See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/320800938

Learning programming from Scratch

Article in Turkish Online Journal of Educational Technology · July 2017

CITATIONS READS

19 26,732

3 authors, including:

Monika Mladenović Sasa Mladenovic

University of Split University of Split
44 PUBLICATIONS 397 CITATIONS 63 PUBLICATIONS 417 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Monika Mladenović on 05 August 2020.

The user has requested enhancement of the downloaded file.

Learning Programming From Scratch

Monika
University of Split, Faculty of Science
Croatia
[email protected]

Divna KRPAN
University of Split, Faculty of Science
Croatia
[email protected]

MLADENOVI
University of Split, Faculty of Science
Croatia
[email protected]

ABSTRACT
The link between problem-solving and programming skills is well known. Students with higher problem-solving
abilities find programming easy and they can master programming with no or little difficulties regardless of the
programming environment. On the contrary, students with lower problem-solving abilities find programming
difficult to understand and are often unable to master it. The before mentioned groups of students usually make up
two thirds of the entire class, the top and lowest thirds, respectively. What about the "middle third" students? This
is probably the most represented group; those are students who can master programming but with some difficulties.
Visual programming language environments are tools designed to engage all students but maybe the
students would gain the most benefit from that approach. In this paper, we explore the educational and motivational
effect of using Scratch for game-based programming on 5th-grade elementary school students based on their
problem-solving abilities. Results presented here confirm the positive effect of using Scratch as the introductory
programming language for game-based programming , compared to Python as the text-
based programming language.

Keywords: programming, Scratch, problem-solving, elementary school

INTRODUCTION
Programming is difficult, , so why should they learn to program?
Programming, as a part of Computer Science, is also a part of everyday life, so learning programming as soon as
possible should be the educational right of the 21st Century (Maloney, Peppler, B. Kafai, Resnick, & Rusk, 2008).
It is well-known that programming requires higher problem-solving abilities and that by programming problem-
solving abilities can be practised. Students with higher problem-solving abilities can master programming with no
or little difficulties, so motivation is crucial to this group of students. On the other hand, students with low problem-
solving abilities are facing a lot of difficulties during learning programming, and are often unable to master it. We
could facilitate their efforts by choosing an appropriate pedagogical approach. Maybe the most interesting group
- ogramming with some difficulties (Armoni, Meerbaum-Salant, &
Ben-Ari, 2015). Motivation and appropriate pedagogical approach can be crucial for these students. Choosing the
proper programming environment for target age may be the key factor.

Textual programming language approach rely

Python, etc. Those languages require high problem-solving skills and precision in syntax, so many children
perceive programming as difficult. Programming is most commonly taught through solving various math
problems. The question is: do children find programming based on this approach repulsive?
Visual programming languages are syntax error free and more attractive to children. Scratch is visual, block-based,
programming language appropriate for elementary school children (Resnick et al., 2009). Using programming
languages like Scratch has the advantage of eliminating syntax problems which allows students to focus on the
language semantics. Besides that, elementary school students have yet to reach the appropriate level of abstract
thinking required to program, which makes learning programming more difficult. Learning programming by using
a visual programming language can provide concrete to abstract experience (Dann & Cooper, 2009), and can thus
be used as a medium for mediated transfer (Dann, Cosgrove, Slater, & Culyba, 2012) .
By teaching programming in Scratch, we can also shift the context of programming from solving math problems,
which is the most commonly used approach in text-based programming, to programming games, storytelling, etc.
The question is: do students learn programming concepts or is it just fun?

Our previous research showed that 5th-grade students with higher

problem-solving abilities achieved better results -based programming language. When it
comes to Scratch that is not the case. Students with intermediate problem-solving abilities achieved better results
in Scratch than in Python. This led us to new questions, can w

the motivational factor of these languages? In this paper, we give answers to these questions. This research was
conducted in classroom settings. We analysed 5th-grade elementary school students
enrollment in elective course Informatics in two schools by both approaches using Python as textual language and
programming games in Scratch. Students didn't have any programming experience, and their problem-solving
abilities were tested prior to the experiment. Based on the problem-solving test results, students were assigned to
one of three groups. Students were learning different programming concepts in two programming languages and
were tested for both approaches. Test results were analyzed and compared for each problem-solving group. The
results of our research are presented in this paper.

BACKGROUND
Novice programmers find it very difficult to master programming. In order to solve a programming problem, we
first need to break the problem into smaller, more manageable steps. This is the process of developing an algorithm.
Students who need to focus more on syntax commonly ignore this phase. Flow charts and trace tables are often
used to take the students through the steps of structured problem solving which leads to identifying sequence,
selection, and repetition (Whitfield, Blakeway, Herterich, & Beaumont, 2007). Afterwards, novices need to
program these steps using some programming language. A programming language is perceived as a major obstacle
for novices (McCracken et al., 2001), especially when it comes to elementary school children because novice often
focus on programming language syntax rather than developing an algorithm. Visual programming languages may
help with this issue since they have simpler syntax, which allows novices to focus on developing an algorithm
(Grover & Pea, 2013).

It has been reported that students with lower mathematical skills can learn problem-solving and programming if
provided with appropriate materials and the use of less complex visual tools (e.g. Java Trainer) before moving to
IDEs which are considered more complex (Whitfield et al., 2007). It is also reported that children weren't aware
that they were programming, they argued that they were making games, stories, interactive presentations in Scratch
(Maloney et al., 2008). This phenomenon (Dann & Cooper, 2009) (Pausch &
Zaslow, 2008).

In one of our previous research, we compared students success in LOGO, which is a text-based, and Scratch,
which is a block-based programming language. Results showed that students success in Scratch was better,
especially with regard to the concept of a nested loop. There were only a few students who were able to fully
understand basic programming concepts while using Logo, but after a switch to Scratch that number increased.
When it comes to motivation, Scratch is far more positively accepted than Logo
.

Therefore, we can conclude that the can benefit from using a visual programming language.
Similar conclusions were made in other studies. It was demonstrated that the use of the Jeliot program animation
system primarily benefited (Ben-Bassat Levy, Ben-Ari, & Uronen, 2003). The same
conclusion was reported in a study where Scratch was used as an introduction to C# (Armoni et al., 2015). Besides,
students can master basic programming concepts more quickly by using a visual programming language (Armoni
et al., 2015) (Price & Barnes, 2015).

However, there are studies that indicate some possible bad habits of programming in Scratch. It that
during programming in Scratch middle-school students developed bottom-up programming and extremely fine-
grained programming bad habits, although researchers were satisfied by motivation and developed technical skills
of students by programming in Scratch (Meerbaum-Salant, Armoni, & Ben-Ari, 2011). These habits may have
been developed with the idea of Scratch
programming approach. These habits concerning (Gordon, Marron, & Meerbaum-Salant, 2012),
especially when Scratch is used as part of formal learning in a classroom setting in which teachers can guide
students.

Researchers analysed a total of 100 projects, and two other bad habits were discovered. The first one refers to
character naming, where most students didn't change default names like Sprite1, Sprite2. Conversely, it was also
reported that user variables are named correctly, i.e., semantically meaningful. A possible explanation is that when
creating a new character, the name is given automatically which is not the case when adding new variables. The
second bad habit is duplicating code in the same project which indicates that abstraction and modularization were
not taught (Moreno & Robles, 2014). Teachers who use visual programming languages in their classroom need to
be aware of bad habits to minimize their occurrence.

METHODOLOGY
In this study, qualitative and quantitative methods were used.

Research design
This research was conducted with the purpose to compare the basic programming concepts understanding,
concerning two programming languages: Python and Scratch. The target group were elementary school students
with no previous programming experience. In the Republic of Croatia Informatics is an elective course (Ministry
of Science Education and Sports of the Republic of Croatia, 2005) from 5th to 8th grade, and programming is only
one of several main topics in each grade. Therefore, 5th-grade students were appropriate for this research. Since
programming is related to problem-solving skills and there is a positive correlation between math and
programming (White & Sivitanides, 2003), the administered pre-test was designed to test student problem-solving
skills. Students were first exposed to Python programming for four weeks with two hours per week which makes
a total of eight hours. The lectures included selected programming concepts: variables, input, print, sequencing
and conditionals. Student skills in Python programming were tested afterwards. Three weeks later (after winter
school break), we introduced students to programming in Scratch. We have selected a game-based approach and
students were required to program simple games. They were introduced to basic programming concepts like
sequencing, conditional and iteration. The lectures were held for two hours per week for three weeks, and
afterwards student understanding of concepts in Scratch was tested. Additionally, they were given a
questionnaire about their attitude towards programming after learning both Scratch and Python.

The participants of the experiment were students from two elementary schools in Split, in both of which the first
author of this paper was the teacher. Prior to the experiment, the teacher had five years experience of teaching
computer science in elementary schools and four years of experience in teaching computer science at the
undergraduate level.
The research design is shown in Table1.
Table 1: Research design
Experiment

week Topic New terms and instructions New concepts

Pre- test
Introducing algorithm term, basic
algorithms: sequencing,
Algorithms: sequencing, Algorithm, sequencing, conditional conditional and iteration with
1
conditional and iteration and iteration examples from real life.
Introducing to Python
programming language.
Basic Python instructions,
Python

2 Variable, input and output Variable, input, print, int variable term and integers with
examples in Python.
Solving simple problems in
input processing, output
Python program using input,
3 process phases of the Arithmetic operations (+,-,*, /)
processing including basic
computer program
arithmetic operations and output.
Solving simple problems
4 Conditional If else including branching algorithm in
Python using if else.
Python test, questionnaire about programming and python

Three-week Christmas holidays

Aquarium simulation
1 forward, left, right, repeat sprites, concurrency, loops
program
Scratch

2 Chasing ghosts game If, variables conditionals

communication by messaging,
3 Simple ricochet game conditional loops, Coordination and loops with conditionals
Synchronisation
Scratch test, questionnaire about programming and programming languages

The primary goal of this research is to find the -tests and tests following
the Python and Scratch lectures. The second goal is to examine the differences in attitudes towards programming
and the programming language used. Based on that, we defined the next hypotheses:
H1 Students with higher problem-solving abilities will be more successful in Python programming than students
with lower problem-solving abilities.
H2 - Students with higher problem-solving abilities will be more successful in Scratch programming than students
with lower problem-solving abilities.
H3 attitude towards programming will be more positive after Scratch than after Python.

Participants
The research sample consisted of 54 5th grade students from two schools during the school year 2014/2015. Since
programming was taught for seven weeks, some students didn't attend all of the lectures or tests. Hence, the final
number of participants is 50, which includes 34 boys and 16 girls. Students had no previous programming
experience, which means that this is their first contact with programming. Non-probability, purposive sampling
(Cohen, Manion, & Morrison, 2013) was used, because our goal was to target pupils with no previous
programming experience in elementary school.
Assessment instruments
The data was collected in three phases. In the first phase, students were tested for problem-solving abilities before
the programming lectures began. In the second phase, they were introduced to basic programming concepts in
Python like variables, input, print, sequencing and conditionals.
tested using Python assignments. They also filled a short questionnaire about the attitude toward programming.
During the last, third phase, students were learning basic programming concepts like sequencing, conditionals and
iteration while programming games in Scratch. Again, after the third phase, their acquired programming concepts
were tested using Scratch assignments. They also filled a short questionnaire about the attitude toward
programming and programming languages.

Data analysis
Results were analysed by qualitative and quantitative techniques that are used for triangulation purpose (Cohen et
al., 2013) to increase the validity of the findings. Kolmogorov-Smirnov test is used to determine the normality of
data. Parametric independent t-test and non-parametric Mann-Whitney U test are used to compare results between
groups. Parametric test paired t-test and non-parametric Wilcoxon Signed Ranks Test are used to compare student
results in different tests. Non-parametric tests are used for data which meet the requirements for using
parametric tests.

RESULTS AND DISCUSSION

Results are presented in this section.

Problem-solving test
The first test was a problem-solving test that was administered before any programming lectures were held. The
maximum test score was 14 points. Based on the achieved score, students were placed in one of three groups:
stronger, intermediate and weaker students. The test can be seen in our previous paper
2016). Table 2 shows distribution of participants by strength groups.

Table 2: Distribution of participants by strength groups

Group N points Mean SD
Stronger 15 >=11 11,93 1,223
Between 7
Intermediate 16 9,69 0,704
and 11
Weaker 19 <8 4,05 2,97
Post-tests
Two post-tests were conducted in order to assess student achievement. The first post-test was administered
following the conclusion of Python lectures, and the second one following conclusion of Scratch lectures. Since
the number of points in each test was different, we decided to use the percentage as a measure of success.
Kolmogorov-Smirnov test showed that there is a normal distribution of data in both Python (p=0,197) and Scratch
(p=0,069) tests, but not in all groups combined. Table 3 shows descriptive statistics results.

Table 3: Descriptive statistics

Python test Scratch test
Shapiro- Shapiro-
Mean SD Wilk Mean SD Wilk
p p
Stronger 81,667 20,5116 0,006 76,953 19,7003 0,095
Intermediate 61,831 22,5395 0,278 71,575 19,0275 0,253
Weaker 38,910 21,9488 0,504 54,537 17,1179 0,505

Because the results of the Python test didn't satisfy the assumption of normal distribution, the Kruskal-Wallis test
was conducted. The test showed statistically significant difference between strength groups ( 2(2)=19.342,
p=0.000). As it can be seen in Table 3, stronger students achieved, statistically significant, better results compared
to intermediate and weaker.

ANOVA test was conducted to compare group results in Scratch test. There was a statistically significant
difference between groups as determined by one-way ANOVA (F(2,47) = 6.943, p = .002). As ANOVA showed
statistically significant difference, we made further analysis by Man-Whitney U and independent t-test differences
between groups whose results are presented in the following sections.

Comparing success based on problem-solving abilities

In order to compare student success based on programming language used, we used the Mann-Whitney U test. In
the analysis, we considered only two-thirds of the participants, those with intermediate and higher problem-solving
skills. Students with lower problem-solving skills were left out of the analysis. When the Mann-Whitney statistic
was calculated to determine whether there was any statistically significant difference in the Python test scores (U
=62,5, z=-2,288, p=0,022), a statistically significant difference was found between students with higher problem-
solving skills and those with intermediate problem-solving skills. From these results, we conclude that the former
group of students achieved better results than the latter. In the case of Scratch test scores, no statistically significant
difference between groups was found (U =99,5, z=-0,828, p=0,408).

These results indicate that with programming language like Scratch -

programming. This finding is consistent with other have the
most benefit from the use of animations (Ben-Bassat Levy et al., 2003) and visual programming languages like
Scratch (Armoni et al., 2015).

Afterwards, we compared intermediate and weaker students by independent t-test for both test results. Results of
the t-test showed that students from intermediate group achieved statistically significant better results on the
Python (t(33)=3,040, p=0,005), and Scratch test (t(33)=2.788, p=0.009). These results indicate that weaker
regardless of the programing language used.

Based on the result we accept H1 because problem-solving abilities are directly related to success in Python
programming language. But, when it comes to Scratch success this is not the case, at least for intermediate students
Based
on these results we can conclude that students with higher problem-solving skills can master programming
regardless of the programming language or method used. An important finding is that by programming games in
visual programming languages like Scratch we can stimulate the motivation of intermediate students. If we add a
motivational factor,
from programming.

Attitude towards programming

In H3 we assumed that a positive attitude towards programming would be higher after using Scratch compared to
Python. After the lessons about programming in Python, students answered a Likert scale question of 5 items about
their attitude towards programming. This question was repeated in the small questionnaire students answered after
the Scratch lessons. The questionnaire was composed of four Likert scale questions regarding their attitude towards
the programming languages used.

Table 4 shows the questions.

Table 4: Survey questions

Questions
After Python Q1 How much do you like programming?
Q2 How much do you like programming?
Q3 How much do you like Python?
After Scratch
Q4 How much do you like Scratch?
Q5 Which programming language do you prefer?
Figure 1 shows frequencies.

Figure 1: frequencies of the answers

From Figure 1 it is obvious that students liked Scratch more and that attitude towards programming is more positive
after using Scratch. Figure 2 shows results for Q5.

Figure 2: Q5 results

We wanted to compare student attitude towards programming after Python (Q1) and after Scratch (Q2). The
Wilcoxon signed test rank showed statistically significant difference between Q1 and Q2 groups (Z=-2,012,
p=0,044). Students had a greater affiliation for programming after Scratch compared to Python. This confirms that
Scratch had a positive effect on student attitude towards programming. Thus we can accept H3 and conclude that
attitude towards programming is more positive after Scratch than after Python.

However, it came as a surprise that even after being introduced to Scratch, a handful of students still preferred
Python. We assumed that these are the students that belong to the top third of the class with regard to their
problem-solving abilities.

In the questionnaire, students had space to write their thoughts about programming. They wrote 32 comments,
only 4 of them were negative. For example:
this is boring oo hard for me
Some of the comments were neutral:
sometimes is boring and sometimes fun

Most of the comments (22) about programming were positive:

Programming is awesome Programming is cool Programming is great and interesting I like
programming, and I would like to learn it again next school year I like programming because I learned
something new ,

Some comments referring to Scratch:

I like Scratch more than Python ; Scratch is extremely fun, . ; t like
programming too much but I had fun while programming in Scratch ;
chosen Python as favorite programming language because I was on Python programming competition I like
both programming languages, Python and Scratch Scratch is awesome.

Since the first author was the teacher in all classes, we can also confirm the observations (Armoni et al., 2015)
which refer to early recognition of though concepts in a second programming language. Furthermore, we also
observed a shortened teaching process which enables the teacher to assist students with weaker programming
abilities, while those with higher programming abilities could explore new features in Scratch.

CONCLUSION
Programming novices, especially those at the elementary school level need a very gentle introduction to
programming. Students should be able to focus on problem-solving and writing algorithms instead of thinking
about syntax. Visual programming languages, like Scratch, offer the experience of syntax free programming which
is suitable for novices. Furthermore, visual programming languages allow the teacher to shift the teaching context
from solving math problems to programming games. Finally, is improves positive attitude towards programming.
Considering statements above we need to be careful not to forget that the main reason for using visual
programming languages is to focus on teaching programming concepts. Scratch should be a media or a tool used
for transfer of those programming concepts in -based programming languages like Python.

Students with high problem-solving abilities can master programming easily, regardless of the programming
language. On the other hand, students with lower problem-solving abilities encounter significant difficulties while
learning programming, and might be unable to truly master it. These two groups usually constitute two-thirds of
students in a class. The "middle third" students are the ones that we can influence the most. This is a group that is
able to master programming with some difficulties. Based on some previous studies we assumed that the use of
Scratch might boost their motivation, attitudes and achievement. With the teacher's help, Scratch can be used as a
tool for mediated transfer of programming concepts from block-based to text-based programming languages and
can improve the motivation for all students.

There is a lack of empirical research which compares the use of text-based and block-based visual programming
languages in school settings at the K-12 level. We conducted research among 50 5th-graders in two elementary
schools. Students were learning programming in Python, and later in Scratch. Results showed that students with
higher problem-solving abilities were more successful in Python programming than students with lower problem-
solving abilities. This is not the case when it comes to Scratch. In the case of Scratch, there were no differences in
the success between better which proves the usefulness of using Scratch to learn
programming. Most students had more positive attitude towards programming after Scratch than after Python. It
is important to note that students learned Python first, which is more difficult than Scratch. This order of
introducing different programming languages might seem inversed, but we believe that it had a positive influence
on the student's motivation. We believe that their motivation would be smaller if the languages were introduced
vice versa.
View publication stats

REFERENCES
Armoni, M., Meerbaum-Salant, O., & Ben- ACM Transactions on
Computing Education, 14(4), 25:1 25:15. https://doi.org/10.1145/2677087
Ben-Bassat Levy, R., Ben-Ari, M., & Uronen, P. A. (2003). The Jeliot 2000 program animation system. In Computers
and Education (Vol. 40, pp. 1 15). https://doi.org/10.1016/S0360-1315(02)00076-3
Cohen, L., Manion, L., & Morrison, K. (2013). Research methods in education. Routledge.
Dann, W., & Cooper, S. (2009). Alice 3: Concrete to Abstract. Communications of the ACM, 52(8), 27.
https://doi.org/10.1145/1536616.1536628
Proceedings of the 43rd
ACM Technical Symposium on Computer Science Education, 141 146. https://doi.org/10.1145/2157136.2157180
Gordon, M., Marron, A., & Meerbaum-Salant, O. (2012). Spaghetti for the main course?: observations on the naturalness
of scenario-based programming. In Proceedings of the 17th ACM annual conference on Innovation and technology
in computer science education - (p. 198). New York, New York, USA: ACM Press.
https://doi.org/10.1145/2325296.2325346
Grover, S., & Pea, R. (2013). Computational Thinking in K--12 A Review of the State of the Field. Educational
Researcher, 42(1), 38 43.
Maloney, J., Peppler, K., B. Kafai, Y., Resnick, M., & Rusk, N. (2008). Programming by choice: urban youth learning
programming with scratch. ACM SIGCSE Bulletin, 40(1), 367 371. https://doi.org/10.1145/1352322.1352260
-national,
multi-institutional study of assessment of programming skills of first-year CS students. In Working group reports
from ITiCSE on Innovation and technology in computer science education - ITiCSE- (p. 125). New York,
New York, USA: ACM Press. https://doi.org/10.1145/572133.572137
Meerbaum-Salant, O., Armoni, M., & Ben-Ari, M. (2011). Habits of programming in Scratch. In Proceedings of the 16th
annual joint conference on Innovation and technology in computer science education (pp. 168 172).
https://doi.org/10.1145/1999747.1999796
Ministry of science education and Sports of the Republic of Croatia. (2005). The curriculum for primary school. Zagreb.

STUDENTS NOVICES BY USING GAME DEVELOPMENT IN PYTHON AND SCRATCH. In

EDULEARN16 Proceedings (pp. 1622 1629). IATED. https://doi.org/10.21125/edulearn.2016.1323

Programming Environment. International Journal of Modern Education and Computer Science, 8(August), 1 10.
https://doi.org/10.5815/ijmecs.2016.08.01
Moreno, J., & Robles, G. (2014). Automatic detection of bad programming habits in scratch: A preliminary study. In
2014 IEEE Frontiers in Education Conference (FIE) Proceedings (pp. 1 4). IEEE.
https://doi.org/10.1109/FIE.2014.7044055
Pausch, R., & Zaslow, J. (2008). Last Lecture. Statistics, 7, 1 18. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/19064375
Price, T. W., & Barnes, T. (2015). Comparing Textual and Block Interfaces in a Novice Programming Environment. In
Proceedings of the eleventh annual International Conference on International Computing Education Research -
(pp. 91 99). New York, New York, USA: ACM Press. https://doi.org/10.1145/2787622.2787712
Resnick, M., Maloney, J., Monroy-
Programming for All. Commun. ACM, 52(11), 60 67. https://doi.org/10.1145/1592761.1592779
White, G., & Sivitanides, M. (2003). An Empirical Investigation of the Relationship Between Success in Mathematics
and Visual Programming Courses. Journal of Information Systems Education, 14(4).
Whitfield, A. K., Blakeway, S., Herterich, G. E., & Beaumont, C. (2007). Programming, disciplines and methods adopted
at Liverpool Hope University. Innovation in Teaching and Learning in Information and Computer Sciences, 6(4),
145 168.