Do Desperate Students Trade Their Privacy for a Hope?
An Evidence of the Privacy Settings Influence on the User Performance
Tomáš Obšívač, Hana Bydžovská and Michal Brandejs
Faculty of Informatics, Masaryk University, Brno, Czech Republic
Keywords: Information Privacy, Social Translucence, Educational Information System, Student Performance.
Abstract: Maintaining people's privacy should be the top priority not only in the context of Information Systems (IS)
design. Sometimes, however, certain level of privacy can be traded for a gain in another IS quality or aspect.
We present a real world example of IS with user maintained level of privacy and an evidence of its usage,
correlated with users' performance. Recent students' and applicants' privacy settings in an educational IS
were examined. According to our findings, a part of students voluntarily disclose their presence in the
courses enrolled and on the examination dates registered. Surprisingly, the study results of the disclosed
students are worse then the results of undisclosed ones. In the correspondence with our thesis, disclosed
applicants have better entrance exam results.
1 INTRODUCTION
Privacy can be simply defined as the human right to
be left alone. While it is individually perceived, all
of us feel a need to stay hidden or unnoticed by
others to some extent. With the advent of the
information age, including popular online services
dedicated to support social networking, our concerns
about information privacy rise. At the same time, the
trust to these technologies is conditioned by the
reduced or nonexistent privacy concerns. Even a
long term positive relationship can be lost by just
one security accident involving privacy breach.
Proper definition of privacy, if possible, is harder
to give, because we are dealing with a very elusive
concept. (Solove, 2010) But it is worth of our
attention as a fundamental value that is under attack
from several quarters. (Wacks, 2010).
One characteristic change in social relations
accompanies life in the information society as we
become more and more connected. Maintaining
one's privacy used to be cheap and publicity
expensive. Now the opposite is true, privacy needs
to be defended, all involved parties should care
deeply about it and become aware of dangers
consequent upon an improper use of the information
technology.
End users are concerned with privacy/security
problems more than they are with other types of
computer problems. (Gross, 2007) The taxonomy of
the regulatory and technological approaches to
protect privacy is available. (Chen, 2012).
Careful behavior is indeed recommended for
young people accustomed to the use of new
technologies, e.g. personal mobile devices, as a part
of their needed e-safety awareness. (Atkinson, 2009)
We aim to provide an analysis of a complex real
world example of an information system (IS) with
user controlled privacy settings and the influence of
these settings to the users' performance, as a non
USA-centric evidence missing by the research
community (Bélange, 2011), although it is yet
another one based on students' behavior.
The rest of the paper is structured as follows. The
second chapter deals with the description of privacy
with focus on an educational IS and the description
of particular optional privacy settings usage data.
The third chapter introduce the hypothesis and its
test on the usage data. The chapter four briefly
discuss the results. The last chapter concludes the
paper, including the possible directions of the future
study.
2 INFORMATION SYSTEM USER
PRIVACY
There are a number of important choices regarding
the information privacy settings inside the majority
of information systems. The very basic ones are if
users are allowed to see each other existence,
presence and activity.
156
Obšíva
ˇ
c T., Bydžovská H. and Brandejs M..
Do Desperate Students Trade Their Privacy for a Hope? - An Evidence of the Privacy Settings Influence on the User Performance.
DOI: 10.5220/0004972101560161
In Proceedings of the 16th International Conference on Enterprise Information Systems (ICEIS-2014), pages 156-161
ISBN: 978-989-758-029-1
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
Visibility is an important determinant of harm to
privacy. Meanwhile, one of the essential social
software features, the social translucence, includes
visibility of participants and their activities to one
another. (Erickson, 2000) It is advised for systems
supporting communication and collaboration among
large groups of people.
In the context of IS design, we must admit, that a
certain level of privacy can be traded for a gain in
another IS quality or aspect, being it any desired IS
feature or functionality, e.g. mediation of
interpersonal communication or better user comfort.
Designers are expected to have a good reason for the
tradeoff, however. They should clearly present and
explain it to the user, in the case of both default and
the user maintained privacy settings.
Basic system usage should be possible without
involving users in the privacy setting, for those who
don't care sufficiently, while the highest possible
level of privacy is preserved by default. For
advanced users, who value control over information,
we have good experience with optional settings,
which can lower one's own privacy when desired.
Individuals are willing to trade off privacy
concerns for economic benefits (Hann, 2002) and we
can confirm such behavior with non-economic and
even indirect benefits.
2.1 User Privacy within an Educational
Information System
For the purpose of this study, the goals of an
Educational Information System (EIS) can be
defined as improving the management of education
and providing the digital learning environment.
Even though it is invaluable for the academic
departments, the major EIS use involve students.
Employes usually get proper instructions how to
treat student information privacy and deal with it
accordingly. (Earp 2001) It is inevitable to balance
students' privacy concerns while increasing their
engagement in computer mediated learning at the
same time. (Siemens 2013) Privacy concerned
students have interest in avoiding or selectively
limiting their exposure.
The largest group of users are applicants, in the
case of EIS administering an online admission
procedure.
2.2 Masaryk University Information
System
Being developed since 1999, Web-based
Information System of Masaryk University (IS MU)
hosts numerous applications utilized for managing
study-related records, e-learning tools and those
facilitating communication inside the University. It
is used by more than 30,000 users (of the total of
44,000 students and staff members) a day. It is also
outsourced to another higher education institutions.
As a basic feature, every user has a customizable
personal profile page, by default visible to logged in
users. Students have access to the private list of
enrolled courses. The IS serves as an educational
environment, e.g. stores study materials, collects
homework, includes discussion forums or runs and
evaluates examination tests.
Admissionçprocedures are a part of IS MU.
About 70,000 students apply each year.
Students' and applicants' privacy is considered to
be important of course. Default high privacy settings
can be changed at the user will. We will examine
three of these opt-in settings and their impact on
users' performance in the following chapters.
2.2.1 Opt-in Visibility among Classmates
As IS MU developers, we value student privacy a
lot, indeed, and therefore we do not provide
complete list of classmates. Since we also want to
encourage communication and collaboration among
students, someones presence has to be disclosed in
the specific situations during learning process, such
as the contribution to the discussion forum or a
shared assignment. Eventually, students would
become acquainted anyway in the corresponding
situations during the in-person education form of the
full-time study. A part of students expect to be
visible to classmates and asking our user support
personnel for navigation to the list of students.
On top of that, since Spring 2011, students are
provided with the choice to reveal one's own course
enrollment, but only to classmates which apply for
the same option. Second and slightly weaker form of
disclosure is to reveal of one's registration for a
shared examination date. Opt-in enrollment/
registration disclosure option is provided for every
single course/exam or globally for all courses/exams
ever attended.
2.2.2 Opt-in Visibility among Applicants
In the case of the admission procedure, applicants
are provided with the choice to reveal their county of
residence, but only to applicants which apply for the
same option and for the same field of study as well.
DoDesperateStudentsTradeTheirPrivacyforaHope?-AnEvidenceofthePrivacySettingsInfluenceontheUser
Performance
157
2.3 Optional Privacy Settings Usage
We present data from IS MU from the several last
semesters. To provide a perspective of the social
networking role of IS MU we can tell that the
percentage of users with at least one “friend”, which
is the standard interpersonal relation agreed by both
users, is 23.4%. We register more than 156,000
friendships.
2.3.1 Course Enrollment Disclosure
The trend in opt-in enrollment disclosure option use
is depicted in Figure 1. We selected courses from all
9 University faculties with more than 1 and 10
disclosed students. The more students disclose, the
greater possible cooperation among them is possible.
One disclosed student cannot make any difference
on the results presented later.
Spring 2011
Autumn 2011
Spring 2012
Autumn 2012
Spring 2013
0%
10%
20%
30%
40%
more than 1 more than 10
Figure 1: The percentage of courses per semester with
more than 1 and 10 disclosed students.
The average number of courses per semester
examined was 9,641 for the group of courses with
more than 1 disclosed students. The second group's
average was 6,767. The number is lower because of
exclusion of courses with less then 10 students.
2.3.2 Examination Date Registration
Disclosure
In the same way to the previous chapter, the trend in
the examination date registration disclosure is
presented in Figure 2.
The trend in exam disclosure popularity as
shown in Figure 2. is similar to the course
enrollment disclosure, although percentages are
about half of it.
The average number of courses in a semester is
same as in the previous section. About 500 to 1350
courses per semester have more than one student
disclosed on the same examination date.
Spring 2011
Autumn 2011
Spring 2012
Autumn 2012
Spring 2013
0%
5%
10%
15%
20%
more than 1 more than 10
Figure 2: The percentage of courses per semester with
more than 1 and 10 students disclosed on an examination
date.
2.3.3 Application for Study at the University
Disclosure
The trend in county of residence disclosure by
applicants which have sit for entrance exam is drawn
in Figure 3.
2011 2012 2013
0%
2%
4%
disclosed applicants
Figure 3: The percentage of disclosed applicants per year.
The average number of applications is about
41,750 a year. Opt-in disclose is thus used in 1100 to
1350 cases every year.
3 FINDINGS
We hypothesize that users which lower their privacy
intentionally expect better performance in return,
and actually achieve it.
When we know that sufficient number of our
students are willing to trade part of their privacy for
better contact with their classmates, so let's examine
if there is any correlation with their study results.
The percentage of application disclosures is smaller
but a correlation with the admission results still can
be there.
3.1 Influence of the Privacy Settings
According to our hypothesis, we are looking for the
connection between disclosures and study or
application results.
ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems
158
3.1.1 Influence of the Course Enrollment
Disclose on the Study Results
We have done two simple analyses to compare study
results of group of disclosed students versus others.
The first comparison is by the percentage of
courses passed. The difference between the groups
was below 1% in almost all semesters, which we
consider insignificant.
The second comparison is by the average grade
assigned. The results are presented in Table 1. We
can see the percentage difference (PD) ranging from
4.8% to 8.7% in examined semesters, with the
average PD of 7.3%.
Table 1: Gained Grades Averages.
Disclosed
Students
Undisclosed
Students
PD
[%]
Spring 2011 1.92 1.77 8.1
Autumn 2011 1.91 1.82 4.8
Spring 2012 1.90 1.78 6.5
Autumn 2012 1.96 1.80 8.5
Spring 2013 1.92 1.76 8.7
Students which disclose their course enrollment
have gained slightly worse average grades.
Table 2 shows the average grades again, but
counted for the courses where the disclose take
place, considering only disclosed students. The
percentage difference ranges from 5.1% to 14.1%
here, with the average PD of 9.2%. Notice the
reversed averages at the Autumn 2012 semester.
Table 2: Disclosed Students Grades Averages.
Courses With
Disclose
Courses w/o
D
isclose
PD
[%]
Spring 2011 1.98 1.72 14.1
Autumn 2011 2.16 1.97 9.2
Spring 2012 2.18 1.95 11.1
Autumn 2012 1.92 2.05 6.5
Spring 2013 2.02 1.92 5.1
The disclose takes place within the courses with
worse average grades assigned to disclosed students.
3.1.2 Influence of the Examination Date
Registration Disclose on the Study
Results
We have conducted the same two comparisons as in
the previous section.
The first one ends with insignificant results
again, with no difference in the percentage of
courses passed between the groups.
The second one is presented it Table 3, with
substantial PD between the groups, ranging from
12.4% to 23.2%, with the average PD of 18.6%.
Table 3: Gained Grades Averages.
Disclosed
Students
Undisclosed
Students
PD
[%]
Spring 2011 2.15 1.77 19.4
Autumn 2011 2.06 1.82 12.4
Spring 2012 2.13 1.77 18.5
Autumn 2012 2.19 1.80 19.6
Spring 2013 2.21 1.75 23.2
Students which disclose their exam registration
have gained notably worse average grades.
Table 4 shows the average grades in the courses
where the disclose take place, considering only
disclosed students. The percentage difference is very
high here, ranges from 6.6% to 38.6%, with the
average PD of 27.2%.
Table 4: Disclosed Students Grades Averages.
Courses With
Disclose
Courses w/o
Disclose
PD
[%]
Spring 2011 2.03 1.90 6.6
Autumn 2011 2.75 1.86 38.6
Spring 2012 2.70 1.84 37.9
Autumn 2012 2.37 1.80 27.3
Spring 2013 2.37 1.83 25.7
The discloses takes place at the courses with
higher average grades assigned to disclosed students.
3.1.3 Influence of an Application for Study
Disclose on the Admission Success
Applicant's capacity to study is tested during the
entrance exams. Test is the same for majority of
faculties. Table 5 consists of the disclosed and
undisclosed applicants entrance exam results and
their percentage difference.
Table 5: The entrance exam results (higher is better).
Disclosed
Applicants
Undisclosed
Applicants
PD
[%]
2011 59.00 50.55 15.4
2012 59.04 49.87 16.8
2013 60.58 53.92 11.6
Applicants which disclose their county of
residence have better results of the entrance exam.
The average percentage of overall application
success and the percentage difference is in Table 6.
Applicants which disclose their county of
residence have grater probability of application
success.
DoDesperateStudentsTradeTheirPrivacyforaHope?-AnEvidenceofthePrivacySettingsInfluenceontheUser
Performance
159
Table 6: The application success [%].
Disclosed
Applicants
Undisclosed
Applicants
PD
[%]
2011 47.15 35.83 27.3
2012 48.27 32.85 38.0
2013 47.55 33.85 33.7
4 DISCUSSION
The enrollment/registration/application disclosure
option is the advanced IS MU feature. It is not
visually strongly proposed in the system graphical
user interface or promoted outside the system. The
feature has found its users despite not being widely
known.
So far we have found no evidence in the study
results to support the change from opt-in visibility to
opt-out. We cannot proceed to trade user privacy for
IS functionality by default, unless a clear evidence
of massive positive influence can be proven.
On the other hand, we cannot exclude the
possibility of positive influence of the optional
privacy settings in individual cases. Although the
majority of students are not involved, there are
probably numerous students benefiting from it.
The percentages of courses with both types of
disclosure are different among the University
faculties. The Faculty of Economics and the Faculty
of Informatics have the most courses with disclosed
students. The reason may be the high number of
massively attended courses with final written tests.
On the opposite side, the Faculty of Education and
the Faculty of Medicine have the least courses with
disclosed students. The reason may be the high
proportion of the oral final exams.
Teachers have an educational intent to support
acquittance among students, since some study fields
allow a lot of subject choices and e-learning coupled
with massively attended courses becomes widely
employed. Students with individual curriculum meet
more people but usually only for a brief time and
thus have a lower chance to familiarize each other.
5 CONCLUSIONS
Our hypothesis appears to be wrong in the courses
disclosure and especially in the exams disclosure,
where the results show the opposite phenomenon.
Our best explanation is that students in need of help
look for a classmate assistance. Unfortunately, we
cannot prove it from the presented data. The only
way to answer the question “Do Desperate Students
Trade their Privacy for a Hope?” could be to ask
them. We are currently preparing such survey.
Regarding the application disclosure, the results
support our hypothesis. As soon as the disclosed
applicant is successful in the entrance exam, he or
she may use the feature to the better start of the
study, e.g. to find mates for the commuting to the
University.
5.1 Future Study
We would like to propagate the disclosure options
and to attract students to use this IS MU feature
widely. The disclosure, an investigated parameter,
can be eventually used for the student characteristics
definition and can therefore result in the student
performance prediction accuracy improvement. We
have shown such technique in our previous work.
(Bayer 2012).
The future study should extend this research in
the area of subgroup discovery. Which types of
students use the disclosure? Variables as the gender,
the type of study, the field of study or the form of
study can influence the percentage of the usage. It
will be also promising to explore the courses in
which students are disclosed. Courses can be of a
high capacity, without seminar groups or more
difficult than others.
ACKNOWLEDGEMENTS
We would like to thank all colleagues of IS MU
development team for the support. This work has
been partially supported by Faculty of Informatics,
Masaryk University.
REFERENCES
Atkinson S., Furnell S., & Phippen A., 2009. Securing the
next generation: enhancing e-safety awareness among
young people. In Computer Fraud & Security,
Elsevier, Volume 2009, Issue 7, Pages 13–19. ISSN
1361-3723.
Bayer, J., Bydžovská, H., Géryk, J., Obšívač, T., &
Popelínský, L., 2012. Predicting drop-out from social
behaviour of students. In Proceedings of the 5th
International Conference on Educational Data Mining
(EDM 2012). Chania, Greece. Pages 103–109. ISBN
978-1-74210-276-4.
Bélanger, F., & Crossler, R. E., 2011. Privacy in the
digital age: a review of information privacy research in
information systems. In MIS Quarterly, Society for
ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems
160
Information Management and The Management
Information Systems Research Center Minneapolis,
MN, USA, Volume 35, Issue 4. ISSN 0276-7783.
Earp, J. B., & Payton, F. C., 2001. Data Protection in the
University Setting: Employee Perceptions of Student
Privacy. In Proceedings of the 34th Annual Hawaii
International Conference on System Sciences, Los
Alamitos, CA: IEEE Computer Society Press.
Erickson, T., & Kellogg, W. A., 2000. Social translucence:
an approach to designing systems that support social
processes. In ACM Transactions on Computer-Human
Interaction, Volume 7, Issue 1, Pages 59–83.
Gross, J. B., & Rosson, M. B. (2007, July). End user
concern about security and privacy threats. In
Proceedings of the 3rd symposium on Usable privacy
and security (pp. 167–168). ACM.
Hann, I. H., Hui, K. L., Lee, S. Y. T., & Png, I. P., 2002,
Online Information Privacy: Measuring the Cost-
Benefit Trade-Off. In ICIS (p. 1).
Chen, T. M., & Fu, Z. J. (2012). Protection of Privacy on
the Web. In I. Management Association (Ed.), Cyber
Crime: Concepts, Methodologies, Tools and
Applications (pp. 83–100). Hershey, PA: Information
Science Reference.
Siemens L., Althaus C., & Stange Ch., 2013. Balancing
Students’ Privacy Concerns While Increasing Student
Engagement in E-learning Environments. In
Increasing Student Engagement and Retention in e-
learning Environments: Web 2.0 and Blended
Learning Technologies, (Cutting-edge Technologies in
Higher Education, Volume 6), Emerald Group
Publishing Limited. Pages 339–357.
Solove, D. J., 2010. Understanding Privacy, Harvard
University Press. ISBN 978-0674035072.
Wacks, R., 2010. Privacy: A Very Short Introduction,
Oxford University Press. ISBN 978-0191609626.
DoDesperateStudentsTradeTheirPrivacyforaHope?-AnEvidenceofthePrivacySettingsInfluenceontheUser
Performance
161
