Education Tech Research Dev

DOI 10.1007/s11423-017-9551-0

DEVELOPMENT ARTICLE

A problem posing-based practicing strategy

for facilitating students’ computer programming skills
in the team-based learning mode

Xiao-Ming Wang1 • Gwo-Jen Hwang2

 Association for Educational Communications and Technology 2017

Abstract Computer programming is a subject that requires problem-solving strategies and

involves a great number of programming logic activities which pose challenges for
learners. Therefore, providing learning support and guidance is important. Collaborative
learning is widely believed to be an effective teaching approach; it can enhance learners’
social interaction and offer a learning environment which provides rich learning experi-
ences. However, the social interaction in collaborative learning does not occur automati-
cally. Without proper guidance strategies or supporting tools for collaborative learning, the
learning effects can be disappointing. To solve such a problem, a problem posing-based
practicing strategy was proposed to support the development of a collaborative learning
activity in a computer programming practice course. The students were guided to raise
computer programming problems to boost the discussion among team members. The
problems raised in each team were then exchanged and solved by another team to examine
the coding and to provide feedback. To investigate the effectiveness of the proposed
approach, an experiment was conducted in a C# programming course. Two classes of
students from a university participated in the experiment. One class with 25 students was
randomly assigned as the experimental group, and learned with a collaborative learning
activity using the problem posing-based practicing strategy; the other class with 28 stu-
dents was the control group, which learned with a conventional collaborative learning
activity. The results show that the proposed strategy benefited the students in terms of
improving their learning achievement, in particular, their programming skills. Moreover, it
was found that the students who learned with the proposed approach had higher self-

& Gwo-Jen Hwang

[Link]@[Link]
Xiao-Ming Wang
zsdwxm@[Link]
1
School of Teacher Education, Zhejiang Normal University, Jinhua 321004, China
2
Graduate Institute of Digital Learning and Education, National Taiwan University of Science and
Technology, Taipei 106, Taiwan, ROC

123
 X. Wang et al.

efficacy and lower cognitive load than those who learned with the conventional collabo-
rative learning approach.

Keywords Computer programming
Problem posing
Cognitive load
Team-based

learning
Self-efficacy of group learning

Introduction

Learning computer programming in teams is highly related to learning how to think

creatively, infer systematically, and communicate and collaborate effectively with peers
(Kalelioglu and Gülbahar 2014). Fostering computer programming skills is of increasing
importance and is gradually becoming one of the core objectives of undergraduate and
graduate programs in various domains, such as the software development and electronic
engineering fields. However, for students, learning to program is a difficult process,
especially for those novices without sufficient computer knowledge (Bravo et al. 2005).
Therefore, how to provide a learning strategy or cognitive tool to help students learn
computer programming is a challenging issue (Robins et al. 2003). In the past decades,
researchers have conducted many studies to boost novices’ understanding of computer
programming and increase their programming skills (Robins et al. 2003; Wang et al. 2011).
Mow (2008) pointed out that learning to program is a task with a high cognitive load, and
to have great programming skills, students need to practice repeatedly. Several researchers
have further indicated that students might lose enthusiasm and interest while programming,
especially as a result of experiencing failure and frustration (Law et al. 2010; Isong 2014;
Zhang and Yan 2014).
Learning with teamwork is deemed by educators to be a prospective approach; it can
help students get through the difficult learning tasks with the help of social interaction (Kao
et al. 2008). Several studies have shown the benefits of team-based computer programming
(Sabin and Sabin 1994; Preston 2005). On the other hand, educators have also found that,
on many occasions, learning in teams does not bring the expected social interactions;
moreover, students’ learning achievement might be disappointing due to the lack of
learning strategies (Brush 1998; Serrano-Cámara et al. 2014). Therefore, how to increase
students’ social interactions in team-based learning environments has become an important
issue.
Scholars have indicated that both cognitive and affective issues need to be taken into
account when engaging students in complex tasks such as developing computer programs
(Jonassen 2000). Regarding the cognitive issue, more complex tasks mean more cognitive
operations, and hence more working memory is required (Antonenko and Niederhauser
2010). This implies that students’ cognitive load could be a factor which affects their
programming performance. Mayer’s (1998) effort-based learning principle shows that
students’ self-efficacy could significantly affect their learning behaviors when dealing with
complex tasks; that is, they might try harder and think more deeply if they have higher self-
efficacy. Therefore, when engaging students in computer programming tasks, it is
important to measure their self-efficacy and cognitive load as well as their programming
outcomes.
Consequently, in this study, a problem posing-based practicing strategy was proposed to
boost social interactions among learners in order to promote their team-based learning
performance and self-efficacy as well as reduce their cognitive load when learning com-
puter programming. Moreover, an experiment was conducted in a C# programming course

123
A problem posing-based practicing strategy for facilitating…

to investigate the effectiveness of the proposed approach. During the learning activity, the
team members took the programming tasks given by the teacher as a reference and started
to pose problems collaboratively to design a new programming task. They then collabo-
rated to complete the task. The teacher examined the results and gave feedback. Following
that, the tasks were exchanged among teams; that is, each team tried to solve the tasks
proposed by other teams. They then started a code review activity (i.e., examining the
programs developed by others), and the teacher examined their tasks as well as the review
results.

Literature review

Computer science education

Computer science education is considered as a fundamental curriculum at all levels of

schooling around the globe (Gordon and Brayshaw 2008). In computer science courses,
students are required to learn not only the basic knowledge of computers, but also the skills
of using computers to deal with practical problems, such as the operation of computer
application software and the development of computer programs (Esteves et al. 2011;
Soloway 1993).
Scholars have indicated that computer programming is a problem-solving task that
usually involves some semantically rich areas, meaning that it requires a great amount of
domain knowledge and expertise to complete the task (Brooks 1983; Piteira and Costa
2013). In particular, the algorithm aspect of computer programming is highly related to
learners’ problem-solving skills (Papert 1980; Fessakis et al. 2013). For example, DiSessa
and Abelson (1986) indicated that in the programming activity, students need to analyze,
organize, convey, and evaluate their thoughts. Therefore, educators have emphasized the
need to promote students’ higher order thinking awareness in programming courses rather
than just teaching them to memorize the programming code or instructions (Hawi 2010;
Blignaut and Naude 2008).
In the past decades, researchers have helped learners gain computer programming skills
via different paths, including building the theories of cognitive processing and under-
standing for computer programming (Brooks 1983), analyzing how programming novices
learn from a psychological angle (Winslow 1996), evaluating the affective factors in
learners’ programming learning (Hawi 2010), and providing strategies, tools, and envi-
ronments for learning (Hadjerrouit 2008; Esteves et al. 2011; Wang et al. 2011; Shaw
2012). Rogalski and Samurçay (1990) pointed out that gaining and enhancing knowledge
in programming is a highly complex process, which involves cognitive activities and
mental representations; such cognitive activities and mental representations have relations
with programming design, understanding, adjustment, and testing. Mow (2008) indicated
that programming is a task with a heavy cognitive load which requires students to practice
many times in order to master it. Moons and de Backer (2013) also reported the importance
of taking cognitive load into account when designing supporting mechanisms for computer
programming. Researchers have further expressed that students might lose enthusiasm and
interest in the process of learning to program, especially when experiencing failure (Law
et al. 2010). Uysal (2014) emphasized that students’ first programming experience could
affect their interest in and willingness to take programming courses. Therefore, when
designing activities for learning to program, it is important to provide learning support for

123
 X. Wang et al.

students to increase their learning participation and reduce their cognitive load (Yang et al.
2015).

Collaborative learning

Collaborative learning means two or more people learn, or attempt to learn, something
together, including knowledge, skills, abilities, and so on. It is usually seen as a process
whereby people would be helped to finish a certain target or a product with designated
goals (Serrano-Cámara et al. 2014). The term ‘‘collaborative learning’’ describes a scenario
in which people hope that certain interactions would occur, and this would trigger the
mechanism of group learning. In the past 10 years, researchers have reported the results of
collaborative learning applied in different courses, such as writing (Wichmann and
Rummel 2013), English (Rick et al. 2002; Law et al. 2015), natural science (Sung and
Hwang 2013), and computer programming (Serrano-Cámara et al. 2014; Bravo et al. 2013;
Preston 2005). Research has confirmed that collaborative learning encourages learners to
use higher level cognitive strategies, critical thinking, in-depth learning and understanding,
and positive attitudes towards learning and peers; such a learning approach provides them
with a more open and flexible way to collaborate with their peers (Wang and Lin 2007;
Laal and Ghodsi 2012; Laal et al. 2013). Meanwhile, researchers also believe that col-
laborative learning is an effective method to help learners solve difficult learning tasks with
the support of peer interaction (Wang and Lin 2011).
However, the benefits of collaborative learning only appear in positive and well-func-
tioning teams. Merely dividing students into learning groups cannot guarantee an effective
collaborative learning activity (Soller 2001). Therefore, how to provide learning strategies,
tools, and an environment to boost learners’ social interaction is the key to their success.

Problem-posing and code review

Problem-posing refers to the cognitive activity in which a new problem is posed with a
given question or certain situation; it also includes the process of problem solving to
reform the problem (Lavy and Bershadsky 2003). The problem-posing strategy is a
learning strategy which engages students in posing problems based on the issue specified
by the teacher during the learning process; such a strategy has been widely applied in
mathematics courses (Silver 2013; Shuk-kwan 2013; Singer and Voica 2013; Unal and
Arikan 2015). In recent years, researchers have also implemented this strategy in other
courses (Casagrande et al. 1998; Hung et al. 2014). For instance, they have adopted the
problem-posing strategy to guide students to learn about local culture in a collaborative
learning environment, with the results indicating that the strategy can effectively increase
both the learning effects and group learning self-efficacy (Sung et al. 2016).
Many studies have shown that the problem-posing strategy can promote learners’
curiosity regarding a certain concept, which is beneficial for knowledge construction and
critical thinking ability improvement (Moses et al. 1990). Researchers have also found that
when the teacher adopts the problem-posing strategy in school settings, the students
usually more actively engage in creating and solving their problems, and hence have better
learning performance (Barlow and Cates, 2006).
Furthermore, code review is an activity which engages students in examining pro-
gramming code. It was originally proposed by Fagan (1976) in the field of software
engineering to examine the quality of computer software (Mantyla and Lassenius 2009;
Liu et al. 2001). The successful application of code review in software engineering has

123
A problem posing-based practicing strategy for facilitating…

encouraged educators to use it as a learning strategy in computer science education (Wang

et al. 2012). For instance, Wang et al. (2012) indicated that engaging students in peer code
review could significantly improve their learning outcomes, including computer pro-
gramming skills, collaborative learning, code norms, time management, and so on.
In this study, a problem posing-based practicing strategy was proposed to help learners
experience collaborative learning in the practice of programming. The strategy is proposed
based on the theories of constructivism and collaboration (Merrill 1991; Yakimovicz and
Murphy 1995), and aims to engage students in higher order thinking activities, as indicated
by Bloom (1994). During the learning process, the problem-posing strategy with follow-up
code review tasks was used to promote the learners’ involvement and interactions in the
collaborative learning tasks. The theoretical basis of the approach can be further explained
by the social constructivist theory, which emphasizes the role of social interactions in the
learning process (Vygotsky 1962). Vygotsky indicated that knowledge is co-constructed;
that is, students learn from interacting with peers. Therefore, he highlighted the importance
of engaging learners in a process in which they can learn from each other and receive
comments or assistance from peers. Au (1998) pointed out the effectiveness of applying the
theory to school settings; for example, teachers can design learning activities to encourage
those students who know the learning tasks and content better to work with those who do
not.
In order to investigate the effectiveness of the proposed approach, an experiment was
conducted in a C# programming course to answer the following research questions:
1. Compared to conventional collaborative learning: Can collaborative learning with the
problem posing-based practicing strategy significantly improve learners’ learning
achievement?
2. Compared to conventional collaborative learning: Can collaborative learning with the
problem posing-based practicing strategy significantly improve learners’ self-efficacy
of group learning?
3. Compared to conventional collaborative learning: Can collaborative learning with the
problem posing-based practicing strategy significantly lower learners’ cognitive load?

Problem posing-based practicing strategy for training computer

programming skills

The online team-based learning environment with the problem posing-based practicing
strategy was developed by integrating the Wiki function module in the Moodle learning
management system. The learning environment allowed each team of students to edit and
provide comments; in addition, it recorded each version of the edited results to enable the
students to compare and see the changes between different versions.
In this study, the teacher used the grouping function to divide the students into learning
teams. Besides, the teacher assigned two teams to work as a pair to exchange the pro-
gramming tasks and conduct the coding review. Each team had two Wiki pages, one for
problem-posing and the other for code review. Take Team i and Team j as an example; the
problem-posing page for Team j was only visited by the members of Team j, but the code
review page for Team j could be visited by the members of both Teams i and j. Each team
needed to complete two programming tasks on these two pages: one was designed by
themselves, and the other was designed by the other team. For instance, the members of

123
 X. Wang et al.

Team i not only posed a new programming task and provided the answer on the problem-
posing page, but were also asked to complete the task proposed by Team j.
Figure 1 presents the application of the problem posing-based practicing strategy for
Team i and Team j. When the students logged into the system and found their own group,
they started posing problems on the problem-posing page.
There are five steps in the problem-posing activity. In the first step, the team members
reviewed the learning contents based on the examples and problem-posing requirements
provided by the teacher. The examples and learning contents given by the teacher for Team
i and Team j were different. For instance, in Fig. 1, the content that Team i needed to pose
problems for was an abstract class in C#, while for Team j, the students needed to pose
problems about the interface of C#.
In the second step, after fully understanding the knowledge and the task requirements,
the team members linked what they had learned with their life experience and voiced their
opinions to settle the topic of the programming task. They then discussed the details of the
task. For instance, the problem that Team i needed to pose was about an abstract class.
They chose to calculate the area of a shape as the topic and requested to calculate the area
of a circle and a sphere in their derived class. After posing the most preliminary problem,
the team members entered the content of the task into the problem-posing page, as pre-
sented in Fig. 2. In should be noted that the learning content and system interface were in
Chinese.

Team i Team j
Log in Log in Log in

Step 1 Step 1
Review learning contents based on the Review learning contents based on the
examples and problem-posing examples and problem-posing
requirements. requirements.

Step 2 Step 2
Pose the preliminary problems. Pose the preliminary problems.

Step 3 Step 3
Gradually finish problem-posing based Gradually finish problem-posing based
on the teacher’s feedback. on the teacher’s feedback.

Step 4 Step 4
Finish the preliminary answers to the Finish the preliminary answers to the
problem. problem.

Step 5 Step 5
Give a thorough answer based on the Give a thorough answer based on the
teacher’s feedback. teacher’s feedback.
Problem-posing stage Problem-posing stage
Team i Team j

Team j and Team i

Give answers to the correct each other’s Give answers to the
tasks posed by Team j. answers and give tasks posed by Team i.
feedback.
Between-team interacons

Fig. 1 Problem posing-based practicing strategy mechanism for Team i and Team j

123
A problem posing-based practicing strategy for facilitating…

Problem-posing requirements
from the teacher

History record Examples given

by the teacher
Eding funcon Comment funcon
for students for the teacher

Outline of the Result from student A Student B gives comments on and

curriculum page correcons to Student A’s result.

Fig. 2 Problem-posing page

In the third step, after the team members had finished the preliminary problem-posing,
using the comment function in Wiki, the teacher gave feedback for correction, checking the
discreteness and difficulty level of the task, and if the posed problem met the requirements.
The students subsequently worked together to improve the task based on the teacher’s
comments, using the Wiki collaboration function.
In the fourth step, after posing the programming task, the team members needed to
complete the task. They could work on the programming alone and then collaborate to
reach the final answer. It could also be conducted by having one member enter the coding
and letting others do the corrections. No matter which way they decided to collaborate,
they all needed to discuss the problem and reach a consensus. The history function in Wiki
recorded the editing process for each member. As shown in Fig. 2, Student B could adjust
the contents written by Student A and the system would record the adjustment for the users
to compare the two versions.
In the fifth step, the teacher gave feedback on the programming, such as the correctness
of the solution process and the code norms. The team members then adjusted the pro-
gramming based on the suggestions given by other groups. Meanwhile, the teacher went
around the class to observe the situation. Once each group had finished the problem-posing
stage, it went on to the code review stage.
In the code review activity, two matched groups exchanged their programming tasks.
For instance, Team i would put their task about abstract class on Team j’s code review
Wiki page, while Team j would put their task about the interface on Team i’s code review
Wiki page. Then, the members of the two groups worked on each other’s tasks. After
finishing the tasks, the two groups gave each other feedback. If the results were too diverse,
the students could discuss them and the teacher could also make comments.

123
 X. Wang et al.

Methodology

Learning materials

To evaluate the effectiveness of the problem posing-based practicing strategy, a quasi-

experiment was conducted in a C# course. This course requires students to gain C# object-
oriented programming knowledge (i.e., knowing how a computer program works and the
syntax of the programming language) and programming skills (i.e., knowing how to
analyze a task and develop a program for dealing with the task). The object-oriented
concept is closely linked to the real world by providing several features in programming,
such as ‘‘encapsulation,’’ ‘‘inheritance,’’ and ‘‘polymorphism’’ (Winslow 1996).

Participants and design

A quasi-experiment was adopted, and the participants were 53 sophomores (second year
students) from two classes at a university located in eastern China. The learning activity
was conducted in a C# programming course. They had already studied C programming in
their freshman year and knew the procedure-oriented programming method. One class with
25 students (5 males and 20 females) was assigned to be the treatment group and the other
class with 28 students (6 males and 22 females) was the control group. The ratios of male
and female students for the two groups were very close; that is, 20 and 80% for the
treatment group, and 21 and 79% for the control group. The students in both groups were
majoring in Educational Information Technology, and had an average age of 20. All of the
students in the treatment group and the control group were taught by the same teacher who
has 15 years of teaching experience. During the quasi-experiment, the students’ personal
information was hidden so as to protect their privacy; moreover, they were informed that
joining the experiment was voluntary and would not affect their grades, and they were
allowed to withdraw from the experiment at any stage. It should be noted that in the
department of Educational Information Technology in many Asian countries, the students
need to take courses in both computer technology and educational technology.
In this experiment, through the implementation of the problem posing-based practicing
strategy in the collaborative learning activity, learners were guided to link the program-
ming learning with the real world to cultivate their problem-solving conceptions in the real
world and to enhance their object-oriented programming thinking. Meanwhile, compared
to conventional collaborative learning, whether such an approach can improve students’
learning was observed, in terms of learning achievement, cognitive load, and self-efficacy
for group learning.

Experimental procedure

Figure 3 presents the experiment process. Before the learning activity, all of the students
were instructed by the same teacher for 7 weeks in the C# programming course, including
the basic syntax and class of C#. Then, all of the students took a pre-test and completed a
pre-questionnaire. The purpose of the pre-test was to evaluate the students’ prior knowl-
edge of C# programming and programming skills, while the pre-questionnaire included an
assessment of the students’ self-efficacy for group learning.
Following that, all of the students received the instruction on the ‘‘inheritance’’ concept
in C# programming with corresponding examples for 2 weeks. Then, the treatment group

123
A problem posing-based practicing strategy for facilitating…

Treatment Group Control Group

(n=25) (n=28)
Preliminary learning 7 week
Basic syntax and class of C#
Programming knowledge and skill pre-test
60 mins
Basic syntax and class of C#
Pre-questionnaire
10 mins
Self-efficacy of group learning

Instruction on Inheritance and Polymorphism of C# 2 weeks

Programming design practice Programming design practice 5 weeks

with problem posing-based with conventional
collaborative learning collaborative learning

Learning achievement post-test 80 mins

Inheritance and polymorphism of C#
Post-questionnaire 10 mins
Self-efficacy of group learning and cognitive load
Fig. 3 Experimental design of the learning activity

and the control group learned with the two different teaching strategies in the coming
5 weeks.
The students in the treatment group learned with the problem posing-based practicing
strategy, using the Wiki page in Moodle to engage in the collaborative problem-posing and
code review activities. During the learning activity, the members of the treatment group
were requested to finish two programming tasks on inheritance, one of which was proposed
by the team members, based on the examples given by the teacher in class, while the other
was proposed by the other matching team.
On the other hand, the students in the control group were guided to complete the
learning tasks in teams with the conventional collaborative learning approach. After log-
ging into the Moodle system and collaborating with their team members via discussion,
they directly completed two programming tasks given by the teacher, and then submitted
their results. The teacher then gave feedback on the students’ programs and provided
example programs to them. The students then reviewed the code provided by the teacher,
and examined and corrected their own programs by referring to the feedback provided by
the teacher and the example programs. Following that, the teacher answered the questions
raised by the students. Such a collaborative learning approach had been employed in the
selected school for years.
After the learning activity, a post-test and a post-questionnaire were conducted. The
content of the post-test included inheritance and polymorphism of C#, while that of the
post-questionnaire included group learning self-efficacy and cognitive load.

123
 X. Wang et al.

Measuring tools

The measuring tools in this study were composed of the pre-test and post-test of learning
achievement, the pre-questionnaire of self-efficacy for group learning, and the post-
questionnaire of self-efficacy for group learning and cognitive load.
The purpose of the pre-test was to evaluate whether the students in both the treatment
group and the control group met the same minimum standard for learning the programming
knowledge and skills, while that of the post-test was to investigate whether any difference
existed as a result of the two groups of students adopting different learning strategies.
There were 14 multiple-choice items (70%) and one design item (30%) in the pre-test, and
8 multiple-choice items (40%) and two design items (60%) in the post-test, both with a
perfect score of 100. The multiple-choice items aimed to test the students’ C# program-
ming knowledge, while the design items were to test the students’ problem-solving pro-
gramming skills. Both the pre-test and the post-test were designed by two experienced
teachers, each with 15 years of programming teaching experience. The Cohen’s kappa
coefficients of the scores graded by the two teachers were 0.97 and 0.98 for the pre-test and
post-test, respectively, showing high consistency between their grading.
The self-efficacy of the group learning questionnaire was adopted from the one
developed by Wang and Lin (2007), which was revised from the original questionnaire of
Pintrich et al. (1991). There were eight items in the questionnaire, which was used to test
the individuals’ judgement of group ability and the evaluation of the group ability for the
upcoming tasks, such as ‘‘I am certain I can understand the most difficult material pre-
sented in the readings for this course’’ and ‘‘I am confident I can understand the basic
concepts taught in this course.’’ A 5-point Likert scheme was used, with 5 representing
Strongly agree, 4 representing Agree, 3 representing Neutral, 2 representing Disagree, and
1 representing Strongly disagree. The Cronbach’s alpha value for the self-efficacy of group
learning is .93, showing good internal consistency of the items, as indicated by Christmann
and van Aelst (2006) and Sijtsma (2009).
The cognitive load questionnaire was modified by Hwang et al. (2013) based on the
cognitive load measures proposed by Paas (1992) and Sweller et al. (1998). It includes two
dimensions, mental load (e.g., ‘‘The learning content in this learning activity was difficult
for me’’ and ‘‘I had to put a lot of effort into answering the questions in this learning
activity’’) and mental efforts (e.g., ‘‘The instructional way in the learning activity was
difficult to follow and understand’’). The former refers to the intrinsic cognitive load
caused by the difficulty or large amount of the learning content, while the latter refers to
the extraneous cognitive load caused by the teaching strategy (i.e., the way the learning
content is organized and presented). There are five items for mental load and three for
mental efforts. A 7-point Likert scheme was adopted for each item, with 7 representing
Strongly agree and 1 representing Disagree. A higher value means that the students have
higher cognitive loading in the learning process. The Cronbach’s alpha values for the two
dimensions are .92 and .84, respectively.

123
A problem posing-based practicing strategy for facilitating…

Results

Analysis of learning achievement

The analysis of covariance (ANCOVA) was employed by using the pre-test scores as the
covariant, the post-test scores as the dependent variable, and the learning approach as the
independent variable in order to investigate whether the treatment and control groups
showed significant differences in their learning achievements by excluding the interaction
effects of the pre-test.
The test of the homogeneity of regression coefficients on the two groups’ pre-test scores
showed that the homogeneity of variance was not violated with F = 0.000 and
p = 0.998 [ 0.05. Following that, ANCOVA was then conducted on the two groups’ post-
test scores by excluding the impact of the pre-test scores, as shown in Table 1. The results
show that there was a significant difference in the scores of the students who adopted the
different learning strategies for collaborative learning (F = 5.09, p = 0.028 \ 0.05), with
the treatment group scoring higher than the control group. Therefore, it is concluded that
the problem posing-based practicing strategy used in collaborative learning can signifi-
cantly improve students’ learning achievement compared with conventional collaborative
learning. The Eta squared value g2 = 0.09 implies a medium to large effect size.
To further investigate the effects of the different collaborative learning strategies on the
students’ learning of programming knowledge and programming skills, the scores for the
multiple-choice items and design items in the pre- and post-tests were analyzed.
The test of the homogeneity of regression coefficients on the two groups’ pre-test scores
of programming knowledge and skills showed that the homogeneity of variance was not
violated with F = 0.68 (p = 0.42 [ 0.05) and F = 0.32 (p = 0.57 [ 0.05), respectively.
That is, ANCOVA can be used to compare the two groups’ scores of programming
knowledge and skills.
Table 2 shows the ANCOVA results of the two groups’ programming knowledge and
skills. It can be seen that the adjusted mean scores of the post-test for the multiple-choice
items in the treatment group and the control group were 25.77 and 24.52, respectively,
meaning that with the different collaborative learning strategies, the students still had
similar basic knowledge of C# (F = 0.686, p = 0.41 [ 0.05). On the other hand, the
ANCOVA results showed that the adjusted mean scores of the post-test for the design
items in the treatment group and the control group were 32.46 and 24.31, meaning that
with the different collaborative learning strategies, a significant difference was reached in
the design items of the post-test score, with the mean score of the treatment group higher
than that of the control group (F = 5.67, p = 0.021 \ 0.05). The Eta squared value
g2 = 0.1 implies a medium to large effect size.

Table 1 The ANCOVA result of the learning achievements post-test for the two groups
Group N Mean SD Adjusted Mean SE F g2

Treatment group 25 57.88 15.48 57.85 2.83 5.09* 0.09

Control group 28 49.11 17.27 49.14 2.68

* p \.05

123
 X. Wang et al.

Table 2 The ANCOVA result of the programming knowledge and skills post-test for the two groups
Variance Group N Mean SD Adjusted mean SE F g2

Programming knowledge Treatment 25 25.77 5.72 25.75 1.09 0.69 0.01

Control 28 24.52 5.50 24.51 1.03
Programming skill Treatment 25 32.08 12.47 32.46 2.49 5.67* 0.10
Control 28 24.64 14.65 24.31 2.35
* p \.05

Analysis of cognitive load

An independent t test analysis was conducted on the post-questionnaire of cognitive load,

as presented in Table 3. It can be seen that the mean score of cognitive load for the control
team was 4.66, while that for the treatment group was 3.70, meaning that the students who
learned with the problem posing-based practicing strategy had significantly lower cognitive
load than the students in the control group (t = 2.69, p = 0.01 \ 0.05). The Cohen’s d is
0.75, showing a medium to large effect size.

Analysis of self-efficacy of group learning

ANCOVA was employed by using the pre-questionnaire ratings as the covariant, the post-
questionnaire ratings as the dependent variable, and the learning approach as the inde-
pendent variable in order to investigate whether the treatment and control groups showed
significant differences in their group learning self-efficacy by excluding the interaction
effects of the pre-questionnaire ratings.
The test of the homogeneity of regression coefficients on the two groups’ pre-ques-
tionnaire ratings of self-efficacy of group learning showed that the homogeneity of vari-
ance was not violated with F = 0.453 with p [ 0.05. ANCOVA was then conducted on the
post-questionnaire ratings of self-efficacy of group learning. Table 4 shows the results for
the two groups. It can be seen that the adjusted mean scores for the treatment group and the
control group were 4.18 and 3.76, meaning that the treatment group significantly improved
their self-efficacy of group learning, compared to the control group (F = 4.90,
p = 0.03 \ 0.05). The Eta squared value g2 = 0.09 implies a medium to large effect size.
The results showed that the students learning with the problem posing-based practicing
strategy collaborative learning showed better self-efficacy for group learning than those
learning with the conventional approach. This result matches the result that the treatment
group students had higher learning achievement and lower cognitive load.

Table 3 Independent sample t test of the cognitive loads of the two groups
Group N Mean SD t p d

Treatment 25 3.70 1.21 2.69* 0.01 0.75

Control 28 4.66 1.39

* p \.05

123
A problem posing-based practicing strategy for facilitating…

Table 4 The ANCOVA result of the self-efficacy of group learning post-test for the two groups
Group N Mean SD Adjusted mean SE F g2

Treatment 25 4.20 0.50 4.18 0.14 4.90* 0.09

Control 28 3.74 0.77 3.76 0.13

* p \.05

Discussion and conclusions

A problem posing-based practicing strategy was proposed in this study. The results present
that, compared to conventional collaborative learning, the problem posing-based practicing
strategy can significantly improve students’ programming skills and group learning self-
efficacy, while also lowering their cognitive load.
In terms of learning achievement, some studies showed that learners can gain positive
learning achievement by having course content-related learning tasks and guiding them to
actively engage in construction, creation, and reflection activities to promote peer inter-
action as well as their problem-solving abilities (Lai and Hwang 2015; Hardy et al. 2014;
Auttawutikul et al. 2014). It is inferred that the problem-posing with follow-up code review
activities was in line with the social constructivist theory, which emphasizes the impacts of
social interactions on students’ cognitive process. In this study, the aim of the proposed
approach was to engage learners in a process whereby they could learn from each other and
receive comments or assistance from their peers. Such an approach offered the treatment
group students not only opportunities to deal with the problems raised by their peers and to
receive comments from their peers, but also to review their peers’ work, which situated
them in a learning environment to learn from each other. Such an activity reinforced the
learners’ active participation and responsible attitude towards engaging in learning. Fur-
thermore, the code review activity improved the learners’ interaction and gave them a
chance to reflect on and reinforce their understanding of programming problems and
coding. As indicated by Wang et al. (2012), engaging students in peer-review tasks enables
them to reflect on and reinforce their understanding of the learning content. In the learning
process of the problem posing-based practicing strategy, the learners continued partici-
pating in the interaction of knowledge construction, problem creation, problem solving,
comment giving, and reflection. This might be the reason why the treatment group out-
performed the control group in terms of learning achievement.
In terms of cognitive load, Antonenko and Niederhauser (2010) indicated that, in
addition to intrinsic load and extraneous load, germane load also contributes to the total
load caused by learning materials or tasks. They also indicated that, while germane load
improved cognition and learning, extraneous load did the opposite owing to the limitations
caused by the learning design. In this study, after teacher instruction in the control group,
the students followed the programming tasks given by the teacher to find the answer;
following that, they reviewed their own programming code as well as the feedback and
standard answer provided by the teacher to correct their code. On the other hand, the
students in the treatment group were provided with example tasks at the beginning of the
learning activity. With such scaffolding, they could pose a problem within a certain
knowledge range and collaborate to solve it. During the process, the learners were not
limited to one single task and could be more actively engaged and creative, which can meet

123
 X. Wang et al.

the goal-free effect of cognitive load. This might be the reason why the treatment group
was able to have not only better learning achievement, but also lower cognitive load. The
study of Antonenko and Niederhauser (2010) also supports this point; that is, the goal-free
effect allows learners to express themselves without limitation, triggers creation, and hence
lowers their cognitive load and helps them achieve meaningful learning (Sweller
1994, 2010).
It should be noted that there are several limitations to this study. First, the sample size of
the participants was not large, and hence the findings cannot be generalized to represent the
cases with a large number of learners. Second, in the present study, the proposed approach
was applied to the selected units of the programming course for 7 weeks; therefore, the
impacts of the strategy cannot be generalized to all of the computer programming units.
Third, the effect sizes for the statistical results of cognitive load and self-efficacy were not
large, implying that the generalization of the findings in these dimensions is limited. In
addition, in the proposed strategy, it is assumed that the learners have equivalent basic
programming knowledge for posing problems in the first stage as well as reviewing peer
work in the second stage. For other courses in which learners are from different depart-
ments with different levels of basic knowledge or skills, additional considerations and
treatments could be required.
For future research, it is suggested that long-term experiments with large sample sizes
are required. It is also worth applying the strategy to other computer programming units to
investigate its impacts from different aspects, such as the tendencies of collaboration and
communication in team-based programming.

Acknowledgments This study is supported in part by the Key Research Base of Humanities and Social
Sciences in Colleges and Universities of Zhejiang Province (Zhejiang Normal University Education Base).

Compliance with ethical standards

Conflict of interest The authors would like to declare that there is no conflict of interest in this study.

References
Antonenko, P. D., & Niederhauser, D. S. (2010). The influence of leads on cognitive load and learning in a
hypertext environment. Computers in Human Behavior, 26(2), 140–150.
Au, K. H. (1998). Social constructivism and the school literacy learning of students of diverse backgrounds.
Journal of literacy research, 30(2), 297–319.
Auttawutikul, S., Wiwitkunkasem, K., & Smith, D. R. (2014). Use of weblogs to enhance group learning and
design creativity amongst students at a Thai University. Innovations in Education and Teaching
International, 51(4), 378–388.
Barlow, A. T., & Cates, J. M. (2006). The impacts of problem posing on elementary teachers’ belief about
mathematics and mathematics teaching. School Science and Mathematics, 106, 64–73.
Blignaut, P., & Naude, A. (2008). The influence of temperament style on a student’s choice of and per-
formance in a computer programming course. Computers in Human Behavior, 24(3), 1010–1020.
Bloom, B. S. (1994). Reflections on the development and use of the taxonomy. Yearbook: National Society
for the Study of Education, 92(2), 1–8.
Bravo, C., Duque, R., & Gallardo, J. (2013). A groupware system to support collaborative programming:
Design and experiences. Journal of Systems and Software, 86(7), 1759–1771.
Bravo, C., Marcelino, M. J., Gomes, A. J., Esteves, M., & Mendes, A. J. (2005). Integrating educational
tools for collaborative computer programming learning. J. UCS, 11(9), 1505–1517.
Brooks, R. (1983). Towards a theory of the comprehension of computer programs. International Journal of
Man-Machine Studies, 18(6), 543–554.
Brush, T. A. (1998). Embedding cooperative learning into the design of integrated learning systems:
rationale and guidelines. Educational Technology Research and Development, 46(3), 5–18.

123
A problem posing-based practicing strategy for facilitating…

Casagrande, L. D. R., Caron-Ruffino, M., Rodrigues, R. A. P., Vendrusculo, D. M. S., Takayanagui, A.

M. M., Zago, M. M. F., et al. (1998). Problem-posing in education: Transformation of the practice of
the health professional. Patient Education and Counseling, 33(2), 161–167.
Christmann, A., & Van Aelst, S. (2006). Robust estimation of Cronbach’s alpha. Journal of Multivariate
Analysis, 97(7), 1660–1674.
diSessa, A. A., & Abelson, H. (1986). Boxer: A reconstructible computational medium. Communications of
the ACM, 29(9), 859–868.
Esteves, M., Fonseca, B., Morgado, L., & Martins, P. (2011). Improving teaching and learning of computer
programming through the use of the Second Life virtual world. British Journal of Educational
Technology, 42(4), 624–637.
Fagan, M. E. (1976). Design and code inspections to reduce errors in program development. IBM Journal of
Research and Development, 15(3), 182.
Fessakis, G., Gouli, E., & Mavroudi, E. (2013). Problem solving by 5–6 years old kindergarten children in a
computer programming environment: A case study. Computers & Education, 63, 87–97.
Gordon, N. A., & Brayshaw, M. (2008). Inquiry-based learning in computer science teaching in higher
education. Innovations in Teaching and Learning in Information and Computer Sciences, 7(1), 22–33.
Hadjerrouit, S. (2008). Towards a blended learning model for teaching and learning computer programming:
A case study. Informatics in Education, 7(2), 181–210.
Hardy, J., Bates, S. P., Casey, M. M., Galloway, K. W., Galloway, R. K., Kay, A. E.,… & McQueen, H. A.
(2014). Student-generated content: Enhancing learning through sharing multiple-choice ques-
tions. International Journal of Science Education, 36(13), 2180-2194
Hawi, N. (2010). Causal attributions of success and failure made by undergraduate students in an intro-
ductory-level computer programming course. Computers & Education, 54(4), 1127–1136.
Hung, C. M., Hwang, G. J., & Wang, S. Y. (2014). Effects of an integrated mind–mapping and problem–
posing approach on students’ in–field mobile learning performance in a natural science course. In-
ternational Journal of Mobile Learning and Organisation, 8(3–4), 187–200.
Hwang, G. J., Yang, L. H., & Wang, S. Y. (2013). A concept map-embedded educational computer game for
improving students’ learning performance in natural science courses. Computers & Education, 69,
121–130.
Isong, B. (2014). A methodology for teaching computer programming: first year students’ perspective.
International Journal of Modern Education and Computer Science, 6(9), 15.
Jonassen, D. H. (2000). Toward a design theory of problem solving. Educational technology research and
development, 48(4), 63–85.’’ based on this comment.
Kalelioglu, F., & Gülbahar, Y. (2014). The effects of teaching programming via scratch on problem solving
skills: A discussion from learners’ perspective. Informatics in Education, 13(1), 33.
Kao, G. Y. M., Lin, S. S., & Sun, C. T. (2008). Beyond sharing: Engaging students in cooperative and
competitive active learning. Educational Technology & Society, 11(3), 82–96.
Laal, M., & Ghodsi, S. M. (2012). Benefits of collaborative learning. Procedia-Social and Behavioral
Sciences, 31, 486–490.
Laal, M., Naseri, A. S., Laal, M., & Khattami-Kermanshahi, Z. (2013). What do we achieve from learning in
collaboration? Procedia-Social and Behavioral Sciences, 93, 1427–1432.
Lai, C. L., & Hwang, G. J. (2015). An interactive peer-assessment criteria development approach to
improving students’ art design performance using handheld devices. Computers & Education, 85,
149–159.
Lavy, I., & Bershadsky, I. (2003). Problem posing via ‘‘what if not?’’ strategy in solid geometry—a case
study. The Journal of Mathematical Behavior, 22(4), 369–387.
Law, Q. P. S., Chung, J. W. Y., Leung, C. C., & Wong, T. K. S. (2015). Enhancement of self-efficacy and
interest in learning english of undergraduate students with low english proficiency through a collab-
orative learning programme. American Journal of Educational Research, 3(10), 1284–1290.
Law, K. M., Lee, V. C., & Yu, Y. T. (2010). Learning motivation in e-learning facilitated computer
programming courses. Computers & Education, 55(1), 218–228.
Liu, E. Z. F., Lin, S. S., Chiu, C. H., & Yuan, S. M. (2001). Web-based peer review: the learner as both
adapter and reviewer. IEEE Transactions on Education, 44(3), 246–251.
Mantyla, M. V., & Lassenius, C. (2009). What types of defects are really discovered in code reviews? IEEE
Transactions on Software Engineering, 35(3), 430–448.
Mayer, R. E. (1998). Cognitive, metacognitive, and motivational aspects of problem solving. Instructional
Science, 26(1–2), 49–63.
Merrill, D. (1991). Constructivism and instructional design. Educational Technology, 31(5), 45–53.

123
 X. Wang et al.

Moons, J., & De Backer, C. (2013). The design and pilot evaluation of an interactive learning environment
for introductory programming influenced by cognitive load theory and constructivism. Computers &
Education, 60(1), 368–384.
Moses, B., Bjork, E., & Goldenberg, E. P. (1990). Beyond problem solving: Problem posing. In T. J. Cooney
(Ed.), Teaching and learning mathematics in the 1990s (pp. 82–91). Reston, VA: National Council of
Teachers of Mathematics.
Mow, I. C. (2008). Issues and difficulties in teaching novice computer programming. In Innovative tech-
niques in instruction technology, e-learning, e-assessment, and education (pp. 199–204). Berlin:
Springer.
Paas, F. G. W. (1992). Training strategies for attaining transfer of problem-solving skill in statistics: a
cognitive load approach. Journal of Educational Psychology, 84(4), 429–434.
Papert, S. (1980). Mindstorms: Children, computers, and powerful ideas. Basic Books, Inc.
Pintrich, P.R., Smith, D.A.F., Garcia, T., & McKeachie, W.J. (1991). A manual for the use of the motivated
strategies for learning questionnaire (MSLQ). MI: National Center for Research to Improve Postsec-
ondary Teaching and Learning. (ERIC Document Reproduction Service No. ED 338122)
Piteira, M., & Costa, C. (2013, July). Learning computer programming: study of difficulties in learning
programming. In Proceedings of the 2013 International Conference on Information Systems and
Design of Communication (pp. 75–80). ACM.
Preston, D. (2005). Pair programming as a model of collaborative learning: a review of the research. Journal
of Computing Sciences in Colleges, 20(4), 39–45.
Rick, J., Guzdial, M., Holloway-Attaway, K. C. L., & Walker, B. (2002, January). Collaborative learning at
low cost: CoWeb use in English composition. In: Proceedings of the Conference on Computer Support
for Collaborative Learning: Foundations for a CSCL Community (pp. 435–442). International Society
of the Learning Sciences.
Robins, A., Rountree, J., & Rountree, N. (2003). Learning and teaching programming: A review and
discussion. Computer Science Education, 13(2), 137–172.
Rogalski, J., & Samurçay, R. (1990). Acquisition of programming knowledge and skills. Psychology of
Programming, 18, 157–174.
Sabin, R. E., & Sabin, E. P. (1994, March). Collaborative learning in an introductory computer science
course. In: ACM SIGCSE Bulletin (Vol. 26, No. 1, pp. 304–308). ACM.
Serrano-Cámara, L. M., Paredes-Velasco, M., Alcover, C. M., & Velazquez-Iturbide, J. Á. (2014). An
evaluation of students’ motivation in computer-supported collaborative learning of programming
concepts. Computers in Human Behavior, 31, 499–508.
Shaw, R. S. (2012). A study of the relationships among learning styles, participation types, and performance
in programming language learning supported by online forums. Computers & Education, 58(1),
111–120.
Shuk-kwan, S. L. (2013). Teachers implementing mathematical problem posing in the classroom: challenges
and strategies. Educational Studies in Mathematics, 83(1), 103–116.
Sijtsma, K. (2009). On the use, the misuse, and the very limited usefulness of Cronbach’s alpha. Psy-
chometrika, 74(1), 107.
Silver, E. A. (2013). Problem-posing research in mathematics education: Looking back, looking around, and
looking ahead. Educational Studies in Mathematics, 83(1), 157–162.
Singer, F. M., & Voica, C. (2013). A problem-solving conceptual framework and its implications in
designing problem-posing tasks. Educational Studies in Mathematics, 83(1), 9–26.
Soller, A. (2001). Supporting social interaction in an intelligent collaborative learning system. International
Journal of Artificial Intelligence in Education, 12, 40–62.
Soloway, E. (1993). Should we teach students to program? Communications of the ACM, 36(10), 21–25.
Sung, H. Y., & Hwang, G. J. (2013). A collaborative game-based learning approach to improving students’
learning performance in science courses. Computers & Education, 63, 43–51.
Sung, H. Y., Hwang, G. J., & Chang, Y. C. (2016). Development of a mobile learning system based on a
collaborative problem-posing strategy. Interactive Learning Environments, 24(3), 456–471.
Sweller, J. (1994). Cognitive load theory, learning difficulty, and instructional design. Learning and
Instruction, 4(4), 295–312.
Sweller, J. (2010). Element interactivity and intrinsic, extraneous, and germane cognitive load. Educational
psychology review, 22(2), 123–138.
Sweller, J., Van Merriënboer, J. J. G., & Paas, F. G. W. C. (1998). Cognitive architecture and instructional
design. Educational Psychology Review, 10(3), 251–297.
Unal, H., & Arikan, E. E. (2015). An investigation of eighth grade students’ problem posing skills (Turkey
sample). International Journal of Research in Education and Science, 1(1), 23–30.

123
A problem posing-based practicing strategy for facilitating…

Uysal, M. P. (2014). Improving first computer programming experiences: The case of adapting a web-
supported and well-structured problem-solving method to a traditional course. Contemporary Edu-
cational Technology, 5(3), 198–217.
Vygotsky, L. S. (1962). Thought and language. Cambridge, MA: MIT Press.
Wang, Y., Li, H., Feng, Y., Jiang, Y., & Liu, Y. (2012). Assessment of programming language learning
based on peer code review model: Implementation and experience report. Computers & Education,
59(2), 412–422.
Wang, S. L., & Lin, S. S. (2007). The effects of group composition of self-efficacy and collective efficacy on
computer-supported collaborative learning. Computers in Human Behavior, 23(5), 2256–2268.
Wang, T., Su, X., Ma, P., Wang, Y., & Wang, K. (2011). Ability-training-oriented automated assessment in
introductory programming course. Computers & Education, 56(1), 220–226.
Wichmann, A., & Rummel, N. (2013). Improving revision in wiki-based writing: Coordination pays off.
Computers & Education, 62, 262–270.
Winslow, L. E. (1996). Programming pedagogy-a psychological overview. ACM SIGCSE Bulletin, 28(3),
17–22.
Yakimovicz, A. D., & Murphy, K. L. (1995). Constructivism and collaboration on the Internet: case study of
a graduate class experience. Computers & Education, 24(3), 203–209.
Yang, T. C., Hwang, G. J., Yang, S. J., & Hwang, G. H. (2015). A two-tier test-based approach to improving
students’ computer-programming skills in a web-based learning environment. Journal of Educational
Technology & Society, 18(1), 198.
Zhang, Y., & Yan, D. (2014). Curriculum reform of C language programming and cultivation of compu-
tational thinking. Advances in Natural Science, 7(4), 49–53.

Xiao-Ming Wang is a lecture in the School of Teacher Education, Zhejiang Normal University, China. His
research interests include computer education, computational thinking and web-based learning.

Gwo-Jen Hwang is a Chair Professor at National Taiwan University of Science and Technology, Taipei,
Taiwan. Dr. Hwang has published more than 330 academic papers. His research interests include e-learning,
computer mediated communication, intelligent tutoring system, and expert system. Owing to the good
reputation in academic research and innovative inventions of e-learning, in 2007 and 2011, he received the
annual Most Outstanding Researcher Award from the National Science Council in Taiwan.

123