Research Status in Computational Thinking in STEM Education

Abstract

Computational thinking (CT) is an approach to problem-solving that has its roots in computer science. However, its inherent value in the science, technology, engineering, and mathematics (STEM) disciplines cannot be over-emphasized, considering that we are in the fourth industrial revolution. The chapter draws attention to its close affinity to problem-solving and programming, and the impact of computational thinking on the labour market, and in turn the digital economy is highlighted. A global overview of recent research findings and initiatives to implement CT education in school curricula are discussed. Because of the importance of STEM education, and the inherent value of CT, it is necessary to explore the status and inclinations of CT in STEM disciplines. Hence, a snapshot of research over the last two years was used in a systematic review to determine the trends and challenges for integrating CT in the curriculum of STEM related fields. Using the ERIC database of journals, and specific criteria for selection of publications, 31 articles were examined in this study. Overall, it was found several tools and instructional strategies are used to develop CT, but more needs to be done to increase teachers’ knowledge and enactment for CT in the STEM fields.

1. Introduction

Computational Thinking is fundamental for many, if not all occupations, particularly science, technology, engineering, and mathematics (STEM). STEM related fields play a significant role in economies by driving innovation to meet the demands of the fourth industrial revolution era. However, STEM fields of study are often perceived as difficult and many students drop out of these subject areas as a result, impeding career opportunities in the related fields [1]. Accordingly, institutions of higher learning play a crucial role in preparing the people for STEM employment to meet the exigences of the 4^th industrial revolution. While the acronym, STEM, was coined as a general and appropriate word for Science, Technology, Engineering and Mathematics fields of study, STEM often relate to all sciences (astronomy, physics, computing fields and the like). These fields often depend on computational tools for modeling and simulation, data analysis and visualization, creating computer programs to solve problems, and understanding a system as an aggregate of parts; these are characteristics of computational thinking. Hence, it can be inferred that Computational thinking (CT) is inherent to STEM practices [2]. Moreover, computational thinking is widely considered to be an important and necessary twenty-first century skill that contributes to the development of solving complex problems. As a result, a growing body of literature has investigated the tools and interventions to foster and develop CT in education [3, 4, 5]. Overall, such studies highlight the need for a consistent outcome of the interventions implemented.

Moreover, with the growing importance of STEM education world-wide [6], it is not surprising for the need to foreground research, not just in general STEM education, but in the integration of computational thinking in the STEM fields. The need for CT to be enhanced and fostered in STEM fields is imperative. Therefore, this chapter seeks to assess the research status and inclinations of computational thinking in STEM fields and determines the key findings for future implications.

The rest of the chapter proceeds as follows: a) Literature review highlighting the key concepts of STEM, CT and its close affinity to programming and AI, b) Methodology), c) Results, d) Discussion of findings, and d) Conclusion. This chapter contextualises the research by providing background literature on CT, programming and AI, and the global status of CT education. The chapter then discusses the specific methods by which the research and analyses were conducted, followed by a discussion and conclusion section.

2. Literature review

2.1 What is computational thinking

It is important to re-iterate that while computational thinking has its roots in computer science (CS), it is not computer thinking and reasoning, or programming either. Though there is minimal consensus on the definition of computational thinking, there is agreement in the literature that CT involves decomposition, abstraction, pattern recognition, and algorithmic thinking, which when expanded entails the following:

Decomposition: Splitting a composite problem or system into smaller components and solving each component and then logically organizing and analyzing data and making deductions.

Abstraction: Managing complexity so that the complicated and difficult aspects can be put aside into a black box so others can work on the details of it, focusing on the relevant information only, while temporarily ignoring the detail in the black box. Abstraction is at the heart of CT.

Pattern recognition: Looking at how people have solved similar problems drawing on that experience and identify similarities among and within problems.

Algorithmic thinking: Formulating a set of steps to achieve the objective, i.e., a set of instructions.

Generalizing this problem-solving process: Translating trends and patterns into rules, principles, or insights to apply to a wide variety of problems.

After igniting the importance of computational thinking advocacy in 2006, ten years later Jeanette Wing [7] defined CT as

“… the thought processes involved in formulating problems and their solutions so that the solutions are represented in a form that can be effectively carried out by an information-processing agent” (p. 8).

On examining this definition, two aspects emerge that are of importance for STEM education.

Computational Thinking (CT) is a thinking procedure; therefore, CT can be independent from the use of technology. Supporting this view, Sysło [8] writes that CT “is a collection of key mental tools and practices originated in computing but addressed to all areas far beyond computer science” (p. 1). Furthermore, there is strong evidence that unplugged approaches can be just as effective, if not better in advancing computational thinking skills and thereby facilitating students’ ability to program [3]. For example, in their study [9] found that when comparing plugged to unplugged (non-computer use) approaches as students learned programming, those who learned using unplugged approaches were more confident in understanding the concepts than those who used the plugged approach. Hence, it can be inferred that CT can be used in other STEM fields, where the use of computers is not required. It is, therefore, not surprising that [4] in an earlier study described computational thinking as, “… a specific type of problem-solving that entails distinct abilities, e.g., being able to design solutions that can be executed by a computer, a human, or a combination of both”. Thus, to the non-specialist in computing, good sense suggests that computational thinking may be explained as a way of reasoning and of solving problems in a modern-day world characterized by up-to-date technology.

2.2 Importance of computational thinking (CT)

While the genesis of computational thinking can be traced back to Papert [5] for his work in Logo programming, it was only when Wing [10], the former Vice President of Microsoft Research, published her seminal paper on CT, did research in CT begin to gain momentum. As we have been ushered into the 4th Industrial revolution, this increased attention to CT could not have been timelier. In the last decade, there has been a surge of interest in the effects of computational thinking. While teaching children thinking is a key competence that education should inculcate, developing CT has become even more crucial in this era of digital economy.

The advancement of digital technology has heightened the use of computational thinking and this trend is growing across all industries, which certainly has implications for our students and the labour market. This situation in turn should impact our education and STEM curriculum.

Hence, solving unusual problems in the current era of digital technology is an important competence. Our current students live a life greatly predisposed to information technology (IT), and many will work in areas that involve or are impacted by IT. The unprecedented advancement of technology, and its various forms of communication via the Internet, have not only permeated our lives in many respects, but is hugely impacting the digital economy. To name a few examples; in healthcare – operating rooms’ efficiency depends on computing, and it has enabled advances and inventions such as contact lenses that detect levels of insulin in people with diabetes; in space – there is a move to develop and use a generation of robots to explore where humans cannot now; in households – people have begun to automate every-day phenomena like the heating and lighting systems, and the use of robots to clean floors and carpets; on our road travels – we depend on navigation systems to get us to our various destinations, and now manufacturers are experimenting with self-driving cars. Hence, computational thinking has been recognised as a twenty-first century skill [6, 11].

Wing [7] in a later study asserts that:

“Everyone should be given the opportunity to gain competences in the field of computational thinking to allow them to successfully participate in a digitalized world”.

This excerpt gives importance to the claim that computational thinking is fundamental, not only for computer science but for all sciences in agreement with [12]. This claim raises new charges and challenges for schools. Furthermore, the increasing rate of technology users or consumers far exceeds that of creators or digital innovators disproportionately. This imbalance can have some adverse effects on the labour market and consequently the digital economy. How do we close this disproportionate gap? This situation, therefore, calls urgent attention to the development of computational thinking generally, and more specifically in STEM fields. While much research has been generated in computational thinking and in STEM education [1], comparatively there are limited studies on the integration of computational thinking and its use in STEM fields. The question therefore arises: how do we know whether our students really possess this skill to meet the twenty-first century skills set? Moreover, with the rise of artificial intelligence awareness, tools, and applications, computational thinking is even more crucial in this era. Several studies advocate programming and AI for all, inferring that coding, or programming is needed in most STEM related fields. What follows, is a discussion of the link between CT, programming, and AI.

2.3 Computational thinking, programming and artificial intelligence

Computing is at its heart a creative subject. The best way to learn is to make something by getting involved. Children learn by playing and experimenting with technology, in keeping with Seymour Papert’s theory of learning—Constructionism. Several efforts to develop students’ CT has tended to use activities or tools that are directly linked to programming skills in educational settings (e.g. [13, 14, 15]). However, this close kinship to computer science or IT, does not automatically imply that computational thinking is exclusively the domain of computer science or programming. Still, it is not injudicious to consider the relationship between CT and programming.

As mentioned earlier, one of the aspects of CT is algorithmic thinking. An algorithm is an unambiguous defined step by step guide to solve a problem or achieve a particular objective. Programming, however, may be broadly regarded as a two-step process – first, a set of steps to solve a problem, its algorithm and second, the coding of the algorithm into a specific system. How one solves the problem to achieve the solution involves the creative thinking. One then has the task of making those thoughts come into action by translating that algorithm into a set of symbols according to the computation system one is using, i.e., the language. Both are hard and both require creativity.

So, the best way to learn is by doing and the practical experience of programming [is] almost certainly the best way for primary school pupils to learn about computing and Computational thinking. This has been observed anecdotally and from the literature that is beginning to show evidence of this trend.

More recently, Computational Thinking (CT) and programming skills are now deemed as being as essential as numeracy and literacy by many scholars [16]. In short, getting computers to help us to solve problems is a two-step process. First, we think about the steps to solve a problem or the rules that govern a system. Second, we use our technical skills to get the computer working on the problem. Computational Thinking is the first of these. It describes the concepts, processes, and approaches we draw on, when thinking about problems or systems in a way that a computer can help us with these. It is really that first stage that we ought to be concentrating on and it is very much of what we do in computing. For example, if one is going to write up something – one thinks of the idea, structure, the content, etc. before one starts to type it up.

Having discussed the association between CT and programming, I will now move on to discuss their association to AI, a technological revolution in terms of innovation.

Heintz [17], a specialist in AI, indicated in his talk that “Computational thinking develops techniques for people to solve problems in a way that allows computers to help. Artificial intelligence develops techniques for computers to solve problems.” CT captures what we need to be good at to leverage all the artificial intelligence (AI) and other computational tools that are available.

What is interesting here is that as we learn more about AI and as these tools are developed based on for example machine learning we need to be able to leverage them through our CT skills – by being able to understand how the computational processes work and how we can benefit from those when we solve problems. There are several interesting cases where people use machine learning or other tools as part of their own problem-solving process.

Artificial intelligence platforms involve the use of machines to mimic human reasoning. In an attempt to mimic human cognition, these platforms model human reasoning, problem-solving and intelligence, both social and general of which computational thinking is part.

Referring to Papert’s [5] paper, he said “…technology is something children themselves will learn to manipulate, extend, to apply to projects, thereby gaining a greater and more articulate mastery of the world, a sense of the power of applied knowledge” (p 353). Using the power of computing to make a meaningful social impact, children are empowered to make an impact by doing activities situated in context that are personally relevant. Hence, in our world of AI, computational thinking skills should be a core competence for all students.

Returning to the subject of computational thinking, in the section that follows, I review and summarize the global status of CT education.

2.4 Global status of computational thinking (CT) education

There has been much research on interventions to include CT in the curriculum – these interventions invariably involve some aspects of programming. However, in his paper, Yadav [18] foregrounds how CT nudges students past operational and technical skills, creating problem solvers instead of consumers of software and technology. In a later study, [19] assert that because of its capacity for automation and enactment, programming appears to be a natural vehicle for learning computational thinking, but with some amendments in the approach to focus on CT.

A noteworthy finding in Taslibeyaz, Kursun, and Karaman’s [20] study is that the development of CT skills is predominantly examined with programming content tools, such as Scratch and robotics for school level students. Additionally, it was found that studies on the development of CT for university students were carried out in other content areas besides programming education. While some studies indicated that teaching CT does improve programming education, others have found that programming courses develop CT.

Ministries of education in many countries have recognized the importance of computational thinking for their economy [16]. In a joint report compiled by JCR, it was envisaged that both CT and programming are key competencies for compulsory education. This report indicates the countries that have recognised the importance of CT and its implementation in the school curricular. While Table 1 is not meant to be exhaustive or comprehensive, it provides an overview of the countries that have implemented CT as compulsory education in the school curriculum at the time of writing this chapter. The piloting and revision of the implementation of CT has been done in the years prior to what is indicated.

Country (not meant to be exhaustive)	Implement Year	Current state of new policy initiatives
Asian countries (Taiwan, Japan, and Korea)	2020	For primary in 2020 and -2022 for secondary
Australia	2018	Digital technologies compulsory with CT education
European countries (17 countries—Austria, Czech Rep Finland, Denmark, France, Greece, Hungary, Ireland, Italy Lithuania, Norway, Poland, Portugal, Russia, Sweden, Switzerland, Turkey)	2018	Denmark and Norway introducing CT education as a permanent elective.; Piloting began in 2016
UK (England, Scotland, Netherlands, Wales)	2018	England is the first country to make CT education compulsory in 2014
USA	2018	By 2018 all 50 states have policy for CT in schools
Hong Kong	2017	CT supplement introduced to Primary from P4 to P6
New Zealand	2017	The digital curriculum was reformed to include CT
Singapore	2014	CT implemented as optional in different education sectors—called smart Nation initiative
Ghana		Part of a code club – commitment to implement CT- no formal policy
Nigeria		no formal policy- focus on enrichment
India		Advocacy phase- CSpathshala initiative
South Africa		A Pilot was planned for 2021.

Table 1.

Global status of CT education.

England (part of UK) was one of the first countries to include CT as part of the mandatory course in the school curriculum as early as 2014. Interestingly, by 2018 all 50 states in the US had implemented CT education in the school curriculum as policy. In countries like Ghana and Nigeria – there is commitment to include CT as a mandatory aspect –but it has not yet become policy. In India – in the rural areas Computer science (CS) has become compulsory – they have worked with CSpathshala, an ACM India initiative to bring a computing curriculum to Indian schools in 11 different states. At the time of writing this chapter, it was determined that they are teaching it in the mathematic curriculum, while there is a huge drive to include all schools.

While South Africa has the commitment to implement CT in the school curricula – they have not yet successfully piloted the implementation, which was due to have started in 2021 in 1000 schools. It is necessary to note the huge challenges of infrastructure and teacher development knowledge that exists. It is envisaged that universities as stakeholders for preparing students for the workplace, have a role to play in addressing these challenges.

In their report, Bocconi, et al. [4] review the ministries of education of several countries regarding the status of CT in the school curriculum. In attempting to embed CT, CS, or coding in the school curriculum, Figure 1 summarizes the rationale for including CT in the curricula across the globe. It was found that most countries introduced CT in mainstream education at secondary level [4]. However, an emerging trend indicates integration of CT in primary school levels.

Figure 1.

Rationale for integrating CT in the curriculum (adapted from [16]).

What is important to recognize here, is that fostering logical thinking and problem-solving skills are the two most common reasons for including CT in the curriculum, followed by other competences and programming. From the literature, it can be summarized that the three main reasons for including CT in compulsory education are: developing CT to increase economic growth, occupy ICT related vacancies, and groom for future work or occupations [9].

To further determine the status of Computational thinking related to studies in STEM education – the fields that drive the economy, a systematic review methodology was employed to ascertain the trends of these studies. The details of the review are presented in the Methodology section. I synthesized this review with existing ones to inform the global status of CT in STEM education and the empirical evidence of CT development.

3. Methodology

Following the PRISMA model (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) [21], a systematic literature review was conducted. In the current study, the following steps were taken to achieve the relevant set of articles for analysis.

1. Specifying the research questions to guide the search.

2. Determining the database(s) for the search procedure

3. Specifying the inclusion and exclusion criteria

4. Choosing the studies

5. Conducting a pre-analysis and extraction of data

3.1 The research questions

To guide this review research the following three research questions were formulated:

• What is the status of CT in STEM education research from 2020 to the end of 2021 based on refereed journal publications?

• What research methods did authors tend to use in conducting CT in STEM education research?

• What key themes had emerged in CT in STEM education research based on the journal publications?

3.2 Determining the database(s) for the relevant studies

The review focused on publications that appeared beyond specific discipline-based journals. Due to the plethora of studies generated in the last few years on the specific aspects of CT and STEM separately, it was decided to examine the most recent research on this exciting and relevant field during the pandemic, CT in STEM fields. It was therefore prudent to consider a snapshot of studies that was extracted from January 2020 until December 2021.

It was assumed that articles on CT in STEM education have been published in journals that involve more than one conventional discipline. Since there are too many conventional discipline-based education journals, journals were not selected but emerged as part of the search results. Since this review is embedded in the education context, it was considered viable to look at the ERIC (Education Resources Information Center) database of journals. The ERIC database is a comprehensive database with information and studies in all disciplines related to education, consisting of several education journals. Using the EBSCOhost Research Databases interface, the advanced search option was selected to include the database: ERIC. The automation search strategy provided, enabled me to expand the search and to apply limitations to suit the study. The results of the search yielded source types from 76 academic journals and 76 reports.

3.3 Exclusion and inclusion criteria used

It was necessary to select studies that would be appropriate to the goal of this review, based on the review questions. The following criteria were used in selecting the studies:

3.3.1 Inclusion criteria

• Studies must be empirical (quantitative, qualitative, or mixed methods) in an educational environment.

• Any STEM related subjects studied with CT

• The search was expanded to apply related words and equivalent subjects.

• The article is a peer-reviewed study

• The momentum of studies generated regarding CT and studies regarding STEM have increased exponentially. Hence, the scoped articles were from 2020 until 2021.

3.3.2 Exclusion criteria

• Studies written in a non-English language.

• Studies only published as an abstract

• Conference papers—Since journal publications are acknowledged as one of the quality sources of research ideas and outputs (e.g., [22, 23]), articles published only in journals were considered, all other publications, including conference proceedings and grey matter were excluded.

3.4 Identifying articles

As was pointed out in the introduction to this paper, the acronym STEM relates to all sciences relating to the core subjects of science, technology, engineering, and Mathematics, which are being recognized as a global interdisciplinary field for our students to learn. Moreover, computational thinking has also been recognised as a twenty-first century skill in the current 4th industrial revolution. Using the phrase “computational thinking” together with STEM related fields as identifiers following the methodology of other researchers [24, 25], a set of relevant research articles were obtained. Specifically, the Boolean phrase used was “Computational Thinking AND (mathematics or computer science or engineering or technology or STEM)”.

3.5 Pre-analysis and extraction of data

Additional criteria were imposed in the present literature research: the article abstracts were screened for empirical interventions and outcomes related to CT.

Based on relevance to the research questions, 30 articles were excluded. The remaining 46 were scoped for further information. Studies that did not include the specific interventions to develop CT in any STEM field education were excluded from the review. The remaining 31 articles were then examined in detail in relevance to the criteria and research questions. Each article was read twice to note and understand the content, procedures, and methods used, and outcomes. The PRISMA process that I followed is depicted in Figure 2.

Figure 2.

Process used to obtain studies.

The 31 articles that composed the final dataset were included in the systematic review as shown in Table 2. The studies were examined for contexts, content areas or interventions, variables, and their relationships with each other.

SN	Author(s) and year	Context/subject	level	Approach
1	Hébert and Jenson (2020) [26]	science	Secondary	qualitative
2	Lyon and Magana (2021) [27]	Engineering	Undergraduate/tertiary	qualitative
3	Ardito, Czerkawski and Scollins (2020) [28]	Programming/engineering	Primary	Qualitative
4	Zha, Jin and Moore (2020) [29]	Programming	Pre-service teachers/tertiary	mixed method
5	Kynigos and Grizioti (2020) [30]	programming/gaming	Secondary	qualitative
6	Hunsaker and West (2020) [31]	Interdisciplinary project—CT and coding	Preservice/tertiary	qualitative
7	Deniz, Kaya and Yesilyurt (2021) [32]	integrated stem	Primary/secondary	qualitative
8	Lapawi and Husnin (2020) [33]	Science module	Secondary	quantitative
9	Kukul and Çakir (2020) [34]	programming	primary, undergraduate /tertiary	qualitative
10	Çevik, Baris and Sirin (2021) [35]	Science		mixed method
11	Ilic (2021) [36]	Technologies course	pre-service teachers/tertiary	mixed method
12	Pürbudak and Usta (2021) [37]	Information Technology	Primary	quantitative
13	Ntourou, Kalogiannakis and Psycharis (2021) [38]	Science	Primary	quantitative
14	Yildiz and Seferoglu (2021) [39]	programming	Secondary	quantitative
15	Usengül and Bahçeci (2020) [40]	programming /science	Primary	quantitative
16	Avcu and Er (2020) [41]	programming	Primary	quantitative
17	Emara, Hutchins and Grover (2021) [42]	science, computing	Secondary	qualitative
18	Eryilmaz and Deniz (2021) [43]	programming	Secondary	mixed method
19	Karakasis and Xinogalos (2020) [44]	programming	Teachers	qualitative
20	Türker and Pala (2020) [45]	Programming-Computer education	pre-service/tertiary	mixed method
21	Min and Kim (2020) [46]	Computing software	Primary	qualitative
22	Threekunprapam and Yasri (2020) [47]	programming	Secondary	qualitative
23	Tsakeni (2021) [48]	science	pre-service teachers/tertiary	qualitative
24	Delal and Oner (2020) [49]	Computing software	Primary/secondary	qualitative
25	Chongo, Osmanand Nayan (2021) [50]	Chemistry	Secondary	quantitative
26	Hijón Neira, García-Iruela and Connolly (2021) [51]	programming	Secondary	qualitative
27	Robertson, Gray, Martin and Booth (2020) [52]	programming	Primary	qualitative
28	Kopcha, Ocak, and Qian (2021) [53]		Primary	qualitative
29	Noh and Lee (2020) [54]	programming	Primary	mixed method
30	Avcu and Ayverdi (2020) [55]		Secondary	quantitative
31	Chongo, Osman and Nayan (2020) [56]	STEM	Secondary	quantitative

Table 2.

Articles included in the review.

3.6 Data analysis

To address the research questions, first the keywords were examined to determine the most common keywords and its importance to the research studies. Other descriptive statistics were determined. To address research question 2, all 31 identified publications were examined for the approaches used (1) qualitative, (2) quantitative, (3) mixed methods, and (4) non-empirical studies (including theoretical or conceptual papers, and literature reviews).

Thereafter, the themes related to the interventions or tools were identified and used in the review of publications identifying intervention tools and its relationship to the variables identified.

4. Results

4.1 Initial findings and descriptive analyses

To obtain an overview of the main themes researched, a word cloud of keywords of the 31 articles was performed. The top five keywords that emerged were computational thinking, programming, coding, learning skills, and education (Figure 3). These initial findings inferred the main aspects of CT in STEM are the education of CT and its relation to programming and coding.

Figure 3.

Keywords related to all the articles under study.

Regarding the participants in the studies reviewed, it can be seen from Table 2, that there were more studies undertaken with primary (14) and secondary (14) students than tertiary (6) students or teachers in service (1). In some cases, the participants covered all three groups of participants, giving a sum of studies more than the total number of articles reviewed. This may be related to the need to develop school pupils’ CT skills early before joining the job market or entering institutions of higher learning. Regarding university students (tertiary), they have been involved more as research subjects in STEM subjects and most studies involved teacher education students to determine their knowledge of CT to develop their pedagogical practices. These studies focused on educational implementation of computational thinking that was central to the study. Considering Table 2, it shows that the studies were carried out in the most common context, programming (F=15) and in all other STEM related fields (F= 16).

4.2 Approach used in the data Analyses

This section examines the approaches used to analyze data in the 31 studies that were included in the systematic review. Most of the studies were found to use a qualitative approach (16), some utilized a quantitative approach (8), and others used a mixed method approach (7). Where studies used some form of intervention to teach CT, a qualitative approach appeared to be appropriate to obtain rich data and understanding of the nuances that emerged. Where quantitative analyses were used, it was generally regarding the perceptions or a scale of the self-efficacies of the main aspects of CT or programming that were needed. The mixed method approach used in some studies allowed for a better understanding of their findings by confirming one set of results with another, especially when design-based or evaluative research was used.

4.3 Interventions and outcomes

To understand the status of CT implementation in teaching, the contexts of interventions and or outcomes in teaching CT in STEM fields were examined. Table 3 shows the interventions and outcomes of the studies.

SN	Refs.	Interventions	Outcome
1	[26]	Making and maker spaces for making projects supports STEM and developed CT	Positive effect
2	[27]	design of model-building -Throughout the building of the model, students exhibited the use of computational thinking, mainly abstraction, algorithmic thinking, evaluation, generalization, and decomposition.	Positive effect
3	[28]	Findings suggest that this process is a gendered one, with the boys focused more on the operational aspects of building and coding their robots while the girls focused more on group dynamics. Lego robotics	Different effects on girls and boys- using Lego robotics
4	[29]	Organisation of the technology and instructional methods, such as team-based learning, flipped classroom, and pair programming, to help develop CT using Blocked based Hopscotch	Positive effect
5	[30]	Using modifying games with ChoiCo—elements of context-aware integrated CT connecting otherwise fragmented areas such as databases, block-based programming, GIS design.	Positive effect of intervention on CT
6	[31]	The badges, tutorials and some related resources were compiled into the Tech with Kids web to understand CT.	Positive effect
7	[32]	Used computational thinking to build animatronic zoo with coding. Used engineering design as well as coding	CT enabled the design and building
8	[33]	3-week instruction using the science module that had embedded use of CT skills to teach Science.	CT improved the science achievement
9	[34]	game programming activities used to scaffold students to learn CT skills. This intervention contributed positively to students' CT skills	game programming has a positive effect on students' CT skills
10	[35]	web 2.0 tools used for digital activities -a significant increase in the participants' technology awareness and computational thinking	Positive but weak effect on the intended skill. Mainly due to covid
11	[36]	The applications conducted in the Instructional Technologies course and pre-service teachers stated that Scratch applications contributed to the acquisition of Computational Thinking-using Scratch	Significant, positive correlation between CT and academic achievement
12	[37]	learning styles of Web 2.0 based collaborative group activities was used to examine the effects on academic achievement, online cooperative learning attitude level, computer thinking skill level	web 2.0 learning style increased CT
13	[38]	the use of Arduino and Scratch for Arduino (S4A),to study their effect on self-efficacy and motivation towards Science Education, Computational Thinking (CT) and about electricity	Positive effect in view of the conceptual understanding of electricity and CT
14	[39]	To determine the effect of coding instruction performed with the Lego Mindstorms EV3 robotic set on students' attitudes towards coding and their perceptions of computational thinking skills self-efficacy.	positive attitudes towards coding – a significant POSITIVE change in CT skills and self-efficacy perceptions
15	[40]	Attitudes, academic achievements and computational thinking skills of the experimental group students, who received robotic-assisted science education, toward science course differed significantly compared to the students in the control group LEGOWeDo2.0	Positive attitudes, academic achievements, and computational thinking skills with robotic-assisted science education
16	[41]	develop an instructional design that focuses on programming teaching for gifted and talented students and to investigate its effects on the teaching process.	Positive effect the instructional design was effective on CT and creative thinking skills
17	[42]	The open-ended, problem-solving nature of the task requires groups to grapple with the combination of two domains (science and computing) as they collaboratively construct computational models.	CT challenges afford opportunities for students to explore resource-intensive processes, -trial and error, debugging model errors -positive effect
18	[43]	To determine the effect of Tinkercad use in computer programming education on students ' CT skills and perceptions. they were highly motivated for interest and appreciation and found Tinkercad to be generally useful and easy to use	A positive moderate-level relationship between their perception of Tinkercad and their CT skills.
19	[44]	BlocklyScript an EG aims to help students develop their CT by learning basic programming concepts, designing algorithms and correcting mistakes. During the designing phase different EGs were taken under consideration.	The positive results of this pilot evaluation show that BlocklyScript is expected to help students understand CT
20	[45]	the effect of algorithm education on pre-service teachers' computational thinking skills, and computer programming self-efficacy perceptions were examined.10 different algorithmic problems were presented each week, and they were asked to solve these problems using flow chart	There was no effect on CT in general algorithm education had a positive and significant effect only on students' algorithmic thinking
21	[46]	Designed and applied physical computing lessons for elementary 6th-grade students based on the software education guidelines. supported the active interaction of the digital world and the physical world by constructing a physical model using specific media and controlling it with a program.	physical computing lessons materialize students' computational concepts through computational practices, and improve their computational perspectives through the use of authentic contexts
22	[47]	Developed unplugged coding activities using flowcharts for high school students to learn computer science concepts, and to promote their CT skills.	self-directed learning approach used unplugged activities to promote CT
23	[48]	Explored how preservice science teachers used computational thinking as a problem-solving strategy when facilitating IBPW in multiple-deprived classrooms.	positive effect – using CT, they solved problems that otherwise they could not
24	[49]	Examined the role of using unplugged computing activities (based on the Bebras) challenge on developing computational thinking (CT) skills, to promote CT and informatics among school students of all ages.	Students' post-test scores were significantly higher than their pre-test scores
25	[50]	Study aimed to identify the effectiveness of the Chemistry Computational Thinking (CT-CHEM) Module on achievement in chemistry.	Combination of unplugged and plugged-in activities is more effective CT improves achievement in Chemistry
26	[51]	Incorporating a visual execution environment (VEE) and Scratch project for secondary school students as a method to teach and assess computational thinking. Scratch	Knowledge gain on computational and programming concepts and translate CT experiences into reality.
27	[52]	Programming and Debugging—correlation with teacher's rating of executive functions—(EF) is an umbrella term for higher order cognitive functions linked with the frontal lobes of the human brain.	Cognition of CT correlates with programming and debugging activities positively.
28	[53]	Exploring how the CT of two fifth grade learners emerged as an embodied phenomenon during an educational robotics activity.	Robotics activity and embodiment of math concepts, CT emerged
29	[54]	Course in programming a robot for elementary school students and investigated its effectiveness by implementing it in actual classes.	Significantly improved CT thinking and creativity
30	[55]	Examined the correlation between the computer programming self-efficacy and computational thinking skills of students.	Positive effect
31	[56]	The relationship between CT skills and mathematics achievement was statistically significant.	Mathematical logic improves CT skills -positive effect

Table 3.

Interventions and outcomes of the reviewed articles.

While many studies associated CT with programming, other STEM contexts were investigated to develop CT. As can be seen in Table 4, almost half of the studies reviewed were conducted in a programming context, while the rest of the studies were conducted in non-programming contexts, but within the STEM fields of specialisation.

Context	Intervention/Tools	Frequency
Programming Context	Operational aspects of building and coding their robots; Orchestration of the Technology and instructional methods, such as team-based learning, flipped classroom, and pair programming; Lego (WedO2.0, Mindstorms, robotics); Robotics; Block programming- Hopscotch, Scratch; Educational games (BlockyScript); Creative programming and debugging	15
Non-programming Context	The badges, tutorials and some related resources were compiled into the Tech with Kids web; Integrated STEM; Mathematical logic; Engineering; Science; Computing Software; Chemistry	16

Table 4.

Summary of the context of the studies.

What is interesting is that CT was developed using unplugged activities in some studies (2) successfully, which has implications for schools that do not have the computing infrastructure.

Studies that involved students from schools, used a variety of tools and interventions to teach programming with a view to develop CT skills. In most studies this was found to be positive. The following interventions were used: robots, Lego (Wedo2.0, Mindstorms, robotics), Robotics (2), Hopscotch, Scratch, Andruino Scratch, Tinkercard, Educational games (Blocklyscript), and Digital game modding (ChoiCho), as can been seen in Figure 3. The outcomes of these studies showed a positive effect on CT development.

In addition to these tools, the studies reviewed indicated the teaching strategies used to develop CT directly in STEM contexts. These interventions are shown on the left of Figure 3 as independent concepts, namely, Engineering design, Science module, Instructional design, constructing physical model, Web 2.0 tools for digital activities, Unplugged Activities (2) and plugged-in (1), Making projects, Game programming, Design-based learning, Mathematic logic, embodied interaction with technology, Computational modelling, and Tech with Kids web. While it was determined that unplugged activities are able to develop CT [47, 49] effectively, in another study, unplugged activities coupled with plugged-in activities was found to be more effective in improving CT [50].

The claim for introducing CT in compulsory education is based on the notion of transferability of cognitive skills (e.g., logical reasoning and problem-solving abilities). In reviewing the studies, an important aspect in CT advancement is development tools, which concentrates on concepts that assist in understanding CT developmental process, like [20]’s study. Following their approach, the associations between content development tools and concepts, referred to in this article as dependent and independent variables were determined. The findings are presented in Figure 3. The main themes (variables) that emerged from the review were computational thinking (CT), programming skills, problem solving, and learning. While some of these studies only considered CT skills, others focussed on the concepts (variables) associated with CT, such as problem-solving, programming, thinking.

To understand the process of CT skill development, the variables before and after the interventions were considered as well as the relationships between the variables. The dependent and independent variables affecting CT were obtained from the examined studies, separated into themes by content analysis, and then the variable groups and their relationships were determined. The relationships between the variables are shown in Figure 3.

The studies addressed dependent variables, such as CT skills, problem solving, and programming skills to measure CT skills. Most of these interventions were based on CT skills and problem-solving variables. However, as shown in Figure 4, CT was the dependent variable most frequently studied, followed by programming skills, and problem-solving. Non-programming independent variables, such as the use of digital making and Mathematics learning, games, and Competitive tactile game were mostly used in the studies which included problem-solving skills, as well as CT. The studies that included computer programming as an active independent variable examined the effects on all dependent variables (CT skills, problem solving, and programming skills). Similar, to computer programming, the use of robotics was also associated with most dependent variables. In all studies reviewed, where either robotics or computer programming were used as the intervention tool to promote CT, it was shown to have had a positive effect on CT skills.

Figure 4.

Relationship of dependent and independent concepts related to CT intervention.

In this review, only one study [45], reported no effect on CT skills. This may be related to the duration of the training or intervention, student interest, or the quality of the course. While most studies examined the effect of the intervention on CT, there were three studies [27, 38, 50] that used CT as part of the teaching strategy to determine its effect of students’ performance in the related subjects, such as engineering, science, and chemistry. In all three case cases, the outcome was positive.

5. Discussion of results

Several reports have shown the need for the development of computational skills among our current students. They further determined the status of implementing CT in the school curricular. Many studies have shown that coupled with CT, coding or programming or computer science were introduced as compulsory components in the school curriculum at different levels. However, Table 1, indicates that while CT is the identifying driver, the context in which it was introduced is the programming or computer science context. From the current review of studies, it was determined that Computational thinking skills can be used in many disciplines, specifically STEM disciplines and is beneficial to all students studying in any field.

Whether it is through an individualized CT course or module, an already existing subject or just as a once-off event, CT can be taught in an enjoyable and engaging way whilst teaching students vital skills which can be applied across the curriculum as well as in daily and work life.

There is no doubt that CT is important and must be developed early in our students. While there are many unplugged activities that can be used, it has been shown that programming is a natural vehicle to develop this skill. In younger learners, use of robotic programming or coding as the buzz word can help to inculcate computational thinking. As has been determined in the reviewed studies, programming or coding and robotics appear to be a major player in the development of CT skills. Since it has been agreed that CT should be developed early in the learner, coding or programming should be taught as compulsory aspects in the curriculum to develop CT skills. However, [30] argue that in practice, it is mostly taught with a narrow focus with just common exercises and testing. An implication is that the teaching strategy needs to change to foreground CT development, with appropriate assessments during programming and coding teaching.

Most if not all sectors of the job market will require some form of coding or programming. Hence, it is important that they can work with algorithmic problem solving and computational methods and tools, a process that should begin in schools. The successful integration of computational thinking concepts into the curriculum requires endeavors in two paths. First, educational policy must be amended to cater to this need and secondly, overcoming infrastructure hurdles, such as the need for teacher resources and teacher education and training. Some emphasis is placed on teacher education regarding the development of CT in STEM related fields [27, 29, 31, 34, 36, 45, 48].

In building problem-solving skills for students, the use of relevant and real problems enhances their understanding, to creatively think of computational steps towards a solution. Hence, it is important to design a learning tool that allow users to teach/learn programming concepts through CT approach while abstracting problems that are familiar within a context.

5.1 Limitations

Some limitations need to be acknowledged. Firstly, a small number of publications (31) qualified for inclusion, and the database search was restricted to just the ERIC database as explained earlier. Hence, other databases of repute should be used as well and more research needs to be carried out to confirm or otherwise, the findings. A potential bias for the study is the influence the researcher had upon the analysis, despite screening the articles at least three times for the final review. Although the current study is based on a limited sample, this work offers valuable insights into the status of CT in STEM and lessons in developing CT in our students.

6. Conclusion

The study set out to determine the trends and status of computational thinking in STEM fields, by problematizing the lack of development strategies of CT within STEM fields. A systematic review was followed. What has become clear is awareness of the need for increasing the uptake of STEM subjects and the acknowledgement of computational skill as a twenty-first century skill. It has been established that CT is a necessary skill to develop in previous studies and in the current view. Several interventions to teach and develop CT in the STEM fields have been explored, indicating robotics as a driver for primary school children to learn CT. It has become abundantly clear that programming and coding with robotics appear to be most used for the development of CT are key to fostering of CT skills. To conclude, the results of this study indicate that there is much work to do regarding teacher education to promote CT skills in their curricular. Despite the emerging research specifically in CT in STEM fields, more needs to be done. Research in CT disciplinary pedagogical studies and transdisciplinary studies need to be conducted.