Peer-learning and Talents Exchange in Programming: Experiences and
Challenges
Corinna Kr
¨
ohn and Barbara Sabitzer
School Of Education, Johannes Kepler University Linz, Altenbergerstraße 69, Linz, Austria
Keywords:
STEM, Programming, Peer-learning, Talents Exchange, COOL, Neurodidactics.
Abstract:
“Programming is difficult” society tells us. However, this is only partly due to the subject itself. Very often,
the tasks in software development are related to mathematical problems, which themselves are considered as
difficult because they have no relation to the student’s world of experience. Programming exams contribute
to high drop-out rates and often students choose a different subject or even stop their university education. A
possible gap could be closed by supporting beginners with the help of the COOL (cooperative open learning)
concept that has already found its way into one course in the Business Informatics bachelor program at our
university. In addition to this mandatory class, a peer-teaching course has been establishes. Students meet on
a weekly basis to work together on the problems of software development based on the COOL concept. Two
PhD candidates supervise the students, although they only act as tutors and do not provide any solutions. The
approach described above is now being carried out for the third semester in a row and the drop-out rate has
been reduced. Of course, further observation must be done to be able to draw concrete conclusions.
1 INTRODUCTION
Whether at school or university, learning to code can
be very difficult. High drop-out rates and a major
gender-gap verify these experiences. There already
exist several different approaches to support begin-
ners like tutoring systems or extra training courses but
nonetheless there are still many students that do not
pass exams.
Figure 1: An information processing model (Wolfe, 2010)
adapted by the authors.
One possible way to improve the learning out-
come is to consider how the brain works and ap-
ply teaching methods investigated by Neurodidac-
tics (Sabitzer et al., 2020). Neurodidactics combines
Neuroscience and Didactics and presents a relatively
young science that is yet not very established. Of
course Neurodidactics cannot be transferred to higher
education lessons one to one but it can give propos-
als for effective learning and teaching and can pro-
vide information of how the brain works (Sabitzer and
Pasterk, 2015).
One of the key findings from Neurodidactics is
that knowledge cannot be transferred but has to be
newly created in each student’s brain. That proposes
that the learning process is an active one which also
agrees with findings in pedagogy and constructivism.
Our memory is a highly complex system (see figure
1) that also includes unconscious processes. The hu-
man long-term memory seems to be static but in fact
is dynamic which concludes that each recall causes a
re-encoding (Sabitzer, 2011).
Teaching methods suggested for effective learning
and stimulating the long-term memory are the follow-
ing (Sabitzer et al., 2013):
Discovery Learning: rests on finding solutions on
your own and not getting instructions from teach-
ers.
Social or Observational Learning: is based on so-
called mirror neurons that are active when stu-
dents observe or imitate others.
466
Kröhn, C. and Sabitzer, B.
Peer-learning and Talents Exchange in Programming: Experiences and Challenges.
DOI: 10.5220/0009472004660471
In Proceedings of the 12th International Conference on Computer Supported Education (CSEDU 2020) - Volume 1, pages 466-471
ISBN: 978-989-758-417-6
Copyright
c
2020 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
Learning by Teaching: Students slip into the role of a
teacher and explain things to others. This method
restarts the memory process and enhances reten-
tion.
Learning by Doing: inspires the student to be more
active and is more effective than teacher-centered
training.
COOL Cooperative Open Learning: was initiated
by a team of Austrian teachers and is described
in one of the following sections.
This paper describes the evolution of a tutoring
group to a peer-teaching class for programming be-
ginners based on COOL concepts at our university.
Sections 2 and 3 describe the underlying theoretical
ideas, whereas section 4 concentrates on the detailed
explanation of the development process of the differ-
ent approaches to help students. It also presents expe-
riences of students and teachers and concludes with
ideas to further integrate the tested concepts in other
university classes.
2 PEER-LEARNING AND
ASSESSMENT
2.1 Definitions
Peer-learning is the use of teaching and learning
strategies in which students learn with and from each
other without any direct intervention of a teacher. Ex-
amples of peer-learning are student-led workshops,
team projects, study groups or student-to-student
learning partnerships. In reciprocal peer-learning
students act as both teachers and learners (Boud et al.,
1999).
In contrast to peer-learning peer-teaching also in-
volves advanced students in class who act as tutors.
The term “collaborative learning” is used to refer to
peer-teaching as well (Boud et al., 1999).
In Pair programming two programmers work to-
gether at one computer on the same algorithm. One
person is the “driver” whereas the other is the “nav-
igator”. The driver types at the computer and per-
forms the physical input while the navigator observes
and comments the role of the driver. Literature rec-
ommends to alternate roles during the software devel-
opment process. Research has shown that this con-
cept improves design quality, produces higher qual-
ity code, and improves communication among team
members. But it has to be mentioned that pair pro-
gramming only works if people are matched with a
partner of comparable ability (Han et al., 2010).
2.2 Why Focusing on Peer-learning?
There are pragmatic as well as principled reasons for
concentrating on peer-learning at university. One ma-
jor problem of universities is lack of staff. As peer-
learning and peer-teaching both concentrate on stu-
dents, universities can easily increase the number of
students in classes without more input from staff. Of
course, it is also being argued that collective forms of
learning may better suit women and minority groups
than the traditional individualistic teaching methods
because they concentrate on cooperation rather than
competition (Boud et al., 1999). Considering the
gender-gap, more and more of these teaching and
learning concepts should be taken into examination.
Another important advantage is to foster students’
practice in planning and teamwork because in peer-
learning it is necessary to work together. It also en-
gages reflection as well as exploration of ideas when
the teacher is not immediately present. Moreover,
students achieve more practice in the communication
in the subject area compared to traditional teaching
methods (Boud et al., 1999).
2.3 Assessment
When considering peer-learning, assessment needs to
be taken into account. Assessment works as a form
of academic currency and provides compensation for
efforts of students. Of course it is also important not
to misuse assessment as an instrument to get students
to join peer-learning activities (Boud et al., 1999).
Assessing peer-learning has to be done with cau-
tion. The teachers will not be able to assess the out-
come of peer-learning groups directly as it restricts
the benefits of peer-learning itself. But there is a tradi-
tion of individual and competitive assessment in most
universities which implies competition against others,
rather than cooperation and collaboration is seen as
cheating (Kohn, 1992).
Boud et al. recommend the following forms of
assessment when peer-learning or peer-tutoring is in-
volved (Boud et al., 1999):
Group Assessment: If teamwork and collaborative
working is valued then each member of the
group should receive the same grade. To prevent
freeloading, the group assessment could also be
the sum of each member’s assessment.
Peer Feedback and Self-assessment: Switching out-
comes of given tasks and grading classmates (or
oneself) is a common approach
Assessment of Process: shifting the main focus from
the learning outcomes towards the learning pro-
cess
Peer-learning and Talents Exchange in Programming: Experiences and Challenges
467
Negotiated Assessment: all parties involved agree on
the assessment process
Use of Cumulative Rather than Weighted Assessment:
If the weighting of one element of the course is
very low, students tend to ignore it or put little
effort in it. Peer-learning is often one of those
aspects with little weight. But if it is only possible
to pass the course by completing all of the tasks,
motivation could rise to attain positive results in
the smaller tasks as well.
3 COOL CONCEPTS: COOL
INFORMATICS AND COOL
PROGRAMMING
COOL is an acronym for Cooperative Open Learn-
ing, Computer-(science)-supported Open Learning,
as well as cool with the meaning of interesting or
“in” (Sabitzer, 2014). The COOL concept was de-
veloped by Austrian teachers in 1996 and then ex-
tended to COOL Informatics which was firstly pre-
sented in the habilitation thesis of one of the authors
(Sabitzer, 2014). The four main principles are Dis-
covery, Cooperation, Individuality, and Activity (see
figure 2). Each of those concepts is then related to
various teaching and learning methods that rest upon
Neurodidactics (Sabitzer and Pasterk, 2015).
COOL Programming concentrates on these de-
scribed neurodidactical concepts and was installed in
one of the beginners programming classes in our uni-
versity’s Business Informatics bachelor program. The
lessons as well as tasks were designed based upon the
following findings:
1. New content is always built on existing knowl-
edge and learning occurs through association
(Sousa, 2009) (Spitzer, 2006).
2. Knowledge cannot be transferred but must be con-
structed in each student’s brain (Roth, 2004).
3. Learning occurs through imitation (Sousa, 2009).
4. The brain recognizes and produces patterns, cate-
gories, and rules itself (Spitzer, 2006).
5. The instruction method impacts the retention of
new information (Sousa, 2009).
6. Double-coded is double-saved (multimedia ef-
fect) (Roth, 2004).
Each of the 90-minutes-lessons in the designed course
was divided into three parts: Questioning (10-15 min-
utes), Discovering (10-15 minutes), and Pair Pro-
gramming (60-70 minutes). More precisely, the stu-
dents started with an open round of questions that
Figure 2: Brain-based teaching concepts (Sabitzer, 2014)
adapted by the authors.
they could ask their professor. The next phase was
dedicated to extract structures, rules, and other essen-
tial elements from Java Code by using reading cor-
ners, sample solutions, puzzles (see figure 3), or step-
by-step examples (see figure 4). Afterwards students
worked together on various programming tasks, fol-
lowing the rules of pair-programming (Sabitzer et al.,
2020).
Figure 3: Datatype puzzle (Sabitzer et al., 2019).
CSEDU 2020 - 12th International Conference on Computer Supported Education
468
Figure 4: Excerpt from a step-by-step programming exam-
ple (Sabitzer et al., 2019).
4 DEVELOPMENT OF A
PEER-LEARNING/
PEER-TEACHING GROUP
Even though one of the beginners programming
classes switched to the concepts of COOL Program-
ming, there were many other courses that required
support for students who had a hard time starting to
code. More and more students created learning and
tutoring groups on their own but they lacked loca-
tions, a fixed time, and experienced tutors. There-
fore, students came to our department and asked if
we could help them. Consequently, in winter term of
2018 we created a peer-learning group that met each
week in the conference room of our department to dis-
cuss the upcoming programming tasks. One student
who already worked as a software developer in differ-
ent companies supported them as tutor. They met on
a regular basis but the number of students varied de-
pending on how hard the task was or if the exam was
close. Accordingly, some of the students had to alter
programming partners and sometimes there was only
one student and the tutor present.
In summer term 2019 the first official peer-
learning class was installed. Students could take that
class voluntarily but still get ECTS for it. In that
semester there were only 8 students who enrolled for
class. Students were expected to actively participate
and attendance was mandatory. The group split up
on their own in various smaller groups most of the
time pairs to work together on the given program-
ming tasks. Two PhD-students acted as tutors but did
not solve the coding tasks step-by-step. The problem
with this class was that the students attended three dif-
ferent mandatory programming courses at university.
One was the beginners class from the Computer Sci-
ence bachelor program, the second one was the class
of Business Informatics described in section 3 that
is based on COOL Informatics, and the third group
consisted of students from an advanced software de-
velopment class. The problem occurring was obvi-
ous: whether the students nor the teachers could an-
swer all the upcoming questions with such a variety
of given tasks. But still the survey at the end of the
course showed that all students approved and wanted
the class to be continued. Fortunately, each one of
them passed the final programming exam.
The following winter term the class was restricted
to students who were enrolled in either the begin-
ners programming class of the Computer Science or
the Business Informatics bachelor program. The sys-
tem also changed from peer-learning to peer-teaching,
as two advanced students were involved in the same
group. At the beginning 23 students attended the
class, whereas only 14 completed it due to lack of
time or interest. The group was split into three smaller
study groups: one group consisted only of Computer
Science students, the other two of Business Informat-
ics students. Each of the two Business Informatics
groups was assisted by one of the advanced students.
Due to the lack of one more experienced student the
group of students from Computer Science had no des-
ignated external tutor. Of course the two PhD candi-
dates who oversaw the class provided help to each of
the groups but still the main focus was on the peers.
As already described in section 2.3, assessment of
such peer-learning or peer-teaching classes is hard.
As all students of the course participated actively in
class, the teachers decided on assessing them accord-
ing to their attendance. The end results of all the final
exams are not published yet and therefore questioning
students about this class has to await the outcome but
we already know that at least half of the group passed.
5 EXPERIENCES OF STUDENTS
AND TEACHERS
The main motivation for changing the ongoing system
of programming education at our university was the
close contact to students. At least once a semester the
professor (and one of the authors) officially opens her
office for first semesters and talks about their prob-
lems and wishes. Students as well as teachers ex-
pressed their experiences towards the implemented
courses:
Peer-learning and Talents Exchange in Programming: Experiences and Challenges
469
5.1 Report of a Female Student
“I started the studies of Computer Science with no
prior programming experiences. As the introductory
software-engineering course started, I soon found out
that I was (1) one of the few students with no prior
programming experiences and (2) one of the very few
girls, which was quite intimidating. Because of the
high demands and rapid pace, students dropped out
week per week. In my opinion, these are the main
reasons for the high drop-out rate:
Collaboration: Students had to complete weekly as-
signments independently. Collaboration was not
allowed and in case of similar codes, the home-
work of all the students involved did not count.
However, especially beginners benefit from col-
laboration and improve their skills by learning
from and talking to each other. It could be ar-
gued, that students can indeed collaborate, they
only have to develop different algorithms. Nev-
ertheless, for students who are also not familiar
with algorithmic thinking, this can be a huge chal-
lenge. And in reality, aren’t software project usu-
ally planned and realized in teams?
Pace and Demands: As a programming beginner, I
needed to dedicate the major part of my time to
this course to be able to keep up the speed. Be-
sides my job and other courses, this meant very
long nights and long weekends. Furthermore, stu-
dents had to positively complete 80% of the as-
signments and at the same time, reach a certain
amount of points. These demands are very hard
to achieve and put a high pressure on the students.
Fortunately, this was the first semester, where a
weekly peer-learning group was offered for CS
students of the teacher training program. In these
weekly meetings, we worked on our own algo-
rithms, but in a very supportive environment. In
other words, in case we were stuck, there was a
tutor who led us to the right direction and gave us
very helpful advice. Luckily, in my case it was
worth the effort and I passed the course.
However, success very much depends on external sup-
port. Without professional support, like we had in the
peer-learning group, I would not have been able to
complete the course, which probably would also have
been the end of my computer science career.
5.2 Report of a Teacher of the
Peer-learning Class
As a teacher of the new peer-tutoring course, I
learned software engineering the classical way and
was challenged with difficult tasks, an extremely high
drop-out rate and hours of searching through the In-
ternet for answers. Teaching this course was a com-
plete new role for me. As a teacher I usually give
answers to questions. In this course I had to help the
students to find the answers on their own. This lim-
ited way to support the students worked better than
expected. It showed that a lot of difficulties were
the results of theoretical problems like “how does a
list work”. Furthermore, students started to help each
other, since they worked on the same task, the prob-
lems were quite similar. The course was offered to
students of two different Computer Science classes,
the basic one and the advanced. During the course I
realised that it is very challenging to stay up to date
with all tasks and to switch during the lesson in an-
swering questions to the basic and the advanced tasks.
Altogether, the new way in teaching showed me, that
students can help their fellow students in a very ef-
fective way and they have a lot of patience in helping
each other.
6 CONCLUSION AND OUTLOOK
In this paper we have presented the journey from a
tutoring group to a peer-teaching course in program-
ming. We had high drop-out rates in our program-
ming classes for beginners at our university. More and
more students created learning groups on their own
but they lacked experienced tutors. At first attempt,
we installed a peer tutoring group that met weekly. In
summer term 2019 the first peer-learning class based
on COOL Informatics and COOL Programming was
created that could be visited voluntarily. After two
semesters, we can already identify improvement in
terms of drop-out rates, exam attendance, and home-
work submissions.
We have identified the need for more collabora-
tion than just working alone on tasks therefore com-
petition has to be decreased. One remaining question
is the assessment of the programming tasks. If stu-
dents work in pairs or in groups on one project they
will hand in similar or even equal solutions. How can
a tutor distinguish between group work and cheating?
In the upcoming months, we will concentrate on
evaluating the survey and the exam grades to further
investigate the needs of programming beginners. We
also need to collect more data about the background
of our students, as their personal factors influenced
their learning outcomes. We do not have enough in-
formation to answer questions like:
1. How do prior programming knowledge and drop-
out likelihood correlate?
CSEDU 2020 - 12th International Conference on Computer Supported Education
470
2. How does the type of high-school-diploma and
drop-out rate correlate?
3. Does peer-learning or peer-teaching reduce the
gender-gap?
Summarizing, we want to expand our concepts to ad-
vanced programming courses and provide more than
one peer-teaching class.
REFERENCES
Boud, D., Cohen, R., and Sampson, J. (1999). Peer learning
and assessment. Assessment and Evaluation in Higher
Education, 24(4):413–426.
Han, K. W., Lee, E. K., and Lee, Y. J. (2010). The Impact
of a Peer-Learning Agent Based on Pair Programming
in a Programming Course. IEEE Transactions on Ed-
ucation, 53(2):318–327.
Kohn, A. (1992). No Contest: The Case Against Competi-
tion. Houghton Mifflin, Boston.
Roth, G. (2004). Warum sind lehren und lernen so
schwierig? Zeitschrift f
¨
ur P
¨
adagogik, 50.
Sabitzer, B. (2011). Neurodidactics – a new stimulus in ict.
Sabitzer, B. (2014). A Neurodidactical Approach to Coop-
erative and Cross-Curricular Open Learning: COOL
Informatics. Habilitation, Alpen-Adria-University
Klagenfurt.
Sabitzer, B., Groher, I., Sametinger, J., and Demarle-
Meusel, H. (2020). Cool programming improving
introductory programming education through cooper-
ative open learning. ICEIT 2020, Oxford, UK.
Sabitzer, B. and Pasterk, S. (2015). Brain-Based Program-
ming Continued: Effective Teaching in Programming
Courses. In Proceedings Frontiers in Education
Conference, FIE, volume 2015-February. Institute of
Electrical and Electronics Engineers Inc.
Sabitzer, B., Pasterk, S., and Elsenbaumer, S. (2013). Brain-
based teaching in computer science: Neurodidactical
proposals for effective teaching. In Proceedings of the
13th Koli Calling International Conference on Com-
puting Education Research, Koli Calling ’13, page
197–198, New York, NY, USA. Association for Com-
puting Machinery.
Sabitzer, B., Spieß-Knafl, S., Pasterk, S., and Kr
¨
ohn, C.
(2019). Entdecken Sie Java! Programmieren lernen
und
¨
uben mit Musterl
¨
osungen. Linz, Austria.
Sousa, D. A. (2009). How The Gifted Brain Learns. Corwin
Press.
Spitzer, M. (2006). Lernen Gehirnforschung und die
Schule des Lebens. Springer Spektrum.
Wolfe, P. (2010). Brain Matters: Translating Research into
Classroom Practice. Assn for supervision & curricu.
Peer-learning and Talents Exchange in Programming: Experiences and Challenges
471