS), pages
111–116. IEEE Computer Society Press, June 2017
•
R. Pryss, T. Probst, W. Schlee, J. Schobel, B. Langguth, P. Neff, M. Spiliopoulou,
and M. Reichert. Mobile Crowdsensing for the Juxtaposition of Realtime Assess-
ments and Retrospective Reporting for Neuropsychiatric Symptoms. In 30th IEEE
Int’l Symp on Computer-Based Medical Systems (CBMS). IEEE Computer Society
Press, June 2017
2018
•
M. Schickler, R. Pryss, W. Schlee, T. Probst, B. Langguth, J. Schobel, and M. Re-
ichert. Usability Study on Mobile Processes Enabling Remote Therapeutic Inter-
ventions. In 31th IEEE Int’l Symp on Computer-Based Medical Systems (CBMS).
IEEE Computer Society Press, June 2018
Books
2015
•
M. Schickler, M. Reichert, R. Pryss, J. Schobel, W. Schlee, and B. Langguth.
Entwicklung mobiler Apps: Konzepte, Anwendungsbausteine und Werkzeuge im
Business und E-Health. eXamen.press. Springer Vieweg, October 2015
105
C Complete List of Publications
Book Chapters
2017
•J. Schobel and M. Reichert. Business Process Intelligence Tools. In G. Grambow,
R. Oberhauser, and M. Reichert, editors, Advances in Intelligent Process-Aware In-
formation Systems: Concepts, Methods, and Technologies, volume 123 of Intelligent
Systems Reference Library, pages 225–249. Springer, May 2017
•
J. Schobel and M. Reichert. A Predictive Approach Enabling Process Execution
Recommendations. In G. Grambow, R. Oberhauser, and M. Reichert, editors,
Advances in Intelligent Process-Aware Information Systems: Concepts, Methods,
and Technologies, volume 123 of Intelligent Systems Reference Library, pages 155–
170. Springer, May 2017
Demo Tracks
2016
•
J. Schobel, R. Pryss, M. Schickler, and M. Reichert. A Configurator Component
for End-User Defined Mobile Data Collection Processes. In Demo Track of the 14th
Int’l Conf on Service Oriented Computing (ICSOC), October 2016
Workshops
2018
•
R. Pryss, J. Schobel, and M. Reichert. Requirements for a Flexible and Generic API
Enabling Mobile Crowdsensing mHealth Applications. In 4th IEEE Int’l Workshop
on Requirements Engineering for Self-Adaptive, Collaborative, and Cyber Physical
Systems (RESACS). IEEE Computer Society Press, August 2018
106
D
Discussion of Personal Contribution
The following is a full list of all researchers contributing to this cumulative Ph. D. thesis
(alphabetically ordered).
Prof. Dr. Thomas Elbert Universität Konstanz
Fachbereich Psychologie
Feuersteinstr. 55, Haus 22
78479 Reichenau (Germany)
Prof. Dr. Thomas Probst Donau-Universität Krems
Department für Psychotherapie und Biopsychosoziale
Gesundheit
Dr.-Karl-Dorrek-Straße 30, Trakt H, EG, Raum H 0.51
3500 Krems an der Donau (Austria)
Dr. Rüdiger Pryss Universität Ulm
Institut für Datenbanken und Informationssysteme
James-Franck-Ring, Geb. O27/5101
89081 Ulm (Germany)
Prof. Dr. Manfred Reichert Universität Ulm
Institut für Datenbanken und Informationssysteme
James-Franck-Ring, Geb. O27/523
89081 Ulm (Germany)
107
D Discussion of Personal Contribution
Dr. Martina Ruf-Leuschner Universität Konstanz
Fachbereich Psychologie
Feuersteinstr. 55, Haus 22
78479 Reichenau (Germany)
Marc Schickler Universität Ulm
Institut für Datenbanken und Informationssysteme
James-Franck-Ring, Geb. O27/5101
89081 Ulm (Germany)
Dr. Winfried Schlee Universität Regensburg
Lehrstuhl Psychiatrie und Psychotherapie
Universitätsstraße 84, Haus 29, Raum 001
93053 Regensburg (Germany)
The following is a list of students contributing to this cumulative Ph. D. thesis (alphabet-
ically ordered).
Dominic Gebhardt Kolpingstraße 22a
87459 Pfronten (Germany)
Fabian Maier Zeughausgasse 5
89073 Ulm (Germany)
Wolfgang Wipp Am Weiher 1
88709 Meersburg (Germany)
108
D.1 Using Smart Mobile Devices for Collecting Structured Data in Clinical Trials: Results From
a Large-Scale Case Study
The contribution of the individual authors for respective publications that are part of
this Ph. D. thesis are discussed in the following:
D.1 Using Smart Mobile Devices for Collecting Structured
Data in Clinical Trials: Results From a Large-Scale Case
Study
The author of this Ph. D. thesis was responsible for assessing the requirements in the
context of the presented mobile application to assist researchers in collecting data in
their psychological trials. In this context, requirement analysis, coordination of the
development process as well as the deployment was supervised by the Ph. D. candidate.
Dr. Pryss advised the candidate in the development of a proper strategy for dealing with
such long-running projects. Both, Dr. Pryss and Prof. Reichert provided assistance in
proof-reading the draft of the manuscript.
D.2 Process-Driven Data Collection with Smart Mobile
Devices
The author of this Ph. D. thesis was responsible for conducting interviews with experts
from various domains in order to document requirements. In this context, Dr. Pryss gave
valuable suggestions on how to properly structure respective interviews. The candidate
developed a mental model that allows for mapping an instrument to a process model.
Both Prof. Reichert and Dr. Pryss suggested improvements for the developed model.
The Ph. D. candidate developed a first prototype application using this mental model.
All authors helped proof-reading the manuscript.
D.3 A Lightweight Process Engine for Enabling Advanced
Mobile Applications
The Ph. D. candidate was responsible for developing the theoretical architecture of the
lightweight process engine running on smart mobile devices. In this context, various
modules have been designed by the candidate. Moreover, the concept of using Executable
Components that are controlled and coordinated by the process engine was developed by
the Ph. D. candidate. Further, a concept of automatically evaluating the collected data
based on rules were drafted and realized by the candidate. Dr. Pryss assisted the Ph. D.
candidate with designing the software architecture for the mobile process engine and its
data analysis component. All authors drafted and helped proof-reading the manuscript.
109
D Discussion of Personal Contribution
D.4 A Configurator Component for End-User Defined Mobile
Data Collection Processes
The Ph. D. candidate was responsible for implementing respective configurator component
of the QuestionSys framework. Dr. Pryss and Mr. Schickler, in turn, helped the candidate
to develop the architecture of the framework on a conceptual level in numerous meetings.
All authors helped proof-reading the manuscript and gave feedback.
D.5 Towards Flexible Mobile Data Collection in Healthcare
The Ph. D. candidate defined flexibility aspects along different phases of data collection
scenarios. In this context, techniques enabling flexibility in various realized mobile data
collection applications were assessed by the candidate. Dr. Pryss advises the candidate
to further refine and properly classify the discovered techniques. All authors drafted and
revised the publication.
D.6 A Mobile Service Engine Enabling Complex Data
Collection Applications
The Ph. D. candidate was responsible for developing the theoretical architecture of the
lightweight process engine running on smart mobile devices. In this context, various
modules have been designed by the candidate. Moreover, the concept of using Executable
Components that are controlled and coordinated by the process engine was developed by
the Ph. D. candidate. Further, he assisted and supervised Mr. Wipp, who developed the
proof-of-concept implementation. Dr. Pryss assisted the Ph. D. candidate with designing
the software architecture for the mobile process engine. All authors drafted and helped
proof-reading the manuscript.
D.7 End-User Programming of Mobile Services: Empowering
Domain Experts to Implement Mobile Data Collection
Applications
The Ph. D. candidate was responsible for developing a sophisticated lifecycle comprising
common phases of various data collection scenarios. Most of these scenarios, in turn, were
supported by the candidate by either implementing mobile data collection applications
or providing management assistance and guidance in order to properly realize such real-
world scenarios. Dr. Ruf-Leuschner and Prof. Elbert helped by providing sophisticated
insights into various psychological studies and assisted the candidate in elaborating
110
D.8 Towards Patterns for Defining and Changing Data Collection Instruments in Mobile
Healthcare Scenarios
domain specific requirements. Dr. Pryss and Mr. Schickler, in turn, helped the candidate
to develop the architecture of the framework on a conceptual level in numerous meetings.
The Ph. D. candidate implemented the prototypes of the QuestionSys framework. All
authors helped proof-reading the manuscript and gave feedback.
D.8 Towards Patterns for Defining and Changing Data
Collection Instruments in Mobile Healthcare Scenarios
The author of this Ph. D. thesis was responsible for creating an initial list of patterns
allowing for adapting data collection instruments. Moreover, these patterns were extracted
by the candidate by evaluating various realized mobile data collection applications. Dr.
Pryss advised the candidate in classifying the identified patterns and applying a proper
scientific methodology with respect to their validation. All authors were responsible for
refining the initial list of patterns as well as proof-reading the manuscript.
D.9 Development of Mobile Data Collection Applications by
Domain Experts: Experimental Results from a Usability
Study
The Ph. D. candidate was responsible for implementing the required features for the
configurator component in order to assess performance measures of participants. Dr.
Schlee advised the candidate with respect to the study design, whereas Mr. Gebhardt
assisted the candidate during the study procedure and helped collecting data from
recruited participants. Prof. Probst helped with analyzing the collected data and
interpreting the results. Dr. Pryss helped with writing the first draft of the manuscript,
while all authors provided valuable input when proof-reading the manuscript.
D.10 Learnability of a Configurator Empowering End Users to
Create Mobile Data Collection Instruments: Usability
Study
The author of this Ph.D. thesis was responsible for developing the required features
in order to automatically assess performance measures of participants. Further, the
candidate manually assessed all modeled data collection instruments in order to determine
the errors made by respective participants. Dr. Schlee and Prof. Probst advised the Ph. D.
candidate in designing the study and helped with analyzing the collected experiment
data. Dr. Pryss and Prof. Probst helped with writing the first draft of the manuscript,
whereas all authors gave valuable feedback when proof-reading the article.
111
E
Curriculum Vitae
113
E Curriculum Vitae
Curriculum Vitae
Personal Information
Name: Johannes Schobel
Date of Birth: 24th August 1986
Place of Birth: Bregenz (Austria)
Nationality: Austrian
Scientific Education
02/2012 – 09/2018 Ph. D. Thesis
Institute of Databases and Information Systems, Ulm Uni-
versity, Ulm
•
Title: “A Model-Driven Framework for Enabling
Flexible and Robust Mobile Data Collection Applica-
tions”
04/2009 – 01/2012 Studies in Computer Science
Ulm University, Ulm
•Grade: Master of Computer Science
•Degree: M.Sc.
•
Master’s Thesis: “Business Process Intelligence: Ak-
tueller Stand und neue innovative Ansätze zur intel-
ligenten Prozessanalyse”
10/2005 – 03/2009 Studies in Computer Science
Ulm University, Ulm
•Grade: Bachelor of Computer Science
•Degree: B.Sc.
09/2000 – 06/2005 Higher School Graduation
Höhere Technische Bundes- Lehr- und Versuchsanstalt
(HTL), Dornbirn
•Grade: Matura
09/1996 – 07/2000 Higher School Graduation
Gymnasium Gallusstraße, Bregenz
09/1992 – 07/1996 Elemantary School
Volksschule, Höchst
114