Content uploaded by Dymitr Ruta
Author content
All content in this area was uploaded by Dymitr Ruta on Oct 25, 2018
Content may be subject to copyright.
Does Gender Matter for Collaborative Learning?
Ling Cen, Dymitr Ruta, Leigh Powell, and Jason Ng
Etisalat British Telecom Innovation Centre,
Khalifa University of Science, Technology and Research, Abu Dhabi, UAE
Email: {cen.ling, dymitr.ruta, leigh.powell, jason.ng}@kustar.ac.ae
Abstract—In our recent work we have proven quantitatively
that collaborative learning improves students’ knowledge
retention and boosts the quality of attained learning outcomes. In
this research we investigate the role that the students’ gender plays
in their engagement during collaborative learning and their
learning performance as assessed by the teacher. In the context of
the EBTIC developed Collaborative Learning Environment
deployed at Khalifa University along the sequence of three group
courseworks over one semesters, we intend to explore the
differences between the collaborative learning style and quality in
female, male and mixed-gender groups. The series of detailed
cross-gender learning engagement and performance comparisons
indicate that female groups tend to work simultaneously and
achieve better results while male group members engage less and
work in sequence. As a result female groups exploit the added
benefits of collaborative learning more than the male groups.
What is striking, however, the members of the mixed-gender
groups excel the most, significantly improving their engagement,
focus and the quality of groupwork comparing to same-gender
groups. We believe this outcome delivers yet another proof of the
synergies and efficiencies of interactive learning in a diverse group
of students and encourages mixing genders when composing
groups for collaborative learning.
Keywords—Collaborative learning, learning performance,
student genders, big education
I. I
NTRODUCTION
Cooperative learning is defined by Johnson, et al. as the
instructional use of small groups so that students work together
to maximize their own and each other’s learning [1]. It refers to
situations and environments in which learners engage in
common tasks and each individual capitalizes on resources and
skills from each other [2]–[4]. It is based on the model that
knowledge can be created within a population where members
actively interact by sharing experiences and take on asymmetric
roles [5], [6]. With the development of cheap and powerful
knowledge access technologies connecting and enabling
students to carry out ever more learning, coursework and
assessment tasks together, collaborative learning has attracted
increasing interests in recent years [2], [7]–[10].
Although there is a wide body of qualitative evidence
reporting the benefits of collaborative learning, the thorough
quantitative analysis is clearly lagging behind. This is perhaps
due to the difficulties with formal knowledge representation and
the usual lack of data capturing the complete process of
collaborative learning in a sufficient detail. To address this, our
recent work [11] has numerically proven that collaborative
learning improves students’ knowledge retention and boosts the
quality of attained learning outcomes. In CLE, learning outcome
is reflected not only by the contribution of individual students to
the groupwork but also by the collaboration patterns within
groups, thereby bringing much more complexity and details to
the analysis of collaborative learning in comparison to
individual learning. Students’ gender brings yet another
interesting dimension to this analysis and provides vital
information about the ways students engage, collaborate and
learn along different gender lines [12], [13]. The study carried
out in [14] indicates that females tend to focus more on social
oriented activities, while males clearly focus more on task-
oriented activities. Moreover female students learning together
in the technology-rich environment seem to participate more
actively and persistently regardless of the nature of the task [15].
Similarly it is found that female engineering students choose
collaboration as a successful learning strategy more often
comparing to their male classmates [16].
Significant research has been dedicated to study and
compare the effectiveness of single-gender education and co-
education with various gender compositions of classrooms. The
focus of these studies varied from trying to employ gender-
specific educational strategies to enhance students’ confidence
and skills to trying to improve learning outcomes and achieve
social mobility [14]. In [14], it is found that the overall course
performance for both genders was improved by changing the
software engineering classroom composition from a gender
heterogeneous to a gender homogeneous classroom. However,
the ratios of gender composition in the 4 academic years under
comparison were different, while the initial mixed gender case
was seriously under-represented. For example the records of
only 11 males and 5 females in mixed-gender setup over one
semester were compared against the records of over 100 students
reported over 3 years in gender - separated education. Another
study in [17] found that there is a slight gap between learning
performance of male and female students in secondary education
levels in the United Arab Emirates (UAE). It is therefore not
entirely clear whether the improved academic performance
reported in [14] was achieved through gender segregation or
simply by increased participation of better performing female
students. In turn research outcomes reported in [18] claim that
high school girls performed better in single gender groups when
learning unfamiliar tasks but excelled more in mixed gender
groups when learning familiar tasks. The authors also suggest
that such strategy can improve the development of personal
authority and self-confidence among girls in science and math.
Despite all these valuable findings and constructive
discussions on the role of genders in education reported in the
literature, there is still a lack of clarity on the role that gender
plays in education particularly in the collaborative learning
context. Part of the reason could be that in various investigations
the analysis of the role of gender in education is intertwined with
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 433
the students’ attitude towards learning and the motivation to
accomplish the tasks both of which have been identified as the
measures of gender gap [14]. Following the remarks in [19] we
intend to approach the gender analysis by trying to understand
how female and male students learn in the controlled
collaborative learning environment as opposed to pointing what
they can learn and how effective they are in different areas of
curriculum. Specifically we want to explore the detailed styles
and patterns of collaborative learning in different gender
configurations of group learning participants and monitor their
engagement as well as the learning outcomes. We believe it is a
more objective measure of the possible learning synergies, it is
more independent to the choice of the learning content and better
informs additional efficiencies or deficiencies of learning in a
group along different gender lines. The outcomes of such
analysis might be even more interesting due to the fact that it is
carried out in the Middle East region where the cultural
characteristics and gender sensitivities are likely to result with
more pronounced outcomes than in the other parts of the world.
The slight gap between learning outcomes of male and female
students in the UAE reported in the Policy Brief [17], poses
another challenge, yet it also creates great opportunities for
enhancing economic competitiveness of the UAE if the methods
of bridging this gap are identified as a result of this work.
To address these challenges we conduct our work in the
context of the EBTIC developed Collaborative Learning
Environment deployed at Khalifa University along the sequence
of three courseworks over one semester. We focus our
investigations on the role that the students’ gender plays in their
engagement during learning and the learning performance as
assessed by the teacher. First, we intend to explore the
differences between the collaborative learning style and
outcomes in female, male and mixed-gender groups. This allows
us to compare the engagement and learning performance of
same- and mixed-gender groups, verify the gender gap reported
in [17] and try to understand its possible sources from the
learning engagement patterns. Second, we intend to look at the
journeys of several students who migrated from same- to mixed-
groups and vice-versa and observe the changes in engagement
and performance throughout these migrations. Finally we then
try to isolate the synergies (measured by improvements in
learning outcomes) attained through collaborative learning and
assess how well these synergies are exploited by different gender
compositions of the group of students.
The remainder of the paper is organised as follows. Section
II introduces the CLE, discusses its features and briefly
describes the trial setup. The learning performance of groups
with different gender types is analyzed in Section III. Section IV
investigates the similarity and difference of learning patterns
between genders in CLE. The impact of collaborative learning
on mixing genders is analyzed in Section V, followed with the
conclusions summarizing the contributions of this paper and a
further research scoped in this area.
II. C
OLLABORATIVE
L
EARNING
E
NVIRONMENT
T
RIAL
The Collaborative Learning Environment (CLE) is a system
developed at EBTIC [9], [10] that brings together a collection of
tools and functionalities enabling communication, information
sharing and collaborative document creation within the same
environment. As opposed to individual communication and
sharing tools like Skype, Facebook, or Google Drive which
focus on a specific interaction or activity, CLE is designed to
integrate these different functionalities into one, cohesive
environment for formal education purpose.
CLE is implemented as a set of modules for Moodle, an open
source learning management system (LMS), and as such is able
to capitalise on existing Moodle functionalities like group
creation, file sharing and forums. By leveraging the flexibility of
open-source technologies, CLE integrates seamlessly into the
LMS, providing a workspace that is already familiar to both
students and faculty, thereby reducing cognitive load and
enabling more focus to be placed on collaborative learning.
The aim of CLE is to stimulate the collaborative learning
process and enable instructors to facilitate collaborative
assignments more easily. The whole interaction history is
logged, which provides data enabling a dynamic analysis of
contributions, usage and participation as well as to allow for
more advanced future functions such as knowledge elicitation.
A screenshot of the CLE is shown in Fig. 1.
Fig. 1. Collaborative Learning Environment in action.
Communication features of the CLE include synchronous
text chat and audio/video communication, which allow
participants to exchange ideas and communicate directly with
each other regardless of their geographic location. Additionally,
a collaboration area is provided to allow students to either
synchronously or a-synchronously create an assignment. This
area, called the collaborative editing pad, provides a canvas on
which each student can contribute and revise their ideas. Each
contributor to the pad is assigned a unique colour, so individual
contributions are evident, and each keystroke, whether it is an
”add”, ”edit” or ”delete” is recorded by the pad. Using this data,
the CLE statistics module, screenshot of which is shown in Fig.
2 can output detailed usage statistics to the instructor upon
request, allowing for an in-depth analysis of how an individual
assignment was built and reveal how the group collaborated
together. Beyond statistics, a playback feature is provided,
allowing students and instructors to watch the entire creation of
the assignment, from start to finish, much like watching a video.
Both the statistics and playback features of CLE were used
by instructors of the Freshman Design Engineering Course at
Khalifa University to assist in analysing student group
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 434
adherence to a prescribed engineering design cycle throughout
the Autumn 2012 semester. Overall 122 students participated in
the trial, who were partitioned into groups. The group sizes
varied typically between 3 to 6 students, however there were
instances when only 1 or 2 students contributed. The interactions
of students within groups over a sequence of three courseworks
were recorded at up to 1s time resolution. Each group’s
coursework followed with the teacher assessment and a single
grade allocated to the whole group. All grades considered in this
paper were normalised to the range of [0; 100] to allow
comparisons across different courseworks.
Fig. 2 CLE statistics, sample group contributions graphs.
Both the statistics and playback features of CLE were used
by instructors of the Freshman Design Engineering Course at
Khalifa University to assist in analysing student group
adherence to a prescribed engineering design cycle throughout
the Autumn 2012 semester. Overall 122 students participated in
the trial, who were partitioned into groups. The group sizes
varied typically between 3 to 6 students, however there were
instances when only 1 or 2 students contributed. The interactions
of students within groups over a sequence of three courseworks
were recorded at up to 1s time resolution. Each group’s
coursework followed with the teacher assessment and a single
grade allocated to the whole group. All grades considered in this
paper were normalised to the range of [0; 100] to allow
comparisons across different courseworks.
III. L
EARNING
P
ERFORMANCE OF
D
IFFERENT
G
ENDERS
Fig. 3. Distribution of male, female and mixed groups.
In this trial, the 3 courseworks are attended by in total 122
students partitioned into 26, 26, and 20 groups, respectively.
Most of students attended all of the 3 courseworks. According
to the gender types of students, we separate the groups into
uniform-gender groups including male and female groups, and
mixed groups. The gender distribution of the groups is shown in
Fig. 3. Among 72 group instances, 37 are male-only groups,
which are much larger than the number of female-only and
mixed groups with sizes of 22 and 13, respectively.
It has been found that in schools of United Arab Emirates
(UAE) there is a slight gap in attained performance between
male and female students in several key areas [16]. Consistent
with this finding positioned in individual learning, it is also
found in our case study that the female-only groups perform
better than male-only groups in collaborative learning. Fig. 4
shows the grades achieved in the 3 courseworks by male, female,
and mixed groups, respectively. The 2 lowest grades below 60%
considered as the borderline were achieved by two singleton
male groups.
Fig. 4. Grade distributions among male, female and mixed groups.
The statistics of group grades are presented in Fig. 5, where
the mean, median, maximum, minimum and standard deviation
are given separately for the 3 types of groups. The figure shows
a consistent performance edge of the female groups over male
groups across all the statistics, with also much tighter
distribution reflected in smaller standard deviation of grades for
the female groups.
Fig. 5. Grade statistics of male, female and mixture groups.
IV. G
ENDERS IN
C
OLLABORATIVE
L
EARNING
P
ATTERNS
We formally quantify and analyse the collaboration among
students in a group using a collection of time series reflecting
individual student’s contributions to the coursework. All
individual students’ coursework progress time series have been
merged into a single colour-coded progress timeline. A sample
of such timeline of the collaboration progress is shown in Fig. 6.
The heading of each figure is organized in a standard way
starting from the name of the group, followed with the
coursework ID and the assessment grade allocated by the
teacher. The vertical axis measures the cumulative volume of the
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 435
coursework content change recorded by the CLE, and the
horizontal axis represents the time stamps of these changes.
Fig. 6. Sample patterns of student collaboration time series.
Initial analysis revealed that students’ collaborative work on
the same coursework followed many different patterns. The
contributions of the individual members in some of the groups
are clearly organized in sequences, while others work
simultaneously. The diversity in contribution is also reflected by
the frequency of individual members’ contribution. Some of the
patterns show only single contribution of one student, while in
other cases students contribute interchangeably multiple times.
Continuous focus, self-reflection and live collaboration is
naturally more likely to lead to a better and more refined and
coherent courseworks, that are likely to achieve better marks.
There is also a significant variability in workload distribution
among group members. In some cases the workload splits fairly
evenly within the groups, while it is dominated by only one
member in other cases. The timeline progress patterns are also
excellent at discovering content cut and pasted into the CLE
from outside. Frequent and sizeable vertical jump patterns
signify such activity and it is in direct contrast to some other
patterns of smooth gradual accumulation of the creative
activities captured by the CLE. The size of a group is also an
important factor when looking at the data gathered by the CLE.
It indicates the amount of knowledge exchange and
collaboration available during the learning and content
generation process. The group sizes varied typically between 3
to 6 students with infrequent instances of only 1 or 2 student.
In this section, the collaborative learning patterns captured
by the CLE are analyzed along the lines of different genders.
Specifically, in Section IV-A, we explore the impact of the
collaborative patterns on the learning performance for groups
with different gender types. In Section IV-B, we attempt to
explore the benefit of mixing genders in collaborative learning.
A. Impact of collaboration patterns on learning performance
It was inline with our expectations to observe that the worst
grades were achieved by the groups with only 1 male contributor
as shown in the example in Fig. 7. Both male and female groups
with larger sizes, have more knowledge and resources and
creative diversity at their disposal which makes them more likely
to achieve better performance.
Fig. 7. The groups with 1 male student achieved lowest grades.
Both male and female groups with larger sizes, have more
knowledge and resources and creative diversity at their disposal
which make them more likely to achieve better performance.
Fig. 8 compares 4 groups which changed from multi-member
(upper figures) to single-member (lower figures) and all
experienced a large drop in the quality of learning outcomes.
Fig. 8. Performance drops in groups changed from multi- to single-member.
(a) Balanced contributions within male groups.
(b) Imbalanced contributions within male groups.
Fig. 9. Contributions balance within male groups.
Another common phenomena found in the learning patterns
of both male and female groups is that even distribution of
workload and balanced contribution could improve the learning
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 436
outcome for both genders. Figs. 9 and 10 illustrate several
sample patterns for male and female groups, respectively. In
Figs. 9(a) and 10(a), the progress timelines have relatively
balanced contributions within groups and received full marks,
while the timeline patterns in Fig. 9(b) and 10(b) show clear
patterns of contribution imbalance correlated with lower grade.
(a) Balanced contributions within female groups.
(b) Imbalanced contributions within female groups.
Fig. 10. Contributions balance within female groups.
However, it is interesting to find that the male groups more
often organized individual member contributions in sequences,
while female and mixed groups preferred to work
simultaneously. Fig. 11 shows the patterns of progress timelines
reflecting the two distinct collaboration modes, where in Fig.
11(a) there was only one student working at a time along the
duration of the coursework, while in Fig. 11(b) the group
members contributed simultaneously and collaboratively to the
groupwork. We can see from the figure that organizing the
contribution of individual members simultaneously is more
likely to have better learning performance than in sequences.
This can be partly explained by the fact that more concurrent
work means more cooperation, self-reflection, and much better
refinement of the jointly created content that earns better
performance grades.
(a) Contribution organized in sequence within male groups.
(b) Simultaneously organized contribution within female
groups.
Fig. 11. Contribution organization within groups.
Fig. 12. Simultaneously organized contributions appears better than sequential.
Even in a few cases where female groups organized their
contributions in sequence, it is found that learning outcome
could be worse than simultaneous contribution. Fig. 12
illustrates the learning time series from one female group in 2
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 437
courseworks, respectively. It shows that the group organized its
major contributions either in sequence or simultaneously in
different courseworks. Better learning outcome was achieved
with the simultaneous groupwork contribution pattern for the
group.
B. Learning synergy in mixed gender groups
As shown in Fig. 5, the mixed groups perform almost as well
as female groups. To further explore extra synergy benefit of
mixing genders, the performance between uniform and mixed
groups is compared, which are shown in Fig. 13. In this figure,
the performance statistics are presented separately for the
uniform and mixed groups. As it appears in the figure, the grades
of mixture groups have larger median, mean, and minimum
values with a smaller deviation, which indicates that mixing
genders in CLE could improve students’ learning performance.
Fig. 13. Comparison between uniform and mixed groups.
Another interesting finding is that the mixed groups have
more even distribution of workload than in the uniform groups.
Fig. 14 illustrates some learning patterns of progress timelines
for male groups (Fig. 14(a)), female groups (Fig. 14(b)), and
mixed groups (Fig. 14(c)). As shown in Fig. 14(a) and Fig.
14(b), the patterns were largely dominated by one student
member, while the series shown in Fig. 14(c) shows groups with
even workload distribution typical for mixed gender groups.
Even workload distribution indicates more participation and
cooperation which consequently leads better learning outcomes.
This is reflected in Fig. 14 where the lower grades of 70 and 80
in male and female groups, respectively, are observed for groups
dominated by one member, while groups with even workload
distribution observed in mixed groups achieve higher scores.
(a) One student dominates contributions in male groups.
(b) One student dominates contributions in female groups.
(c) Workload spilts evenly among members of the mixed gender groups.
Fig. 14 Workload distribution samples in male, female and mixed groups.
It is useful to understand the contributions of male and
female parts in mixed groups, and explore who is driving the
performance by inspecting closer the learning patterns of the
mixed groups. Fig. 15 displays the progress time series for
mixed groups in different courseworks. It is obvious that female
students were more active in mixed groups. It can be seen from
these patterns that most of coursework content changes are made
by the female members, and their login time to CLE platform is
also much more than their male teammates. It indicates that the
performance of the mixed groups is mainly driven by their
female members.
Fig. 15. Female stud ents drove the performance in mixed groups.
We also found in our case study that mixing different
genders in one learning group could possibly arouse learning
enthusiasm of students who are willing to contribute more in
learning process. Fig. 16 illustrates 2 female students, named as
ME and SA, migrated from a mixed group to uniform groups in
different courseworks. The upper figure features a pattern of the
mixed group which consists of the two female students
mentioned and 1 additional male student. Each of the lower 2
figures illustrates a time series of a uniform group consisting of
the same 2 female students and 1 additional female student. A
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 438
full mark was achieved in the first 2 coursework, while a relative
lower grade of 90 was got in the last coursework. It is indicated
from Fig. 16(a) that the contributions of the two common female
students are different in the 3 courseworks. Fig. 16(b) compares
their contributions in the 3 courseworks, from which it can be
seen that the sum of the contributions of the 2 female students in
the coursework of 204 is larger than that in the other 2
courseworks. This indicates that learning in a mixed group can
stimulate students’ contributions compared to a uniform group.
(a) The patterns of student collaboration time series in 3 courseworks.
(b) Comparison of contributions of ME and SA in different courseworks.
Fig. 16 Learning in a mixed group stimulated students' contribution.
V. C
OLLABORATIVE
L
EARNING IN
M
IXED
G
ENDER
G
ROUPS
To isolate the impact of the collaboration style from the
individual student qualities on the expected group performance,
we proposed in [11] that the overall group coursework grade is
defined as a linear combination of the individual students’ fixed
quality measures. The deviation between the created group
performance expectation by this way and the actual grade
received is likely to be linked purely to the way students
collaborated together in the group. The impact of collaborative
learning on mixing genders within a group can then be
qualifiedly compared. It is elaborated below.
The grade of the
th
g gr oup in the
th
a
coursework is defined:
,
ˆ
=
ˆ
1=
sasa
ga
NS
s
ga
wqegrad ×
¦
(1)
where
ga
NS
is the number of students in this group,
sa
q
ˆ
represents the fixed quality measure of the
th
s
student, and
sa
w
is the corresponding weight. The quality measure of an
individual student within a group for a particular coursework is
estimated based on the grades of groups that this student
belongs to in this coursework and the coursework who has
completed before, and his/her contributions made in
corresponding courseworks, shown as
,=
ˆ
2
~
222
~asagas
gradegradeqΔ+ (2)
where
2
~
ˆ
as
q is the quality measure of the th
s
~
student in the
th
a
2
coursework, by assuming this student have finished firstly the
th
a
1
and then
th
a
2
courseworks as a member in the
th
g
1
and
th
g
2
groups, respectively. The
2
~
as
gradeΔ in (2) represents the
performance adjustment, defined as
()()
,=
1
~
2
~
22112
~
asasagagas
ctrctrgradegradegrade −×−Δ (3)
where
as
ctr
~
is the percentage of the contribution of the
th
s
~
student to the
th
a coursework, and calculated based on the
absolute number of changes as
||
1=
|
~
||||
~
|
~
==
sa
ga
NS
s
asgaasas
CCCCctr
¦
(4)
The student's performance in the first coursework is modelled
based on the percentage of his/her contribution within the
group. The performance adjustment is calculated as
,=
~~
fullgasas
gradectrgrade ×ΔΔ (5)
where
full
grade denotes the full mark, i.e. 100. In (5),
gas
ctr
~
Δ
is the deviation between the average student contribution to the
th
g group in the
th
a coursework and the actual contribution
made by the
th
s
~
student, expressed as
()
||
1=
|
~
|
~
=sa
ga
NS
s
gaasgas
CctrCctr
¦
−Δ (6)
where
ga
ctr is the average student contribution to the
th
g
group in the
th
a coursework and can be calculated as
gasa
ga
NS
s
ga
NSCctr
||
1=
=
¦
(7)
Overall there was 72 group-coursework pair instances, each
of which represents one group assigned to one coursework,
including 59 instances with uniform groups and 13 with mixed
groups. Fig. 17 shows a comparison between the actual marks
and estimated ones for both uniform and mixed groups. It is
shown that 52.54% among 59 instances with uniform groups
have their actual marks higher than the estimated ones, while 11
of 13 with mixed groups, i.e. 84.62%, achieved higher grades
than their estimation. This indicates that coeducation in groups
could largely improve learning performance.
VI. C
ONCLUSIONS
Summarising, students’ gender turned out to be an important
factor informing both the level of engagement and the quality of
formal group assignments generated in the context of the
collaborative learning environment trial at Khalifa University.
Introducing a cumulative progress timeline for visualisation of
the groupwork evolution we have uncovered interesting patterns
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 439
of student collaboration in the group that were different for
different genders. As we have shown female groups tend to work
simultaneously with rather even distribution of member
contributions while male groups reveal a preference to the
sequence of isolated individual contributions collected together.
Since female groups appear to consistently outperform the male
groups in the attained group coursework we have concluded that
the simultaneous evenly engaged groupwork exercised by the
female groups appears to better exploit the benefits and synergy
of collaborative learning through the mutual review, reflection
and information flow among the diverse group of bright students
sharing and exchanging knowledge in real-time. In fact we have
shown numerically that the payoffs of collaborative learning
with a simultaneous and evenly distributed groupwork,
exercised by the female students are much greater than the
sequences of isolated contributions typically dominated by 1
lead student, that were more visible among the male groups.
(a) Uniform vs mixed genders comparison .
(b) Male-only vs female-only groups comparison.
Fig. 17 Comparison of the expected individual average and group grades.
Another key finding of this work is the observation that
mixed-gender groups tend to perform very well in collaborative
learning exercises - almost as good as the top ranked female-
only groups. Using the introduced measure of how likely the
groupwork excels the average individual member performance,
we have observed that the mixed-gender groups outperform
even female-only groups with the 80% of chances to outperform
the average member. We explained this surprising effect with a
combination of the higher levels of group diversity in terms of
knowledge and creative content generation abilities as well as
the increased engagement and focus during groupwork
stimulated by the gender differences. We believe this outcome
delivers yet another prove of the synergies and efficiencies of
collaborative interactive learning in a diverse group of students
and encourages co-education to maximize group’s performance.
In a subsequent work we intend to extend this research by
exploiting more evidence in a form of the student profiles, data
from their journeys throughout the educational curriculum and
the actual quality of the assessed coursework content. We intend
to merge all these diverse sources of educational data and utilise
them to deliver optimised recommendations of the most suitable
course, module, individual knowledge content along with the
most effective learning style and grouping in collaborative
learning environment that would maximize students’ academic
performance and knowledge retention.
R
EFERENCES
[1] D.W. Johnson, R.T. Johnson, and K.A. Smith, "Cooperative Learning:
Increasing College Faculty Ins tructional Productivity," Washington,
D.C.: The George Washington University, School of Education and
Human Development,1991.
[2] P. Dillenbourg, "Collaborative Learning: Cognitive and Computational
Approaches. Advances in Learning and Instruction Series," New York,
NY: Elsevier Science, Inc., 1999.
[3] K. Bruffee, "Collaborative Learning," Baltimore: The Johns Hopkins
Univers ity Press, 1993.
[4] R. Mitnik, M. Recabarren, M. Nussbaum, and A. Soto, "Collaborative
Robotic Instruction: A Graph Teaching Experience," Computers and
Education, vol. 53, no. 2, pp. 330-342, 2009.
[5] M.M. Chiu, "Group problem solving processes: Social interactions and
individual actions," Journal for the Theory of Social Behaviour, vol. 30,
no. 1, pp. 27-50, 2000.
[6] M.M. Chiu, Flowing toward correct contributions during groups'
mathematics problem solving: A statistical discourse analysis, Journal of
the Learning Sciences, vol. 17, no. 3, pp. 415-463, 2008.
[7] J. M. Dirkx and R. O. Smith, "Online Collaborative Learning," T. S.
Roberts, Ed. IGI Global, 2013.
[8] E.F. Barkle y, K.P. Cross, C.H. Major, "Collaborative learning techniques:
A handbook for college faculty," San Francisco: Jossey-Bass, 2005.
[9] B. Hirsch, G.W. Hitt, L. Powell, K. Khalaf, and S. Balawi, "Collaborative
Learning in Action," Proc. of the IEEE Int. Conf. on Teaching, Assess -
ment and Learning for Engineering (TALE'2013), Bali, Indonesia, 2013.
[10] D. Ruta, L. Powell, D. Wang, B. Hirsch, and J. Ng, "Self-Organising
P2P Learning for 21C Education," International Symposium on Smart
Learning for th e Next Generation, 2013.
[11] L. Cen, D. Ruta, L. Powell, and J. Ng, "Collaborative Learning
Environment: Group Performance Prediction Using Machine Learning
Techniques," Int. Conf. on Interactive Collaborative Learning, 2014.
[12] V. Savicki, M. Kelley, and D. Lingenfelter, "Gender, group composition,
and task type in small task groups using computer-mediated
comunication," Computers in Human Behavior vol. 12, pp. 549-565, 1996.
[13] A. Gordon, "In a class of their own: boys benefit even more than gi rls
from single-sex schools, A-level grades study reveals," The Mail on
Sunday (UK), June 2000, p. 42.
[14] A. Zeid, R. El-Bahey, "Impact of Introducing Single-Gender Classrooms
in Higher Education on Student Achievement Levels- A Case Study in
Software Engine ering Courses in the GCC Region," Proc. of the 41st
ASEE/IEEE Frontiers in Education Conf., 2011.
[15] J. Goldstein, and S. Puntambekar, "The brink of change:Gender in
technology-rich collaborative learning environments," Journal of Science
Education and Technology vol. 13, no. 4,December 2004.
[16] G.S. Stump, J.C. Hilpert, J. Husman, W.T. Chung, and W. Kim,
"Collaborative Learning in Engineering Students: Gender and
Achievement," Journal of Engineering Education 100(3): 475-497 , 2011.
[17] N. Ridge, "The Hidden Gender Gap in Education in the UAE," Policy
Brief, no. 12, August, 2009.
[18] R. Chennabathni, and G. Rgskind, "Gender Issues in Collaborative
Learning," Canadian Women Studies, 17.4 (1998): 44. University of
Central Florida Library. ProQuest: GenderWatch. Web. 20 Oct. 2009.
J. O'Reilly, "Mixed school hits new heights with single-sex classes,"
Sunday Times (London), August 2000.
2014 International Conference of Teachin
g
, Assessment and Learnin
g
(TALE)
978-1-4799-7672-0/14/$31.00 ©2014 IEEE 08-10 December 2014, Wellin
g
ton, New Zealan
d
Pa
g
e 440