View in EDS HTML Full Text PDF Full Text

Integration of Problem-Based Learning in Elementary Computer Science Education: Effects on Computational Thinking and Attitudes

Saved in:
Bibliographic Details
Title: Integration of Problem-Based Learning in Elementary Computer Science Education: Effects on Computational Thinking and Attitudes
Language: English
Authors: Kwon, Kyungbin (ORCID 0000-0001-8646-0144), Ottenbreit-Leftwich, Anne T., Brush, Thomas A., Jeon, Minji (ORCID 0000-0002-0301-2221), Yan, Ge
Source: Educational Technology Research and Development. Oct 2021 69(5):2761-2787.
Availability: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
Peer Reviewed: Y
Page Count: 27
Publication Date: 2021
Document Type: Journal Articles
Reports - Research
Tests/Questionnaires
Education Level: Elementary Education
Grade 6
Intermediate Grades
Middle Schools
Descriptors: Problem Based Learning, Elementary School Curriculum, Elementary School Students, Student Attitudes, Computer Science Education, Programming, Thinking Skills, Instructional Effectiveness, Concept Formation, Retention (Psychology), Knowledge Level, Prior Learning, Grade 6
DOI: 10.1007/s11423-021-10034-3
ISSN: 1042-1629
Abstract: This study investigated how a computer science (CS) problem-based curriculum impacted elementary students' CS learning and attitudes. Four sixth-grade teachers and 200 of their students participated in the study. Researchers developed a CS curriculum in collaboration with the teachers, which consisted of two main units: (1) an introduction to block-based coding and (2) a problem-based learning (PBL) applied coding project. Overall, students significantly improved their knowledge of CT concepts after the introductory block-based coding lessons and retained that knowledge after completing the PBL activities approximately three months later. Results suggest that "Event" and "Parallelism" were challenging concepts for most of the students, whereas "Loop" and "Sequence" were easily grasped by most of the students. Further analysis based on prior knowledge levels revealed that the high-prior knowledge (HK) group outperformed the low-prior knowledge (LK) group on every measure. However, LK narrowed the gap of CT concepts after the introductory block-based coding lessons. Students also communicated relatively positive attitudes towards CS at the conclusion of the PBL unit. These results provide support for further exploring the integration of inquiry-oriented instructional strategies such as PBL to support CS instruction.
Abstractor: As Provided
Entry Date: 2021
Accession Number: EJ1316649
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwGyq6bf5ohIQ-JV2eE-2i8hAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDPPwUREv55toT93MawIBEICBm13Fimc8u71TQpmvVPI5QX_6zJKMRf0miCPwiXrDo1TlHIkgGaaFbzseyeGOtp8Lq6suJewVsBUsfF8fqozOHr-RXSJpVNK-gMGblJCFehyajPawBXQbNF6tpU7Yq5zfCdRZt-ni7FnaVjZkJC7DJLZAp6zengOGJIFkZjtrijAUDVMi0UyG53zLhVUGooLAHXo7lC7LMYkL9KIG
Text:
  Availability: 1
  Value: <anid>AN0153455265;etr01oct.21;2021Nov10.02:41;v2.2.500</anid> <title id="AN0153455265-1">Integration of problem-based learning in elementary computer science education: effects on computational thinking and attitudes </title> <p>This study investigated how a computer science (CS) problem-based curriculum impacted elementary students' CS learning and attitudes. Four sixth-grade teachers and 200 of their students participated in the study. Researchers developed a CS curriculum in collaboration with the teachers, which consisted of two main units: (<reflink idref="bib1" id="ref1">1</reflink>) an introduction to block-based coding and (<reflink idref="bib2" id="ref2">2</reflink>) a problem-based learning (PBL) applied coding project. Overall, students significantly improved their knowledge of CT concepts after the introductory block-based coding lessons and retained that knowledge after completing the PBL activities approximately three months later. Results suggest that Event and Parallelism were challenging concepts for most of the students, whereas Loop and Sequence were easily grasped by most of the students. Further analysis based on prior knowledge levels revealed that the high-prior knowledge (HK) group outperformed the low-prior knowledge (LK) group on every measure. However, LK narrowed the gap of CT concepts after the introductory block-based coding lessons. Students also communicated relatively positive attitudes towards CS at the conclusion of the PBL unit. These results provide support for further exploring the integration of inquiry-oriented instructional strategies such as PBL to support CS instruction.</p> <p>Keywords: Computer science education; Computational thinking; Problem-based learning; Elementary CS education; Block-based programming</p> <hd id="AN0153455265-2">Introduction</hd> <p>K-12 computer science (CS) education has been rapidly growing and significantly transformed toward innovative learning experiences in the United States during the last decade (Hsu et al., [<reflink idref="bib30" id="ref3">30</reflink>]; The White House, [<reflink idref="bib67" id="ref4">67</reflink>]). Scholars have suggested that K-12 CS education has become more prevalent due to several demands, including the need for more CS programmers (National Research Council, [<reflink idref="bib49" id="ref5">49</reflink>]; Richards & Terkanian, [<reflink idref="bib54" id="ref6">54</reflink>]), and the increasing view that CS is a fundamental twenty-first century skill that should be required by all students.</p> <p>However, K-12 schools do not satisfy public demands by offering unsatisfactory CS courses (Google & Gallup, [<reflink idref="bib28" id="ref7">28</reflink>]); furthermore, they face equity and diversity concerns in CS education (Vakil, [<reflink idref="bib68" id="ref8">68</reflink>]). As a national effort in the United States to establish curriculum standards, the Framework for K-12 Computer Science Education (https://<ulink href="http://www.k12cs.org">www.k12cs.org</ulink>) was published in 2016 (K-[<reflink idref="bib35" id="ref9">35</reflink>] Computer Science Framework Steering Committee, [<reflink idref="bib35" id="ref10">35</reflink>]), and organizations such as the Computer Science Teachers Association (CSTA) and the International Society for Technology in Education (ISTE) have released standards (CSTA, [<reflink idref="bib18" id="ref11">18</reflink>]). These standards place an emphasis on understanding the fundamental concepts of CS and practicing problem-solving tasks in an inclusive and diverse computing culture while collaborating with peers.</p> <p>To satisfy the standards, many stakeholders have explored various strategies for integrating CS curricula in K-12 (e.g., Weintrop et al., [<reflink idref="bib71" id="ref12">71</reflink>]). As the goal of K-12 CS education is aligned with Wing's ([<reflink idref="bib72" id="ref13">72</reflink>]) suggestion, "thinking like a computer scientist," learners are encouraged to complete projects that are related to their own interests by applying CS concepts. Many experts in K-12 CS education have focused on the problem-solving practices associated with CS, described as computational thinking (CT), which is defined as "solving problems, designing systems, and understanding human behaviour, by drawing on the concepts fundamental to computer science" (Wing, [<reflink idref="bib72" id="ref14">72</reflink>], p. 33). From this perspective, K-12 CS education has been encouraged to teach students how to solve problems by utilizing the fundamental concepts of CS, as opposed to only teaching students how to program (Brush et al., [<reflink idref="bib14" id="ref15">14</reflink>]). Especially, for young children who do not have prior CT learning experiences, understanding abstract CT concepts and applying them to problem solving tasks can be challenging (Armoni, [<reflink idref="bib2" id="ref16">2</reflink>]; Statter & Armoni, [<reflink idref="bib62" id="ref17">62</reflink>]). Children's cognitive development and social interactions during learning activities need to be considered in designing CT curricula. Thus, it is undeniably useful to explore how children learn CT concepts and practice them in regular classroom activities.</p> <p>Considering the growing needs of CS education in early educational systems, providing evidence-based pedagogical suggestions is becoming necessary. Although efforts introducing and exploring computing initiatives in elementary school settings are getting acknowledged, there is a scarcity of literature empirically revealing the effects of PBL approaches for K-6 CS curricula (Israel et al., [<reflink idref="bib33" id="ref18">33</reflink>]). A more in-depth examination of children's understanding of CT concepts and utilizing the concepts in developing solutions is highly demanded.</p> <p>Considering the needs, the current study aims to investigate how a CS PBL curriculum impacts elementary students' knowledge (CT concepts and practices) and attitudes towards CS. In addition, it is important to understand how different demographics are impacted by this curriculum. We have investigated how gender may have influenced students' experiences with this curriculum (see Kwon et al., [<reflink idref="bib40" id="ref19">40</reflink>]). However, we also wanted to investigate how students with different levels of prior CS knowledge experienced this curriculum. This study will contribute to understanding whether a PBL-integrated CS curriculum could be an effective pedagogical choice for teaching CS at the elementary level. The specific research questions addressed in this study are as follows:</p> <p></p> <ulist> <item> Were there any CT concepts that were significantly harder for elementary students to understand?</item> <p></p> <item> Do high and low prior knowledge elementary students demonstrate different learning patterns in their CT concepts and practices?</item> <p></p> <item> Do high and low prior knowledge elementary students differ in their attitudes towards CS?</item> <p></p> <item> Is there a relationship between elementary students' CT concepts/practices and their attitudes towards CS after experiencing a problem-based learning CS curriculum?</item> </ulist> <p>In the following sections, we will examine the CT concepts and practices that our CS curriculum aimed for and the role of PBL in CS education. We will also review factors that impact the effect of CS education.</p> <hd id="AN0153455265-3">CT concepts and practices</hd> <p>Many K-12 CS education studies have examined students' acquisition of CT skills, i.e., problem-solving skills used to analyze problems and devise solutions that will be carried out by a computer. It is important to understand that CT is a way of solving problems and designing systems by utilizing the fundamental concepts of CS (Wing, [<reflink idref="bib72" id="ref20">72</reflink>]). Computer scientists use CT skills to represent a given problem appropriately and model the relevant aspects of the problem, which produce multiple levels of abstraction. Based on the abstraction, they develop an algorithm that identifies a sequence of logical steps toward a solution that a computer can execute (Lu & Fletcher, [<reflink idref="bib45" id="ref21">45</reflink>]; Riley & Hunt, [<reflink idref="bib55" id="ref22">55</reflink>]). During this process, computer scientists think algorithmically, develop a model by recognizing patterns, and write instructions logically, which enables them to translate human solutions to programable solutions (Riley & Hunt, [<reflink idref="bib55" id="ref23">55</reflink>]).</p> <p>The common constructs of CT are decomposition, pattern recognition, abstraction, and algorithm (Brackmann et al., [<reflink idref="bib11" id="ref24">11</reflink>]; Selby & Woollard, [<reflink idref="bib59" id="ref25">59</reflink>]; Wing, [<reflink idref="bib72" id="ref26">72</reflink>]). These four constructs represent cognitive skills that are required to solve computer science problems: (<reflink idref="bib1" id="ref27">1</reflink>) recognize problems and break them down into smaller and workable pieces (decomposition), (<reflink idref="bib2" id="ref28">2</reflink>) find similarities or characteristics that the decomposed problems share (pattern recognition), (<reflink idref="bib3" id="ref29">3</reflink>) specify core information that is necessary to solve the problems while ignoring irrelevant ones (abstraction), and (<reflink idref="bib4" id="ref30">4</reflink>) develop a sequence of logical rules to solve problems (algorithm).</p> <p>In terms of integrating CT skills into K-12 curricula, Brennan and Resnick ([<reflink idref="bib12" id="ref31">12</reflink>]) described building CT by examining three constructs (concepts, practices, and perspectives) in the context of utilizing block-based programming. When developing a program, Brennan and Resnick found that learners would typically use seven computational concepts (sequences, loops, parallelism, events, conditionals, operators, and data). Computational practices refer to the processes of program design and development, which are being incremental and iterative in designing code, testing and debugging processes, modifying other solutions, and applying abstraction and module into their solutions. Computational practices reveal how students learn computational concepts while developing programs. Computational perspectives are attitudes that students will develop during and after computational practices, where they would express their ideas, build communities, and make inquiries of real-life issues. We explore each of these CT concepts and practices below.</p> <hd id="AN0153455265-4">CT concepts</hd> <p></p> <hd id="AN0153455265-5">Sequence</hd> <p>Sequences are related to the order in which a computer executes a set of predefined instructions. A computer does not read and understand the whole set of instructions before executing them. Instead, it reads the first line of instructions and performs the required computational tasks, then moves to the next line in sequence, and so forth. Sequences, deciding on the correct order of objects or actions, are essential for planning a program (Zelazo et al., [<reflink idref="bib74" id="ref32">74</reflink>]).</p> <hd id="AN0153455265-6">Loop and conditional</hd> <p>Loops and conditionals are related to the control flow that decides which sets of instructions will be executed in what condition. Loops specify how many iterations a set of instructions will run in the same sequence. Conditionals decide which instructions will run in what circumstances. Loops can include a conditional statement to specify a condition to stop the iteration. As the control flow gets complex by adding multiple conditions, it can be more difficult for learners to utilize loops and conditionals (Bers et al., [<reflink idref="bib7" id="ref33">7</reflink>]).</p> <hd id="AN0153455265-7">Event</hd> <p>An event triggers the execution of a set of predefined instructions. Events can be defined by user actions, such as mouse clicks, key presses, or messages sent from other programs. It is quite natural even for novice programmers to understand how a particular event activates predefined instructions (Bruce et al., [<reflink idref="bib13" id="ref34">13</reflink>]). In Scratch, for example, learners simply use "when green flag clicked" or "when sprite clicked" to begin a program. However, if learners use a message (i.e., broadcast block in Scratch) to activate certain instructions in a specific moment, they need to manage multiple event-handlings.</p> <hd id="AN0153455265-8">Parallelism</hd> <p>Parallelism refers to the way a computer executes sequences of instructions simultaneously. Bogaerts ([<reflink idref="bib8" id="ref35">8</reflink>]) makes an analogical comparison between parallelism in programming and real-world work processes, suggesting students can learn core parallelism concepts "if presented in the right context and with the right scaffolding" (p. 1263). However, as parallelism can increase the number of executions to be considered simultaneously, students may have a high cognitive load in dealing with the complexity (Bogaerts, [<reflink idref="bib10" id="ref36">10</reflink>]).</p> <hd id="AN0153455265-9">Variable</hd> <p>Variables are closely related to data in programming. The fundamental function of variables is to save and update values, which allows a program to be employed for addressing real-life problems that are parameterized by alterable elements. Many computational solutions involve data, which are represented by variables. When designing solutions carried out by a computer, learners should be able to define variables, identify values assigned to the variables, and utilize the variables in subsequent computations, which is quite challenging for novice programmers (Kwon et al., [<reflink idref="bib39" id="ref37">39</reflink>]). When learners do not understand the function of variables, they cannot develop proper solutions in a program (Shi et al., [<reflink idref="bib60" id="ref38">60</reflink>]).</p> <hd id="AN0153455265-10">CT practices</hd> <p>Computational practice emphasizes the importance of learning experiences where students apply computational concepts to develop computational artifacts. The crucial question is how to evaluate and guide computational practice in classrooms. Brennan and Resnick ([<reflink idref="bib12" id="ref39">12</reflink>]) observed students' design practices and described a variety of computational artifact creation processes. They found that the processes when students planned and developed code were adaptive. While creating projects, students modified their plans as they made progress toward their final goal, which involved iterative cycles of design, development, and test. The iterative cycles of CT practices were most likely to involve a trial-and-error approach, a common problem-solving method for less experienced students. Brennan and Resnick suggest that students need to demonstrate abstraction and modulization skills in their projects, as these skills are one of the most important objectives of computational practice.</p> <p>Lee et al. ([<reflink idref="bib43" id="ref40">43</reflink>]) suggested a <emph>use-modify-create</emph> framework that represents the three phases of cognitive and practical activities when students use CT to approach novel problems. The framework describes how students initially run pre-made solutions (use) and gain new skills and understandings through a series of modifications and iterative refinements (modify). After gaining skills and confidence, students are encouraged to develop new computational projects that address their own problems (create). This framework suggests that students should have optimal challenges for learning while limiting anxiety by presenting progressively higher design challenges during computational practice.</p> <p>As briefly reviewed, design and reflection are critical skills required in computational practices. To create computational artifacts, students analyze a problem (identify a goal to achieve) and properly break it down into sub-tasks, which can be associated toward a solution (Shute et al., [<reflink idref="bib61" id="ref41">61</reflink>]). This design process requires a communication skill to represent the solution logically and efficiently, which evolves abstraction, algorithm, and automation skills (Kim et al., [<reflink idref="bib37" id="ref42">37</reflink>]). Thus, practicing a design process can be beneficial to understand CT concepts and develop logical thinking in computational practices. In another aspect, evaluations of designs can reveal students' mental models when utilizing CT concepts to solve the problem (Kwon et al., [<reflink idref="bib39" id="ref43">39</reflink>]).</p> <p>While reflecting, students need to evaluate their solutions and fix errors (debug) if they find a discrepancy between the actual result of their code and what they intended. A common debugging strategy students use is a trial-and-error approach (Brennan & Resnick, [<reflink idref="bib12" id="ref44">12</reflink>]). While applying the trial-and-error method, students identify parts of code that cause an error, modify them, then they check the results and repeat this process until the result is satisfying. Although it is easy to adopt this method because of its simplicity, its educational benefits are limited (Rich et al., [<reflink idref="bib53" id="ref45">53</reflink>]). Then, why do students not utilize better debugging strategies? One may suggest that students do not have enough knowledge (i.e., computational concepts). However, Lewis ([<reflink idref="bib44" id="ref46">44</reflink>], p. 127) argued that "a key competence in debugging is learning to identify what elements of program state are important to pay attention to." Lewis suggested that students need to identify a problem that they should pay attention to and identify its cause to figure out how to fix it. This explains why students who have relevant CT concepts do not apply their knowledge while debugging. Thus, computational practices should go beyond learning CT concepts and practice design and foster reflection in authentic learning contexts.</p> <hd id="AN0153455265-11">PBL for CS education</hd> <p></p> <hd id="AN0153455265-12">Educational merits of PBL</hd> <p>Many have pointed out that PBL could be a way to target both knowledge and attitudes. Previous studies on K-12 have shown that PBL curriculum can improve students' knowledge and attitudes, for a wide range of types of students, in a wide range of topics (Belland et al., [<reflink idref="bib6" id="ref47">6</reflink>]; Strobel & van Barneveld, [<reflink idref="bib63" id="ref48">63</reflink>]). PBL is a learner-centered instructional approach that "empowers learners to conduct research, integrates theory and practice, and applies knowledge and skills to develop a viable solution to a defined problem" (Savery, [<reflink idref="bib57" id="ref49">57</reflink>], p. 12). In PBL, ill-structured problems facilitate and motivate learning by presenting the challenges students will face in practice or daily life; learners take responsibility of their learning and identify what they need to know to solve the problems; learning occurs in small groups supported by facilitators, not instructors; as a result, learners develop problem-solving skills and become self-directed learners (Barrows, [<reflink idref="bib5" id="ref50">5</reflink>]). Strobel and van Barneveld ([<reflink idref="bib63" id="ref51">63</reflink>]) found, in their qualitative meta-synthesis of primary studies in the field of medicine, economy, and CS, that PBL was more effective for long-term knowledge retention, skill-oriented performances, and the satisfaction of both students and teachers, whereas a traditional instructional approach was in favor of short-term retention measured by multiple-choice tests.</p> <hd id="AN0153455265-13">Expected merits of PBL in CS education</hd> <p>Considering the general PBL effects on learning, many stakeholders expect positive learning outcomes of integrating PBL components into CS education by utilizing various initiatives, such as integrating robotics, unplugged activities, and block-based programming. Regarding students' interest in and attitude toward CS, the PBL approach has brought promising results, especially at an elementary level (e.g., Lambert & Guiffre, [<reflink idref="bib41" id="ref52">41</reflink>]; Sáez-López et al., [<reflink idref="bib56" id="ref53">56</reflink>]). However, studies have found inconclusive results regarding the learning outcome of CT concepts. For example, Chen et al. ([<reflink idref="bib17" id="ref54">17</reflink>]) developed a robotics curriculum, which guided fifth-grade students to program a robot to carry out basic functions using a block-based program. The curriculum integrated everyday scenarios so that students could make connections between robotics and daily life tasks. The curriculum brought positive learning outcomes in terms of formulating solutions and algorithmic thinking. However, students still struggled with data processing and parallel execution. They also showed inconsistent improvement in everyday reasoning, which represented the transfer of CT skills. In a similar context, Bers et al. ([<reflink idref="bib7" id="ref55">7</reflink>]) integrated a visual robotics program into a kindergarten CS curriculum. They examined how early childhood students learned CT concepts through an engineering design process and debugging, robotic motion and sensing, and programming experience. Although the curriculum carefully considered the young students' developmental status, student achievement decreased as more challenging concepts were introduced.</p> <hd id="AN0153455265-14">Factors to be considered</hd> <p>For an effective CS curriculum for young students, PBL activities and learning environments should be selected in consideration of learners' cognitive developmental status and group dynamics in collaborative learning (Armoni, [<reflink idref="bib2" id="ref56">2</reflink>]). Students' content knowledge and learning strategies can affect their learning in CS education. Given the same problems, students who are older (more knowledgeable) can take them as interesting challenges, while those who are younger (less knowledgeable) may feel frustrated in understanding algorithms (Gibson, [<reflink idref="bib27" id="ref57">27</reflink>]). While interacting with programming environments, young students may have difficulties interpreting symbols and deciding on the directions of movement, which are not closely related to CT concepts (Fessakis et al., [<reflink idref="bib22" id="ref58">22</reflink>]).</p> <p>There are various strategies students apply when solving problems. Some students can apply structured problem-solving processes, such as planning solutions and identifying errors before modifying codes; however, many students simply apply the trial-and-error method (e.g., modifying codes arbitrarily until getting a satisfying result) or give up on their original plan and change their goals to be achievable (Fessakis et al., [<reflink idref="bib22" id="ref59">22</reflink>]; Sullivan & Heffernan, [<reflink idref="bib64" id="ref60">64</reflink>]). When lacking problem-solving experience, procedural knowledge, and reasoning skills, students utilize more procedural strategies (i.e., trial-and-error) shifting toward structured heuristics (involving forward or backward reasoning) as experienced (Gaudiello & Zibetti, [<reflink idref="bib26" id="ref61">26</reflink>]).</p> <p>Group dynamics in solving problems affect performance. When students solve computational problems in pairs, they tend to perform better than when they work alone (Iskrenovic-Momcilovic, [<reflink idref="bib32" id="ref62">32</reflink>]; McDowell et al., [<reflink idref="bib46" id="ref63">46</reflink>]). However, there are many issues to be explored for a better understanding of group dynamics to impact the effectiveness of collaborative learning, such as compatibility among team members, combinations of genders, levels of expertise, preferences toward collaboration, instructions for collaboration, and so on (Denner et al., [<reflink idref="bib21" id="ref64">21</reflink>]; Taylor & Baek, [<reflink idref="bib66" id="ref65">66</reflink>]; Zhong et al., [<reflink idref="bib75" id="ref66">75</reflink>]). For example, higher ability students are more dominant and influential than those with lower ability in collaboration (Dembo & McAuliffe, [<reflink idref="bib19" id="ref67">19</reflink>]).</p> <p>As briefly reviewed, students' knowledge and skills may affect the outcomes of CS education that involves group interactions. Therefore, it seems critical to examine whether PBL integrated CS curriculum bring different outcomes regarding students' prior knowledge.</p> <hd id="AN0153455265-15">Method</hd> <p></p> <hd id="AN0153455265-16">Participants</hd> <p>A total of 200 sixth-grade students (93 girls and 107 boys) from four elementary schools participated in the study. The schools belong to a public-school corporation located in a small city, Midwestern United States. Most students had no previous experience with block-based programming prior to this study. All teachers participating in the study had over ten years of teaching experience, but this was the first year they had been asked to integrate CS instruction into their science classes. Although one teacher had supervised the school's robotics club, no teachers had received formal CS professional development. The teachers also had limited experiences integrating PBL strategies into their classrooms.</p> <hd id="AN0153455265-17">CS curriculum development</hd> <p>The researchers developed a CS curriculum in collaboration with four 6th-grade teachers and an elementary STEM coach. Utilizing frameworks from Brennan and Resnick ([<reflink idref="bib12" id="ref68">12</reflink>]), the CS Framework (K-[<reflink idref="bib35" id="ref69">35</reflink>] Computer Science Framework Steering Committee, [<reflink idref="bib35" id="ref70">35</reflink>]), and the Indiana CS 6–8 standards (Indiana Department of Education, [<reflink idref="bib31" id="ref71">31</reflink>]), our curriculum focused on CT concepts (Sequences, Loops, Event, Condition, Parallelism, Data, Operator) and CT practices (deciding topics, decomposing tasks, developing programs, demonstrating programs). The CS curriculum consisted of two units: (<reflink idref="bib1" id="ref72">1</reflink>) introduction to block-based coding and (<reflink idref="bib2" id="ref73">2</reflink>) PBL applied coding project (see Table 1).</p> <p>Table 1 CS curriculum week-by-week topics and activities</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Week</p></th><th align="left"><p>Topic</p></th><th align="left"><p>Activity</p></th></tr></thead><tbody><tr><td align="left" colspan="3"><p><italic>Introductory Block-based coding</italic></p></td></tr><tr><td align="left"><p><bold>Week 1</bold></p><p>CS Introduction & Foundations</p></td><td align="left"><p>Students are introduced to the basic ideas of computer science, hardware, software, and computer components</p></td><td align="left"><p>Videos and discussions on relationship with humans and machines. Unplugged activities on binary and communication</p></td></tr><tr><td align="left"><p><bold>Weeks 2–4</bold></p><p>Extending CS Knowledge</p></td><td align="left"><p>Students are introduced to Scratch and the functions of different Scratch blocks</p></td><td align="left"><p>Students create at least 5 programs in Scratch (e.g., a dance party, a maze, a quiz game, and a variables game, and functions). Final project has students create their own program incorporating these ideas</p></td></tr><tr><td align="left" colspan="3"><p><italic>PBL applied coding project</italic></p></td></tr><tr><td align="left"><p><bold>Week 5</bold></p><p>Contextualizing the Problem</p></td><td align="left"><p>Students introduced to the PBL problem of "How can we create a culture of kindness in our school?"</p></td><td align="left"><p>Videos and discussions on creating a culture of kindness. Students research how acts of kindness in their school and daily lives can contribute to a culture of kindness</p></td></tr><tr><td align="left"><p><bold>Weeks 6–7</bold></p><p>Research and Design</p></td><td align="left"><p>Students design and develop an "APP" using Scratch to address the PBL problem</p></td><td align="left"><p>Students spend 8 lessons researching, planning, designing, and developing their Scratch project. Multiple check-ins and scaffolds help support development</p></td></tr><tr><td align="left"><p><bold>Week 8</bold></p><p>Presenting</p></td><td align="left"><p>Students present their final Scratch projects</p></td><td align="left"><p>Students present and share their final projects with their peers, teacher, other students, and external visitors</p></td></tr></tbody></table> </ephtml> </p> <hd id="AN0153455265-18">Introduction to block-based coding (Scratch lessons)</hd> <p>The primary goal of the introduction to block-based coding was to introduce the basic CT concepts using a block-based program (Scratch). The curriculum was also designed to address the Indiana 6–8 grade band CS standards (refer to the site for more information: https://tinyurl.com/5x539vhf). In these lessons, the teachers demonstrated how to construct a Scratch project that would incorporate one or two new CT concepts with each progressing lesson. For example, in Lesson 6, students created a quiz program that asked questions and evaluated user inputs to provide feedback accordingly. This lesson covered using if-else blocks to make decisions (condition) and how a score could be updated (variable). Researchers and teachers intentionally designed the lessons to explain CT concepts through Scratch projects that served as effective examples demonstrating how CT concepts were applied (Atkinson et al., [<reflink idref="bib3" id="ref74">3</reflink>]).</p> <p>During the instruction, students would typically observe a teacher's demonstration of a coding task, and then have an opportunity to build their own version of the program. This programming practice was guided by a structured decomposition activity which asked students to identify the topic, draw the main screens, and describe the tasks in the screens. Debugging skills and strategies were demonstrated by the teachers and researchers when students faced errors or could not make progress toward a particular task. Through introductory block-based coding, students were expected to get familiar with the interface of Scratch and demonstrate the basic CT concepts by developing Scratch projects. The nine lessons were implemented over nine to 13 class sessions, depending on extension activities.</p> <hd id="AN0153455265-19">PBL applied coding project</hd> <p>The PBL applied coding project was designed to provide students with an authentic programming experience that would examine a social issue that they encounter in their daily school lives. Teachers and researchers designed the applied coding project around the driving question: <emph>What tools can be designed to support a culture of kindness in our school?</emph> Students applied their newly acquired CT skills to address this question. The teachers selected the driving question because it was aligned with the social-emotional curricular goals and curriculum that the school district had recently adopted.</p> <p>The PBL project included a grabber activity lesson (introduction to the topic), brainstorming and researching activities, designing and creating the program, and finally a presentation of their team's work. Students were provided scaffolds and resources to conduct research on supporting kindness. To support their design, students used the decomposition worksheet they had practiced with during the Scratch lessons. To help guide the projects, teachers created three main types of "apps" that students could design using Scratch to address the culture of kindness: (<reflink idref="bib1" id="ref75">1</reflink>) 'choose your own adventure apps' that would allow users to make choices about being kind, (<reflink idref="bib2" id="ref76">2</reflink>) quiz games that would ask questions and provide suggestions on ways to be kind, or (<reflink idref="bib3" id="ref77">3</reflink>) kindness trackers that would record the number of kind behaviors with words of encouragement. At the end of the unit, students were asked to present their work to others. During their presentation, they explained how their Scratch projects worked and reflected on their programming experiences. The PBL lessons were implemented over approximately six to eight class sessions. For more information about the curriculum, please refer to the site: https://tinyurl.com/4hx7s635.</p> <hd id="AN0153455265-20">Procedures</hd> <p>The research team closely collaborated with four teachers and one district STEM coach in designing this CS curriculum and had a one-day professional development session regarding Scratch and learning activities. While implementing the curriculum, researchers first led the introduction to block-based coding sessions, and then let the teachers take over the PBL applied coding project. A few research team members attended every class and supported the teachers by solving unexpected technical issues and answering students' questions about Scratch codes.</p> <p>The teachers assigned students into PBL groups and allowed students to form groups if they had common interests. Each group had two to three students. There were two individuals who conducted PBL projects alone. Students' prior knowledge was not considered in the grouping process.</p> <p>Student assent and parent consent were obtained in accordance with IRB policies. On the first and last day of Scratch lessons, a pre-test and post-test were administered. The PBL project was implemented about two to four months later due to a winter break and other school curriculum requirements. At the end of the PBL project, students took a second post-test and responded to an attitudinal survey. Students were given 20 min for each test, and everyone completed them within the allocated time.</p> <hd id="AN0153455265-21">Measurements</hd> <p></p> <hd id="AN0153455265-22">Computational thinking tests</hd> <p>The researchers and teachers collaboratively developed a 14-item multiple-choice assessment designed to measure students' knowledge of basic CS and coding principles. These were developed based on CT concepts (Brennan & Resnick, [<reflink idref="bib12" id="ref78">12</reflink>]; CSTA, [<reflink idref="bib18" id="ref79">18</reflink>]; ISTE & CSTA, [<reflink idref="bib34" id="ref80">34</reflink>]). Eight of the test items focused on basic CS concepts (sequences, loops, event, conditional, variables, parallelism), and six of the items focused on CS practices (debugging of errors belonging to one of the six concepts). To get valid assessments of CT concepts and debugging skills, we developed questions (<reflink idref="bib1" id="ref81">1</reflink>) to include only one concept for each and (<reflink idref="bib2" id="ref82">2</reflink>) provide detailed contexts of problems so that students understand problems within the authentic context (see Fig. 1). Each item was scored one or zero for a correct or incorrect response, respectively.</p> <p>Graph: Fig. 1 Example question of CT tests</p> <p>Parallel forms of the test, measuring the same concepts whose item difficulty levels were similar, were administered to students on three occasions: prior to the beginning of the unit (pre-test), immediately after the completion of the Scratch lessons (post-test 1), and immediately after the completion of the PBL portion of the unit (post-test 2). The pre-test and post-test 1 were delivered 3 weeks apart in the fall of 2017. Post-test 2 was delivered in the spring of 2018. See Fig. 1 for an example of the format of the test. The internal reliability of each test was calculated using Cronbach's alpha and resulted in acceptable degrees; the values of pre-test, post-test 1, and post-test 2 are 0.70, 0.77, and 0.77, respectively (Kline, 1999). Between three parallel forms of the tests, Spearman–Brown coefficients were calculated and ranged from a low of 0.77 to a high of 0.82, which confirmed the parallel forms reliability of the tests.</p> <hd id="AN0153455265-23">Attitudinal survey</hd> <p>The 16-item close-ended survey, administered to students after the end of the PBL project, was designed to measure: (<reflink idref="bib1" id="ref83">1</reflink>) self-efficacy toward programming (4 items); (<reflink idref="bib2" id="ref84">2</reflink>) attitude toward CS (6 items); and (<reflink idref="bib3" id="ref85">3</reflink>) confidence in CT skills (6 items). Scoring was based on a 5-point Likert scale ranging from "strongly disagree" to "strongly agree" (1–5), with higher scores indicating a high degree of agreement. The items were adopted from literature and modified in consideration of the characteristics of the target students (Cetin & Ozden, [<reflink idref="bib16" id="ref86">16</reflink>]; Ramalingam & Wiedenbeck, [<reflink idref="bib51" id="ref87">51</reflink>]). Cronbach's alphas for scales were calculated and suggested high internal consistency reliability; self-efficacy, attitude, and confidence were 0.83, 0.84, and 0.90, respectively (see Appendix for survey items).</p> <hd id="AN0153455265-24">Analysis</hd> <p>A repeated measures analysis of variance (ANOVA) was conducted on students' pre-test and both post-test scores of each unit. Additionally, a secondary analysis of the test scores was conducted to determine if participation in the CS-PBL unit has differential effects based on student's prior knowledge of the content. Students who scored below the mean on the pre-test were classified as <emph>low-prior knowledge</emph>, and students scoring at or above the mean were classified as <emph>high-prior knowledge</emph>. This method of determining low- and high-prior knowledge groups for the purposes of a secondary analysis has been utilized in previous research (Acuña et al., [<reflink idref="bib1" id="ref88">1</reflink>]; Belland et al., [<reflink idref="bib6" id="ref89">6</reflink>]). Additional ANOVAs were conducted for the test results from the low prior knowledge and high prior knowledge groups. Effect sizes for each comparison were calculated using partial eta squared.</p> <p>Student attitudes (as measured by the student attitudinal survey) were analyzed via descriptive statistics. Furthermore, one-way ANOVAs were used to determine differences in student attitudes with regard to self-efficacy towards CS based on prior knowledge group.</p> <hd id="AN0153455265-25">Results</hd> <p></p> <hd id="AN0153455265-26">CT knowledge tests based on concepts</hd> <p></p> <hd id="AN0153455265-27">Differences among concepts</hd> <p>Researchers examined whether there were any differences between the six CT concepts assessed in the tests. Because each concept was assessed by a different number of questions (two to three), the percentages of correct answers were used in the analysis. As Table 2 illustrates, considerable variations among the concepts were observed within and between the tests.</p> <p>Table 2 Percentage correct of each concept in CT tests</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Concepts</p></th><th align="left"><p>Event</p></th><th align="left"><p>Parallelism</p></th><th align="left"><p>Conditional</p></th><th align="left"><p>Variable</p></th><th align="left"><p>Sequence</p></th><th align="left"><p>Loop</p></th><th align="left"><p>Comparisons<sup>a</sup></p></th></tr></thead><tbody><tr><td align="left"><p>Pre-test</p></td><td align="left"><p>28.8(33.4)</p></td><td align="left"><p>37.0(34.1)</p></td><td align="left"><p>48.8(29.5)</p></td><td align="left"><p>65.3(36.2)</p></td><td align="left"><p>58.5(37.2)</p></td><td align="left"><p>71.5(30.5)</p></td><td align="left"><p>Event, Parallelism < other concepts</p><p>Conditional < Sequence < Loop</p><p>Conditional < Variable</p></td></tr><tr><td align="left"><p>Post-test I</p></td><td align="left"><p>61.2(29)</p></td><td align="left"><p>63.5(33.1)</p></td><td align="left"><p>75.5(25.6)</p></td><td align="left"><p>68.7(39.6)</p></td><td align="left"><p>88.2(25.6)</p></td><td align="left"><p>83.3(24.1)</p></td><td align="left"><p>Event, Parallelism < Conditional < Sequence, Loop</p><p>Variable < Sequence, Loop</p></td></tr><tr><td align="left"><p>Post-test II</p></td><td align="left"><p>68(33.6)</p></td><td align="left"><p>65.2(33.3)</p></td><td align="left"><p>71.2(29.1)</p></td><td align="left"><p>75.3(35.8)</p></td><td align="left"><p>79.0(31.5)</p></td><td align="left"><p>84.0(21.4)</p></td><td align="left"><p>Event, Parallelism, Conditional < Sequence, Loop</p><p>Parallelism < Variable < Loop</p></td></tr></tbody></table> </ephtml> </p> <p>M(SD), Means represent the average percentages of correct answers of each concept <sups>a</sups>Statistically significant difference at 5% level</p> <p>To examine statistically significant differences among the CT concepts within each test, a multivariate analysis of variance (MANOVA) was conducted with the scores of the six CT concepts as dependent variables. Regarding the pre-test, Wilks's statistic revealed that there were significant differences among the CT concepts, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>F</mi><mfenced close=")" open="("><mn>5</mn><mo>,</mo><mn>195</mn></mfenced><mo>=</mo><mn>61.96</mn><mo>,</mo><mi>p</mi><mo><</mo><mo>.</mo><mn>001</mn><mo>;</mo><mi mathvariant="normal">Λ</mi><mo>=</mo><mo>.</mo><mn>39</mn><mo>,</mo><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo><mo>.</mo><mn>61</mn><mo>.</mo></mrow></math> </ephtml> Separate univariate tests on the CT concepts revealed that all students had lower scores in <emph>Event</emph> and <emph>Parallelism</emph> than the other concepts, which were statistically significant. <emph>Conditional</emph> was lower than <emph>Sequence</emph> which was lower than <emph>Loop</emph>. <emph>Variable</emph> scored higher than <emph>Event</emph>, <emph>Parallelism</emph>, and <emph>Conditional</emph> but did not show significant difference compared to <emph>Sequence</emph> and <emph>Loop</emph>.</p> <p>Regarding post-test 1, Wilks's statistic revealed that there were also significant differences among the CT concepts, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>F</mi><mfenced close=")" open="("><mn>5</mn><mo>,</mo><mn>195</mn></mfenced><mo>=</mo><mn>41.80</mn><mo>,</mo><mi>p</mi><mo><</mo><mo>.</mo><mn>001</mn><mo>;</mo><mi mathvariant="normal">Λ</mi><mo>=</mo><mo>.</mo><mn>48</mn><mo>,</mo><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo><mo>.</mo><mn>52</mn><mo>.</mo></mrow></math> </ephtml> Separate univariate tests on the CT concepts revealed that students had lower scores in <emph>Event</emph> and <emph>Parallelism</emph> in comparison to the other concepts except for <emph>Variable</emph>. <emph>Conditional</emph> and <emph>Variable</emph> scored lower than <emph>Sequence</emph> and <emph>Loop</emph>. No significant differences were confirmed between <emph>Event</emph> and <emph>Parallelism</emph>, <emph>Conditional</emph> and <emph>Variable</emph>, and <emph>Sequence</emph> and <emph>Loop</emph>, respectively.</p> <p>Regarding post-test 2, Wilks's statistic revealed that there were significant differences among the CT concepts, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>F</mi><mfenced close=")" open="("><mn>5</mn><mo>,</mo><mn>195</mn></mfenced><mo>=</mo><mn>20.65</mn><mo>,</mo><mi>p</mi><mo><</mo><mo>.</mo><mn>001</mn><mo>;</mo><mi mathvariant="normal">Λ</mi><mo>=</mo><mo>.</mo><mn>65</mn><mo>,</mo><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo><mo>.</mo><mn>35</mn><mo>.</mo></mrow></math> </ephtml> Separate univariate tests on the CT concepts revealed students had lower scores in <emph>Event</emph>, <emph>Parallelism</emph>, and <emph>Conditional</emph> compared to <emph>Sequence</emph> and <emph>Loop</emph>. <emph>Variable</emph> was higher than <emph>Parallelism</emph> and lower than <emph>Loop</emph>. No significant differences were confirmed among <emph>Event</emph>, <emph>Parallelism</emph> and <emph>Conditional</emph>, and between <emph>Sequence</emph> and <emph>Loop</emph>, respectively. Overall, students continuously achieved the lowest scores in <emph>Event</emph> and <emph>Parallelism</emph> followed by <emph>Conditional</emph> and <emph>Variable</emph>. <emph>Loop</emph> and <emph>Sequence</emph> were the concepts that students grasped most successfully.</p> <hd id="AN0153455265-28">Learning gains over time</hd> <p>Researchers examined the learning gains observed across the tests per concept. Repeated ANOVAs on the effect of time (pre-test, post-test 1, and post-test 2) in each of the CT concepts were carried out. Students showed significant improvements in all of the concepts from the pre-test to the post-test I (mean differences = 11.8% ~ 32.5%, <emph>p'</emph>s <.001), except for <emph>Variable</emph>. Regarding <emph>Variable</emph>, students showed significant improvement from the pre-test to post-test II (mean difference = 10.0%, <emph>p</emph> =.002) but not in post-test I. Regarding <emph>Event</emph>, students demonstrated continuous improvement from post-test I to post-test II (mean difference = 6.8%, <emph>p</emph> =.034). In <emph>Loop, Parallelism</emph>, and <emph>Conditional</emph>, there was no significant difference between the post-tests. Concerning <emph>Sequence</emph>, students showed a significant decline from post-test I to post-test II (mean difference = 9.8%, <emph>p</emph> <.001).</p> <hd id="AN0153455265-29">CT knowledge tests based on prior knowledge levels</hd> <p></p> <hd id="AN0153455265-30">Overall test results</hd> <p>This study examined whether prior knowledge level had any effect on students' knowledge gains of CT concepts and debugging skills. Researchers divided students into two groups (High- vs. Low-prior knowledge) based on the pre-test results (<emph>M</emph> = 7.4, <emph>SD</emph> = 2.93). Students who had higher than 7.4 were assigned to the High-prior knowledge group (HK) (<emph>n</emph> = 91, <emph>M</emph> = 10.0, <emph>SD</emph> = 1.65), and the others were assigned to the Low-prior knowledge group (LK) (<emph>n</emph> = 109, <emph>M</emph> = 5.2, <emph>SD</emph> = 1.71).</p> <p>Based on the groupings, researchers examined whether there were any significant differences between knowledge tests administered three times separately: before any units, after the introduction to block-based coding, and after PBL applied coding project (see Table 3). The result of a mixed ANOVA confirmed that students' overall scores had improved across tests, <emph>F</emph>(<reflink idref="bib2" id="ref90">2</reflink>, 396) = 221.8, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.53. The results also revealed that the students who were classified as HK outperformed the students who were classified as LK in every test. There was also a significant interaction between the times of tests and the groups, <emph>F</emph>(<reflink idref="bib2" id="ref91">2</reflink>, 396) = 23.5, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.11.</p> <p>Table 3 Means and standard deviations of test scores by high and low prior knowledge groups</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>Pre-test</p></th><th align="left"><p>Post-test I</p></th><th align="left"><p>Post-test II</p></th></tr></thead><tbody><tr><td align="left"><p>CT concepts</p></td><td align="left"><p>4.90 (1.84)</p></td><td align="left"><p>6.72 (1.71)</p></td><td align="left"><p>6.69 (1.64)</p></td></tr><tr><td align="left"><p>High</p></td><td align="left"><p>6.37 (1.01)</p></td><td align="left"><p>7.43 (0.88)</p></td><td align="left"><p>7.32 (0.94)</p></td></tr><tr><td align="left"><p>Low</p></td><td align="left"><p>3.68 (1.44)</p></td><td align="left"><p>6.13 (2.00)</p></td><td align="left"><p>6.17 (1.90)</p></td></tr><tr><td align="left"><p>Debugging</p></td><td align="left"><p>2.50 (1.56)</p></td><td align="left"><p>3.68 (1.50)</p></td><td align="left"><p>3.71 (1.68)</p></td></tr><tr><td align="left"><p>High</p></td><td align="left"><p>3.65 (1.28)</p></td><td align="left"><p>4.48 (1.20)</p></td><td align="left"><p>4.77 (1.18)</p></td></tr><tr><td align="left"><p>Low</p></td><td align="left"><p>1.53 (1.02)</p></td><td align="left"><p>3.01 (1.40)</p></td><td align="left"><p>2.83 (1.51)</p></td></tr><tr><td align="left"><p>Total</p></td><td align="left"><p>7.40 (2.93)</p></td><td align="left"><p>10.40 (2.84)</p></td><td align="left"><p>10.41 (2.88)</p></td></tr><tr><td align="left"><p>High</p></td><td align="left"><p>10.02 (1.65)</p></td><td align="left"><p>11.91 (1.74)</p></td><td align="left"><p>12.09 (1.72)</p></td></tr><tr><td align="left"><p>Low</p></td><td align="left"><p>5.21 (1.71)</p></td><td align="left"><p>9.14 (2.97)</p></td><td align="left"><p>9.00 (2.91)</p></td></tr></tbody></table> </ephtml> </p> <p>M(SD), Number of participants: Low = 109 and High = 91</p> <p>Post-hoc tests comparing adjacent tests revealed the significant main effects of time between pre-test and post-test 1, <emph>F</emph>(<reflink idref="bib1" id="ref92">1</reflink>, 198) = 324.4, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.62 and groups between LK and HK, <emph>F</emph>(<reflink idref="bib1" id="ref93">1</reflink>, 198) = 221.7, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.53, and a significant interaction effect of groups and time on scores, <emph>F</emph>(<reflink idref="bib1" id="ref94">1</reflink>, 198) = 39.8, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.17. The results confirmed that students' scores in both groups improved from the pre-test to post-test 1 and that HK outperformed LK consistently. It is noteworthy that there was a significant interaction between groups and times on the scores, which means that LK demonstrated larger learning gains compared with their counterpart.</p> <p>Post-hoc tests comparing the post-tests revealed a significant main effect of groups, <emph>F</emph>(<reflink idref="bib1" id="ref95">1</reflink>, 198) = 88.2, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.31, but not time. No significant interaction effect of groups and time on scores was found. The results revealed that the difference between groups in post-test 1 remained in post-test 2, and no considerable knowledge gain or loss was found in both groups between the tests.</p> <hd id="AN0153455265-31">Conceptual understanding</hd> <p>More specifically, researchers examined the scores of each test type separately using concepts and debugging. Regarding conceptual understanding, the results of a mixed ANOVA revealed significant main effects of time, <emph>F</emph>(<reflink idref="bib2" id="ref96">2</reflink>, 396) = 174.0, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.47. In every test, HK outperformed LK. There was also a significant interaction between the times of tests and the groups, <emph>F</emph>(<reflink idref="bib2" id="ref97">2</reflink>, 396) = 31.5, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.13.</p> <p>Post-hoc tests comparing adjacent times revealed significant main effects of time between pre-test and post-test I, <emph>F</emph>(<reflink idref="bib1" id="ref98">1</reflink>, 198) = 233.9, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.54 and groups between LK and HK, <emph>F</emph>(<reflink idref="bib1" id="ref99">1</reflink>, 198) = 140.5, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.42, and a significant interaction effect of groups and time on scores, <emph>F</emph>(<reflink idref="bib1" id="ref100">1</reflink>, 198) = 37.0, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.16. The results confirmed that students' conceptual understanding in both groups improved from the pre-test to post-test 1, and that HK outperformed LK consistently. There was also a significant interaction between groups and times on the scores.</p> <p>Post-hoc tests comparing post-tests revealed a significant main effect of groups, <emph>F</emph>(<reflink idref="bib1" id="ref101">1</reflink>, 198) = 38.2, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.16, but not time. No significant interaction effect of groups and time on scores was found. The results revealed that the difference between groups in post-test 1 continued in post-test 2, and no knowledge gain or lose was found between the tests.</p> <hd id="AN0153455265-32">Debugging</hd> <p>Regarding debugging skills, the results of a mixed ANOVA revealed significant main effects of time, <emph>F</emph>(<reflink idref="bib2" id="ref102">2</reflink>, 396) = 90.9, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.32. In every test, HK outperformed LK. There was also a significant interaction between the times of tests and the groups, <emph>F</emph>(<reflink idref="bib2" id="ref103">2</reflink>, 396) = 5.38, <emph>p</emph> =.005, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.03.</p> <p>Post-hoc tests comparing adjacent times revealed significant main effects of time between the pre-test and post-test I, <emph>F</emph>(<reflink idref="bib1" id="ref104">1</reflink>, 198) = 134.7, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.41 and groups between LK and HK, <emph>F</emph>(<reflink idref="bib1" id="ref105">1</reflink>, 198) = 159.8, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.44, and a significant interaction effect of groups and time on scores, <emph>F</emph>(<reflink idref="bib1" id="ref106">1</reflink>, 198) = 10.38, <emph>p</emph> =.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.05. The results confirmed that students' debugging skills in both groups improved from the pre-test to post-test I and that HK outperformed LK consistently. There was also a significant interaction between groups and times on the scores.</p> <p>Post-hoc tests comparing the post-tests revealed a significant main effect of groups, <emph>F</emph>(<reflink idref="bib1" id="ref107">1</reflink>, 198) = 38.2, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.16, but not time. A significant interaction effect of groups and time on scores was found, <emph>F</emph>(<reflink idref="bib1" id="ref108">1</reflink>, 198) = 5.99, <emph>p</emph> =.015, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.03. The results revealed that the difference between groups in post-test 1 continued in post-test 2. It is noteworthy that the interaction between groups and times on the scores suggested while HK tended to gain debugging skills during the PBL activities LK decreased debugging skills.</p> <hd id="AN0153455265-33">Attitudinal outcomes</hd> <p>A total of 189 students responded to the survey. The survey revealed that out of a possible range of 1–5 (1 strongly disagree to 5 strongly agree), students had neutral ranges of self-efficacy toward programming (<emph>M</emph> = 3.8), attitude toward CS (<emph>M</emph> = 3.3), and confidence in CT skills (<emph>M</emph> = 3.6).</p> <p>Considering the prior knowledge level, one-way ANOVAs were carried out to examine any differences in the survey results between students classified as LK and HK (See Table 4). Main effects of groups were confirmed in self-efficacy toward programming, <emph>F</emph>(<reflink idref="bib1" id="ref109">1</reflink>, 187) = 23.63, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.11, attitude toward CS, <emph>F</emph>(<reflink idref="bib1" id="ref110">1</reflink>, 187) = 15.59, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.08, and confidence in CT skills, <emph>F</emph>(<reflink idref="bib1" id="ref111">1</reflink>, 187) = 17.17, <emph>p</emph> <.001, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><msubsup><mi>η</mi><mrow><mi>p</mi></mrow><mn>2</mn></msubsup><mo>=</mo></mrow></math> </ephtml> 0.08. As Table 4 illustrates, HK students expressed significantly higher self-efficacy, more positive attitudes toward CS, and better confidence in CT skills than LK students.</p> <p>Table 4 Means and standard deviations of survey responses by high and low prior knowledge groups</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>Self-efficacy</p><p>toward programming</p></th><th align="left"><p>Attitude</p><p>toward CS</p></th><th align="left"><p>Confidence</p><p>in CT skills</p></th></tr></thead><tbody><tr><td align="left"><p>High (n = 88)</p></td><td align="left"><p>4.07 (.66)</p></td><td align="left"><p>3.61 (.91)</p></td><td align="left"><p>3.89 (.78)</p></td></tr><tr><td align="left"><p>Low (n = 101)</p></td><td align="left"><p>3.51 (.90)</p></td><td align="left"><p>3.06 (.98)</p></td><td align="left"><p>3.38 (.89)</p></td></tr><tr><td align="left"><p>Total (n = 189)</p></td><td align="left"><p>3.77 (.85)</p></td><td align="left"><p>3.31 (.98)</p></td><td align="left"><p>3.62 (.87)</p></td></tr></tbody></table> </ephtml> </p> <p>M(SD)</p> <p>Researchers also examined if learning gains (mean differences between pre-test and post-test 2 in overall scores, conceptual understanding, and debugging) after the learning activities showed any relationship with the students' attitudes. A Pearson product-moment correlation was estimated to determine the relationship between learning gains and the three attitudinal aspects. No significant correlation was found (see Table 5).</p> <p>Table 5 Correlation coefficient between learning gains and attitudinal aspects</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>Self-efficacy</p><p>toward programming</p></th><th align="left"><p>Attitude</p><p>toward CS</p></th><th align="left"><p>Confidence</p><p>in CT skills</p></th></tr></thead><tbody><tr><td align="left"><p>Overall scores</p></td><td align="left"><p>−.01</p></td><td align="left"><p>−.05</p></td><td align="left"><p>.00</p></td></tr><tr><td align="left"><p>CT concepts</p></td><td align="left"><p>−.09</p></td><td align="left"><p>−.04</p></td><td align="left"><p>−.04</p></td></tr><tr><td align="left"><p>Debugging</p></td><td align="left"><p>.08</p></td><td align="left"><p>−.03</p></td><td align="left"><p>.04</p></td></tr><tr><td align="left"><p>M</p></td><td align="left"><p>3.77</p></td><td align="left"><p>3.31</p></td><td align="left"><p>3.62</p></td></tr><tr><td align="left"><p>SD</p></td><td align="left"><p>.845</p></td><td align="left"><p>.982</p></td><td align="left"><p>.874</p></td></tr></tbody></table> </ephtml> </p> <p>Means and standard deviations for attitudinal aspects are presented in each column accordingly</p> <p>Considering the final learning performance, researchers further examined any relationship between post-test 2 (overall scores, conceptual understanding, and debugging) and students' attitudes. A Pearson product-moment correlation revealed significant correlations between all post-test 2 subcategories and the attitudinal aspects (see Table 6). This result suggests that students who showed better learning performance after the PBL activities expressed higher self-efficacy, attitude, and confidence toward CT.</p> <p>Table 6 Correlation coefficient between post-test II and attitudinal aspects</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Post-test II</p></th><th align="left"><p>Self-efficacy</p><p>toward programming</p></th><th align="left"><p>Attitude</p><p>toward CS</p></th><th align="left"><p>Confidence</p><p>in CT skills</p></th></tr></thead><tbody><tr><td align="left"><p>Overall scores</p></td><td align="left"><p>.41</p></td><td align="left"><p>.27</p></td><td align="left"><p>.40</p></td></tr><tr><td align="left"><p>CT concepts</p></td><td align="left"><p>.28</p></td><td align="left"><p>.17</p></td><td align="left"><p>.30</p></td></tr><tr><td align="left"><p>Debugging</p></td><td align="left"><p>.43</p></td><td align="left"><p>.31</p></td><td align="left"><p>.39</p></td></tr></tbody></table> </ephtml> </p> <p>All correlation coefficients are statistically significant at 5% level, where <emph>p</emph> values are smaller than.001</p> <p>Based on the results, multiple regressions using a stepwise method were carried out to predict the attitudinal aspects from the post-test 2 subcategories (see Table 7). Regarding self-efficacy and attitude, debugging skills explained a significant amount of the variance, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>F</mi><mfenced close=")" open="("><mn>1</mn><mo>,</mo><mn>187</mn></mfenced><mo>=</mo><mn>42.19</mn><mo>,</mo><mi>p</mi><mo><</mo><mo>.</mo><mn>001</mn><mo>,</mo><msup><mrow><mi>R</mi></mrow><mn>2</mn></msup><mo>=</mo><mo>.</mo><mn>18</mn></mrow></math> </ephtml> , <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>F</mi><mfenced close=")" open="("><mn>1</mn><mo>,</mo><mn>187</mn></mfenced><mo>=</mo><mn>19.77</mn><mo>,</mo><mi>p</mi><mo><</mo><mo>.</mo><mn>001</mn><mo>,</mo><msup><mrow><mi>R</mi></mrow><mn>2</mn></msup><mo>=</mo><mo>.</mo><mn>10</mn></mrow></math> </ephtml> , respectively. With regards to confidence in CT skills, overall post-test 2 scores explained a significant amount of the variance, <ephtml> <math xmlns="http://www.w3.org/1998/Math/MathML"><mrow><mi>F</mi><mfenced close=")" open="("><mn>1</mn><mo>,</mo><mn>187</mn></mfenced><mo>=</mo><mn>34.69</mn><mo>,</mo><mi>p</mi><mo><</mo><mo>.</mo><mn>001</mn><mo>,</mo><msup><mrow><mi>R</mi></mrow><mn>2</mn></msup><mo>=</mo></mrow></math> </ephtml> 0.16.</p> <p>Table 7 Regression models of attitudinal aspects</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left"><p>Dependent variable</p></th><th align="left"><p>Significant predictor</p></th><th align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><mi xmlns="">β</mi></math><inline-graphic href="11423_2021_10034_Article_IEq29.gif" /></p></th><th align="left"><p><italic>t</italic></p></th><th align="left"><p><italic>p</italic></p></th><th align="left"><p><math xmlns="http://www.w3.org/1998/Math/MathML"><msup xmlns=""><mrow><mi>R</mi></mrow><mn>2</mn></msup></math><inline-graphic href="11423_2021_10034_Article_IEq30.gif" /></p></th></tr></thead><tbody><tr><td align="left"><p>Self-efficacy toward programming</p></td><td align="left"><p>Debugging</p></td><td align="left"><p>.43</p></td><td char="." align="char"><p>6.50</p></td><td align="left"><p>.000</p></td><td align="left"><p>.18</p></td></tr><tr><td align="left"><p>Attitude toward CS</p></td><td align="left"><p>Debugging</p></td><td align="left"><p>.31</p></td><td char="." align="char"><p>4.45</p></td><td align="left"><p>.000</p></td><td align="left"><p>.10</p></td></tr><tr><td align="left"><p>Confidence in CT skills</p></td><td align="left"><p>Overall scores</p></td><td align="left"><p>.40</p></td><td char="." align="char"><p>5.89</p></td><td align="left"><p>.000</p></td><td align="left"><p>.16</p></td></tr></tbody></table> </ephtml> </p> <hd id="AN0153455265-34">Discussion</hd> <p>The purpose of this study was to examine how PBL might impact students' CT knowledge and interest at the elementary level. In this study, researchers examined (<reflink idref="bib1" id="ref112">1</reflink>) whether there were differences among CT concepts in terms of student achievement, (<reflink idref="bib2" id="ref113">2</reflink>) whether there were differences in students' CT knowledge, debugging skills, and attitudes towards CS according to their prior knowledge levels of CS, and (<reflink idref="bib3" id="ref114">3</reflink>) whether there is a relationship between elementary students' CT concepts/practices and their attitudes towards CS.</p> <hd id="AN0153455265-35">CT concepts</hd> <p></p> <hd id="AN0153455265-36">Event and parallelism: the most difficult concepts</hd> <p>Most students incorrectly answered questions on <emph>Event</emph> and <emph>Parallelism</emph> in the pre-test; only 29% (<emph>Event)</emph> and 37% (<emph>Parallelism</emph>) of students answered the questions correctly on the pre-test. Although the curriculum specifically targeted all six CT concepts, students still struggled to understand <emph>Event</emph> and <emph>Parallelism;</emph> 68% and 65% of students correctly answered questions on these two concepts on post-test 2. Between 71 and 84% of students were able to answer questions correctly for the other four CT concepts on post-test 2. Based on these results, it is apparent that <emph>Event</emph> and <emph>Parallelism</emph> were challenging concepts for novice programmers.</p> <p>Event is related to the execution of instructions that is triggered by user actions, such as "clicking a green flag," "pushing a key," or "touching an object" in Scratch. Parallelism is about the simultaneous execution of instructions, which allows the computer to process multiple instructions at the same time. Both concepts are closely related to when specific sets of instructions are carried out. To program an <emph>Event</emph>, students should decide how to detect when a specific event occurs and which sets of instructions to carry out (Franklin et al., [<reflink idref="bib24" id="ref115">24</reflink>]). Scratch provides various events, such as "when flag clicked," "when (space) key pressed," "when this sprite clicked," and "when backdrop switches to (backdrop1)." It may not be as challenging to understand the meaning of events and select one as necessary. However, there is another type of <emph>Event</emph> that triggers instructions by sending out a message: broadcast. By using a broadcast block, students can trigger an instruction separately (event) or multiple sets of instructions simultaneously (parallelism) which requires higher cognitive demands (Bogaerts, [<reflink idref="bib10" id="ref116">10</reflink>]). In this case, students should be able to define a message (create a new message for an event), determine when to send out the message (locate a broadcast block within instructions), and develop sets of instructions to be carried out after receiving the message (use when I receive (message1)). These processes require students to consider multiple aspects of programming and are much more complicated than using a simple event (i.e., "when flag clicked").</p> <p>Parallelism requires students to orchestrate simultaneous events triggering the multiple executions of instructions. In Scratch programming, each sprite (character or object) can have multiple sets of instructions triggering (or being triggered by) other events. To orchestrate these simultaneous events, students need to determine the flow of the instructions shared by the sprites. However, it is challenging for novice programmers to design events triggering specific instructions without making conflicts (Ba & Arora, [<reflink idref="bib4" id="ref117">4</reflink>]). If students fail to consider which tasks to execute with parallelism, a program can easily miss a task or trigger an unintended result. For the students in this study, parallelism was cognitively demanding, which required students to consider the broader scope of a program rather than focusing on a particular set of instructions.</p> <p>Results of the current study suggest that students quickly grasped the concepts of event and parallelism after the Scratch lessons. Furthermore, they gradually improved their debugging skills throughout the Scratch lessons and PBL sessions. These findings suggest that students might benefit from the PBL sessions where they practiced debugging while developing programs. Bogaerts ([<reflink idref="bib9" id="ref118">9</reflink>]) suggested that the concepts of parallelism need to be explicitly introduced with specific examples by drawing analogical connections between sprites. It is expected that students encountered parallelism issues while they developed Scratch projects during the PBL sessions. If students had explicit explanations or structured scaffolding regarding the issues, they might have better opportunities to learn and master parallelism (Nam et al., [<reflink idref="bib48" id="ref119">48</reflink>]).</p> <hd id="AN0153455265-37">Conditional and variable: logical expression and handling values</hd> <p>Conditional is a useful method to make a decision based on conditions. A conditional should be expressed in a logical way that its result must be either true or false. Based on the result, the computer carries out the following set of instructions accordingly, which determines the flow of the program. The current study suggests that students determined the flow of the program easily. They could read if-else coding blocks and expected results according to the conditions in the tests. However, they struggled to understand how to develop conditional statements using operators to compare two values. The tests assessed how well students understood conditional statements in particular settings. Thus, the results suggest that students needed more CT practices that provided them with opportunities to use operators in developing conditional statements.</p> <p>Variable is a crucial concept for students to utilize data (Byckling & Sajaniemi, [<reflink idref="bib15" id="ref120">15</reflink>]; Kwon, [<reflink idref="bib38" id="ref121">38</reflink>]). The current study suggests that students grasped basic concepts of variable relatively easily. Scratch may have made the concepts explicit by using natural languages describing the function of a variable (Kwon et al., [<reflink idref="bib39" id="ref122">39</reflink>]; Swidan et al., [<reflink idref="bib65" id="ref123">65</reflink>]). For example, "set (variable) to (value)" describes how to define a variable and initialize its value, and "change (variable) by (value)" expresses how to update the value of the variable. Students could determine when they needed to use variables and how they could change their values accordingly due to the intuitive features of Scratch.</p> <hd id="AN0153455265-38">Sequence and loop: fundamental to begin programming</hd> <p>Computers execute instructions in order; this is a fundamental concept students need to understand. However, novice programmers often have misunderstandings about the sequence of instructions (Pea, [<reflink idref="bib50" id="ref124">50</reflink>]). In contrast to the parallelism executing different sets of instructions at the same time, computers carry out instructions one by one in sequence within each set of instructions. If students (incorrectly) apply the "parallelism bug" to sequential instructions, they will encounter an unexpected error. A loop defines the set of instructions to repeat in sequence. Thus, the concept of sequences is a prerequisite of understanding the loop.</p> <p>In this study, students grasped the concepts of sequences and loops through the Scratch lessons. The interface of Scratch (listing blocks from top to bottom) enabled students to see the actual sequence of execution explicitly. Students also determined how to shorten repeated instructions by using a loop. The findings suggest that students could quickly pick up the basic concepts through the intuitive features of block-based programming (Scratch).</p> <hd id="AN0153455265-39">CT knowledge gain</hd> <p>The findings of this study suggest that a problem-based CS curriculum can effectively increase elementary students' understandings of CT. This curriculum enabled students to learn about basic CT concepts initially through Scratch lessons, and then by designing a larger-scale Scratch project that addressed a specific problem. While other CS curricular initiatives have been implemented at the secondary level (e.g., García-Peñalvo & Mendes, [<reflink idref="bib25" id="ref125">25</reflink>]; Lawanto et al., [<reflink idref="bib42" id="ref126">42</reflink>]), this research provides evidence suggesting that elementary students can engage in higher-level CT activities that involve designing, programming, and implementing more large-scale projects.</p> <p>Overall, students demonstrated a significant improvement in their CT knowledge from pre-test to post-test 1 where the introduction to block-based coding was implemented. Although the CT knowledge, in most concepts, was maintained at the same level afterward, students continuously improved event concepts (one of the most difficult concepts to master) from post-test 1 to post-test 2 where students developed Scratch projects as a PBL activity. In addition, the fact that there were no significant differences between post-test 1 and post-test 2 scores even though the PBL activities occurred approximately one or two months after the concept-oriented lessons can be considered a positive learning outcome. These positive student achievement results are consistent with other research examining the effects of other inquiry-based instruction on student achievement (Ravitz, [<reflink idref="bib52" id="ref127">52</reflink>]; Strobel & van Barneveld, [<reflink idref="bib63" id="ref128">63</reflink>]; Walker & Leary, [<reflink idref="bib69" id="ref129">69</reflink>]; Wirkala & Kuhn, [<reflink idref="bib73" id="ref130">73</reflink>]). More promising are the results suggesting that CS instruction that integrates a PBL model may have positive effects on more difficult concepts (i.e., event) in CS.</p> <p>The current study further examined whether students' prior knowledge affected learning gains in CT concepts and debugging skills. Overall, students who had higher prior knowledge outperformed those who had lower prior knowledge in every test. It is noteworthy, however, that lower prior knowledge students gained relatively more than the higher prior knowledge students from the pre-test to post-test 1. These results may suggest that components of the PBL curricular model may facilitate student engagement and thus have a positive effect on student achievement regardless of students' prior knowledge. These results also are consistent with the work by Belland et al. ([<reflink idref="bib6" id="ref131">6</reflink>]) who conducted a meta-analysis of research examining problem-centered instruction and scaffolding strategies and found that low-achieving students demonstrated achievement gains of over three standard deviations as compared to their higher-achieving peers. Similarly, Harris et al. ([<reflink idref="bib29" id="ref132">29</reflink>]) conducted a randomized controlled trial of a large scale PBL science curriculum. One of their conclusions from this study was that "...students in classrooms with higher concentrations of low-achieving students benefited from the curriculum as much as students in classrooms with lower concentrations of these students" (p. 1379). Thus, the potential for PBL and other inquiry-based instructional models appear to be promising for students having difficulty with mastering CS content. Of course, it is necessary to be cautious in interpreting the findings in consideration of alternative explanations about the phenomenon, such as a ceiling effect and regression toward the mean.</p> <p>Researchers also found unexpected results between two groups: high vs. low prior knowledge. Especially during the PBL activities, gaps in debugging skills between the groups increased. The researchers often observed students who felt more confident in Scratch took a more active role during coding and debugging tasks. Students were also allowed to take assigned roles in a group according to their preferences and abilities, which may have unintentionally given higher prior knowledge students more opportunities to engage in tasks related to debugging. Researchers observed that high-ability students dominated group interactions and exerted a more significant influence on the decisions, which may have resulted in better learning outcomes (Dembo & McAuliffe, [<reflink idref="bib19" id="ref133">19</reflink>]; Webb, [<reflink idref="bib70" id="ref134">70</reflink>]). Regarding this phenomenon, we call for further studies with controlled research settings.</p> <hd id="AN0153455265-40">Interests about CT</hd> <p>The results of the student attitude survey suggested that students had positive self-efficacy and attitudes toward CS as well as confidence in CT skills. However, based on the responses, students' attitudes towards CS did not appear to be significantly impacted by the PBL curriculum. In addition, the fact that student attitudes and self-efficacy towards CS were more positive for high prior knowledge students suggests that the PBL unit may not have produced optimal results with regards to engaging the interests of students who were not as successful with CS content. The results might due to different performances in Scratch projects according to students' varying prior knowledge; advanced students took charge of features requiring high programming demands while novices took care of basic functions (Mishra et al., [<reflink idref="bib47" id="ref135">47</reflink>]).</p> <p>Results of the current study also suggest that debugging skills may be a predictor of student attitudes and self-efficacy. As discussed in the previous section, PBL activities provided students with learning opportunities to practice debugging skills, such as understanding the meaning of codes, identifying code segments involved in the problem (bug), determining the cause of the problem, testing hypotheses, and confirming a solution. Findings suggest that students who had better debugging skills expressed higher self-efficacy and more positive attitudes toward CS. These findings are consistent with other studies in that students who experience and engage in successful problem-solving have increased beliefs in their abilities, which can lead to higher self-efficacy (Schunk & Gunn, [<reflink idref="bib58" id="ref136">58</reflink>]).</p> <p>Considering the results, future iterations of this unit may need to consider more engaging problems that better resonate with all students. Previous research suggests that incorporating problem-solving skills and re-orienting CS curricula around relevant and meaningful problem-solving can help broaden participation in computing (e.g., Denner et al., [<reflink idref="bib20" id="ref137">20</reflink>]; Fields et al., [<reflink idref="bib23" id="ref138">23</reflink>]). The integration of relevant, real-world problem-solving can help broaden participation by shifting students' understanding of the type of work that is done in CS and the value of that work (Khan & Luxton-Reilly, [<reflink idref="bib36" id="ref139">36</reflink>]). In other words, when CS is centered around relevant, real-world problems, more students may be able to make connections between CS content and their own lives.</p> <hd id="AN0153455265-41">Conclusion</hd> <p>Though expanding CS to younger populations has become more prevalent, there is a dearth of research with regard to the proper pedagogy to introduce CS curricula to young learners. This study implemented a problem-based CS curriculum with elementary students and yielded results that suggest the PBL curricular model may effectively increase elementary students' understanding of CT. Students demonstrated significant improvement of CT concepts after the Scratch lessons and retained their learning gains through the completion of PBL activities in which they applied CT skills towards a socially relevant problem.</p> <p>In addition, results of this study suggest different learning trajectories among groups with different prior knowledge. Students with higher prior knowledge in CT performed better on all of the content tests. The differences between high- and low-prior knowledge students in the conceptual understandings of the tests were narrowed down after the Scratch lessons, but the differences increased in the part of the tests measuring debugging skills. These results may have been influenced by unequal participation in the PBL activities. It was observed that the students demonstrated moderate levels of self-efficacy, attitudes towards CS, and confidence in CT, but that the scores on the attitudinal survey were consistently higher among students with higher prior knowledge. These results suggest the necessity of tailored learning activities according to learners' prior knowledge and instructional guidance for collaborative learning. This study calls for follow-up research investigating the effects of CS education for early childhood in more various educational settings.</p> <hd id="AN0153455265-42">Acknowledgements</hd> <p>This material is based upon work supported by the Google Computer Science Education Research Grant. Thanks to the research team members for their contributions to this work, and to the teachers and students who participated in this study.</p> <hd id="AN0153455265-43">Declarations</hd> <p></p> <hd id="AN0153455265-44">Conflict of interest</hd> <p>The authors declare that he has no conflict of interest.</p> <hd id="AN0153455265-45">Ethical approval</hd> <p>All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.</p> <hd id="AN0153455265-46">Informed consent</hd> <p>Researchers obtained a student assent form and a parent consent form from the participants as well as their parents/guardians before initiating research activities.</p> <hd id="AN0153455265-47">Appendix 1: Student Attitude Survey</hd> <p></p> <hd id="AN0153455265-48">Self-efficacy toward programming</hd> <p></p> <ulist> <item> I can write a program that gives me the results I want.</item> <p></p> <item> I am good at programming.</item> <p></p> <item> I can find errors in my program and fix them.</item> <p></p> <item> I can read a program written by others.</item> </ulist> <hd id="AN0153455265-49">Attitude toward CS</hd> <p></p> <ulist> <item> I think programming is important to know.</item> <p></p> <item> I think programming is frustrating.</item> <p></p> <item> I think programming is boring.</item> <p></p> <item> I think programming could help solve problems in my everyday life.</item> <p></p> <item> I want to continue to learn programming in the future.</item> <p></p> <item> I enjoy writing programs.</item> </ulist> <hd id="AN0153455265-50">Confidence in CT skills</hd> <p></p> <ulist> <item> I can break down a problem so I can find a solution.</item> <p></p> <item> I can find a pattern in a series of events or numbers.</item> <p></p> <item> I am good at creating plans to solve complex problems.</item> <p></p> <item> I can create a step-by-step solution to a problem.</item> <p></p> <item> I can apply a plan to solve a problem.</item> <p></p> <item> I can find general rules from a problem I solved that I can use for other problems.</item> </ulist> <hd id="AN0153455265-51">Publisher's Note</hd> <p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p> <ref id="AN0153455265-52"> <title> References </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> Acuña SR, García Rodicio H, Sánchez E. Fostering active processing of instructional explanations of learners with high and low prior knowledge. European Journal of Psychology of Education. 2011; 26: 435-452. 10.1007/s10212-010-0049-y</bibtext> </blist> <blist> <bibl id="bib2" idref="ref2" type="bt">2</bibl> <bibtext> Armoni M. Teaching CS in kindergarten: How early can the pipeline begin?. ACM Inroads. 2012; 3; 4: 18-19. 10.1145/2381083.2381091</bibtext> </blist> <blist> <bibl id="bib3" idref="ref29" type="bt">3</bibl> <bibtext> Atkinson RK, Derry SJ, Renkl A, Wortham D. Learning from examples: Instructional principles from the worked examples research. Review of Educational Research. 2000; 70: 181-214. 10.3102/00346543070002181</bibtext> </blist> <blist> <bibl id="bib4" idref="ref30" type="bt">4</bibl> <bibtext> Ba, T. N, & Arora, R. (2018). Towards developing a repository of logical errors observed in parallel code for teaching code correctness. Paper presented at the 2018 IEEE/ACM Workshop on Education for High-Performance Computing (EduHPC).</bibtext> </blist> <blist> <bibl id="bib5" idref="ref50" type="bt">5</bibl> <bibtext> Barrows HS. Problem-based learning in medicine and beyond: A brief overview. New Directions for Teaching and Learning. 1996; 68: 3-12. 10.1002/tl.37219966804</bibtext> </blist> <blist> <bibl id="bib6" idref="ref47" type="bt">6</bibl> <bibtext> Belland BR, Walker AE, Kim NJ. A Bayesian network meta-analysis to synthesize the influence of contexts of scaffolding use on cognitive outcomes in STEM education. Review of Educational Research. 2017; 87: 1042-1081. 10.3102/0034654317723009</bibtext> </blist> <blist> <bibl id="bib7" idref="ref33" type="bt">7</bibl> <bibtext> Bers MU, Flannery L, Kazakoff ER, Sullivan A. Computational thinking and tinkering: Exploration of an early childhood robotics curriculum. Computers & Education. 2014; 72: 145-157. 10.1016/j.compedu.2013.10.020</bibtext> </blist> <blist> <bibl id="bib8" idref="ref35" type="bt">8</bibl> <bibtext> Bogaerts, S. A. (2013). Hands-on exploration of parallelism for absolute beginners with Scratch. Paper presented at the Proceedings of the 2013 IEEE 27th International Symposium on Parallel and Distributed Processing Workshops and PhD Forum. https://doi.org/10.1109/IPDPSW.2013.63</bibtext> </blist> <blist> <bibl id="bib9" idref="ref118" type="bt">9</bibl> <bibtext> Bogaerts SAPrasad SK, Gupta A, Rosenberg AL, Sussman A, Weems CC. Hands-on parallelism with no prerequisites and little time using Scratch. Topics in parallel and distributed computing. 2015; Morgan Kaufmann: 11-24. 10.1016/B978-0-12-803899-4.00002-X</bibtext> </blist> <blist> <bibtext> Bogaerts SA. One step at a time: Parallelism in an introductory programming course. Journal of Parallel and Distributed Computing. 2017; 105: 4-17. 10.1016/j.jpdc.2016.12.024</bibtext> </blist> <blist> <bibtext> Brackmann, C, Barone, D, Casali, A, Boucinha, R, & Muñoz-Hernandez, S. (2016). Computational thinking: Panorama of the Americas. Paper presented at the 2016 International Symposium on Computers in Education (SIIE).</bibtext> </blist> <blist> <bibtext> Brennan, K, & Resnick, M. (2012). New frameworks for studying and assessing the development of computational thinking. Paper presented at the Proceedings of the 2012 annual meeting of the American Educational Research Association, Vancouver, Canada.</bibtext> </blist> <blist> <bibtext> Bruce KB, Danyluk AP, Murtagh TP. Event-driven programming is simple enough for CS1. ACM SIGCSE Bulletin. 2001; 33; 3: 1-4. 10.1145/507758.377440</bibtext> </blist> <blist> <bibtext> Brush, T, Ottenbreit-Leftwich, A, Kwon, K, & Karlin, M. (2020). Implementing socially relevant problem-based computer science curriculum at the elementary level: students' computer science knowledge and teachers' implementation needs. Journal of Computers in Mathematics and Science Teaching, 39(2), 109–123</bibtext> </blist> <blist> <bibtext> Byckling P, Sajaniemi J. Roles of variables and programming skills improvement. ACM SIGCSE Bulletin. 2006; 38; 1: 413-417. 10.1145/1124706.1121470</bibtext> </blist> <blist> <bibtext> Cetin I, Ozden MY. Development of computer programming attitude scale for university students. Computer Applications in Engineering Education. 2015; 23: 667-672. 10.1002/cae.21639</bibtext> </blist> <blist> <bibtext> Chen G, Shen J, Barth-Cohen L, Jiang S, Huang X, Eltoukhy M. Assessing elementary students' computational thinking in everyday reasoning and robotics programming. Computers & Education. 2017; 109: 162-175. 10.1016/j.compedu.2017.03.001</bibtext> </blist> <blist> <bibtext> CSTA. (2017). CSTA K-12 Computer science standards. Retrieved from https://<ulink href="http://www.csteachers.org/page/standards">www.csteachers.org/page/standards</ulink></bibtext> </blist> <blist> <bibtext> Dembo MH, McAuliffe TJ. Effects of perceived ability and grade status on social interaction and influence in cooperative groups. Journal of Educational Psychology. 1987; 79: 415-423. 10.1037/0022-0663.79.4.415</bibtext> </blist> <blist> <bibtext> Denner J, Martinez J, Lyon LA. Computing for the social good: Engaging Latino/a students in K-12. ACM SIGCAS Computers and Society. 2015; 45; 2: 31-32. 10.1145/2809957.2809964</bibtext> </blist> <blist> <bibtext> Denner J, Werner L, Campe S, Ortiz E. Pair programming: Under what conditions is it advantageous for middle school students?. Journal of Research on Technology in Education. 2014; 46: 277-296. 10.1080/15391523.2014.888272</bibtext> </blist> <blist> <bibtext> Fessakis G, Gouli E, Mavroudi E. Problem solving by 5–6 years old kindergarten children in a computer programming environment: A case study. Computers & Education. 2013; 63: 87-97. 10.1016/j.compedu.2012.11.016</bibtext> </blist> <blist> <bibtext> Fields DA, Kafai Y, Nakajima T, Goode J, Margolis J. Putting making into high school computer science classrooms: Promoting equity in teaching and learning with electronic textiles in exploring computer science. Equity & Excellence in Education. 2018; 51: 21-35. 10.1080/10665684.2018.1436998</bibtext> </blist> <blist> <bibtext> Franklin, D, Conrad, P, Boe, B, Nilsen, K, Hill, C, Len, M, Dreschler, G, Aldana, G, Almeida-Tanaka, P, Kiefer, B, & Laird, C. (2013). Assessment of computer science learning in a scratch-based outreach program. Paper presented at the Proceeding of the 44th ACM technical symposium on Computer science education, Denver, Colorado, USA. https://doi.org/10.1145/2445196.2445304</bibtext> </blist> <blist> <bibtext> García-Peñalvo FJ, Mendes AJ. Exploring the computational thinking effects in pre-university education. Computers in Human Behavior. 2018; 80: 407-411. 10.1016/j.chb.2017.12.005</bibtext> </blist> <blist> <bibtext> Gaudiello I, Zibetti E. Using control heuristics as a means to explore the educational potential of robotics kits. Themes in Science and Technology Education. 2013; 6; 1: 15-28</bibtext> </blist> <blist> <bibtext> Gibson, J. P. (2008). Formal methods: Never too young to start. Paper presented at the Formal Methods in Computer Science Education, Budapest, Hungary.</bibtext> </blist> <blist> <bibtext> Google, & Gallup. (2015). Searching for computer science: Access and Barriers in U.S. K-12 Education. Retrieved from https://goo.gl/oX311J</bibtext> </blist> <blist> <bibtext> Harris CJ, Penuel WR, D'Angelo CM, DeBarger AH, Gallagher LP, Kennedy CA. Impact of project-based curriculum materials on student learning in science: Results of a randomized controlled trial. Journal of Research in Science Teaching. 2015; 52: 1362-1385. 10.1002/tea.21263</bibtext> </blist> <blist> <bibtext> Hsu T-C, Chang S-C, Hung Y-T. How to learn and how to teach computational thinking: Suggestions based on a review of the literature. Computers & Education. 2018; 126: 296-310. 10.1016/j.compedu.2018.07.004</bibtext> </blist> <blist> <bibtext> Indiana Department of Education. (2016). Sixth–Eighth grade computer science standards. Retrieved from https://<ulink href="http://www.doe.in.gov/sites/default/files/standards/indiana-6-8-computer-science-standards-2016-060717.pdf">www.doe.in.gov/sites/default/files/standards/indiana-6-8-computer-science-standards-2016-060717.pdf</ulink></bibtext> </blist> <blist> <bibtext> Iskrenovic-Momcilovic O. Pair programming with scratch. Education and Information Technologies. 2019. 10.1007/s10639-019-09905-3</bibtext> </blist> <blist> <bibtext> Israel M, Pearson JN, Tapia T, Wherfel QM, Reese G. Supporting all learners in school-wide computational thinking: A cross-case qualitative analysis. Computers & Education. 2015; 82: 263-279. 10.1016/j.compedu.2014.11.022</bibtext> </blist> <blist> <bibtext> ISTE, & CSTA. (2011). Operational definition of computational thinking for K-12 education. Retrieved from https://id.iste.org/docs/ct-documents/computational-thinking-operational-definition-flyer.pdf</bibtext> </blist> <blist> <bibtext> K-12 Computer Science Framework Steering Committee. (2016). K-12 Computer Science Framework. Retrieved from https://dl.acm.org/doi/book/10.1145/3079760</bibtext> </blist> <blist> <bibtext> Khan, N. Z, & Luxton-Reilly, A. (2016). Is computing for social good the solution to closing the gender gap in computer science? Paper presented at the Proceedings of the Australasian Computer Science Week Multiconference, Canberra, Australia. https://doi.org/10.1145/2843043.2843069</bibtext> </blist> <blist> <bibtext> Kim B, Kim T, Kim J. Paper-and-pencil programming strategy toward computational thinking for non-majors: Design your solution. Journal of Educational Computing Research. 2013; 49: 437-459. 10.2190/EC.49.4.b</bibtext> </blist> <blist> <bibtext> Kwon K. Student's misconception of programming reflected on problem-solving plans. International Journal of Computer Science Education in Schools. 2017; 1; 4: 14-24. 10.21585/ijcses.v1i4.19</bibtext> </blist> <blist> <bibtext> Kwon K, Lee SJ, Chung J. Computational concepts reflected on Scratch programs. International Journal of Computer Science Education in Schools. 2018; 2; 3: 1-15. 10.21585/ijcses.v2i3.33</bibtext> </blist> <blist> <bibtext> Kwon, K, Leftwich, A, Brush, T, & Jeon, M. (2020). Effects of problem-based learning curriculum for computer science education in an elementary school. Paper presented at the AERA Annual Meeting, San Francisco, CA. <ulink href="http://tinyurl.com/wl5lak9">http://tinyurl.com/wl5lak9</ulink> (Conference Canceled)</bibtext> </blist> <blist> <bibtext> Lambert L, Guiffre H. Computer science outreach in an elementary school. Journal of Computing Sciences in Colleges. 2009; 24: 118-124</bibtext> </blist> <blist> <bibtext> Lawanto K, Close K, Ames C, Brasiel SRich PJ, Hodges CB. Exploring strengths and weaknesses in middle school students' computational thinking in Scratch. Emerging research, practice, and policy on computational thinking. 2017; Springer International Publishing: 307-326. 10.1007/978-3-319-52691-1_19</bibtext> </blist> <blist> <bibtext> Lee I, Martin F, Denner J, Coulter B, Allan W, Erickson J. Computational thinking for youth in practice. ACM Inroads. 2011; 2; 1: 32-37. 10.1145/1929887.1929902</bibtext> </blist> <blist> <bibtext> Lewis CM. The importance of students' attention to program state: A case study of debugging behavior. Paper Presented at the the Ninth Annual International Conference on International Computing Education Research, Auckland, New Zealand. 2012. 10.1145/2361276.2361301</bibtext> </blist> <blist> <bibtext> Lu, J. J, & Fletcher, G. H. L. (2009). Thinking about computational thinking. Paper presented at the Proceedings of the 40th ACM technical symposium on Computer science education, Chattanooga, TN, USA. https://doi.org/10.1145/1508865.1508959</bibtext> </blist> <blist> <bibtext> McDowell C, Werner L, Bullock HE, Fernald J. Pair programming improves student retention, confidence, and program quality. Communication of the ACM. 2006; 49; 8: 90-95. 10.1145/1145287.1145293</bibtext> </blist> <blist> <bibtext> Mishra, S, Balan, S, Iyer, S, & Murthy, S. (2014). Effect of a 2-week scratch intervention in CS1 on learners with varying prior knowledge. Paper presented at the the 2014 conference on Innovation & technology in computer science education, Uppsala, Sweden. https://doi.org/10.1145/2591708.2591733</bibtext> </blist> <blist> <bibtext> Nam, D, Kim, Y, & Lee, T. (2010). The effects of scaffolding-based courseware for the Scratch programming learning on student problem solving skill. Paper presented at the the 18th International Conference on Computers in Education, Putrajaya, Malaysia.</bibtext> </blist> <blist> <bibtext> National Research Council. (2012). Report of a workshop on science, technology, engineering, and mathematics (STEM) workforce needs for the U.S. Department of Defense and the U.S. Defense Industrial Base: The National Academies Press.</bibtext> </blist> <blist> <bibtext> Pea RD. Language-independent conceptual "bugs" in novice programming. Journal of Educational Computing Research. 1986; 2: 25-36. 10.2190/689T-1R2A-X4W4-29J2</bibtext> </blist> <blist> <bibtext> Ramalingam V, Wiedenbeck S. Development and validation of scores on a computer programming self-efficacy scale and group analyses of novice programmer self-efficacy. Journal of Educational Computing Research. 1998; 19: 367-381. 10.2190/c670-y3c8-ltj1-ct3p</bibtext> </blist> <blist> <bibtext> Ravitz, J. (2009). Introduction: Summarizing findings and looking ahead to a new generation of PBL research. Interdisciplinary Journal of Problem-based Learning.https://doi.org/10.7771/1541-5015.1088</bibtext> </blist> <blist> <bibtext> Rich, K. M, Strickland, C, Binkowski, T. A, & Franklin, D. (2019). A K-8 debugging learning trajectory derived from research literature. Paper presented at the the 50th ACM Technical Symposium on Computer Science Education.</bibtext> </blist> <blist> <bibtext> Richards, E, & Terkanian, D. (2013). Occupational employment projections to 2022. Monthly Labor Review. Retrieved from https://<ulink href="http://www.bls.gov/opub/mlr/2013/article/occupational-employment-projections-to-2022.htm">www.bls.gov/opub/mlr/2013/article/occupational-employment-projections-to-2022.htm</ulink></bibtext> </blist> <blist> <bibtext> Riley DD, Hunt KA. Computational thinking for the modern problem solver. 2014; CRC Press. 10.1201/b16688</bibtext> </blist> <blist> <bibtext> Sáez-López J-M, Román-González M, Vázquez-Cano E. Visual programming languages integrated across the curriculum in elementary school: A two year case study using "Scratch" in five schools. Computers & Education. 2016; 97: 129-141. 10.1016/j.compedu.2016.03.003</bibtext> </blist> <blist> <bibtext> Savery JR. Overview of problem-based learning: Definitions and distinctions. Interdisciplinary Journal of Problem-Based Learning. 2006; 1; 1: 9-20. 10.7771/1541-5015.1002</bibtext> </blist> <blist> <bibtext> Schunk DH, Gunn TP. Self-efficacy and skill development: Influence of task strategies and attributions. The Journal of Educational Research. 1986; 79: 238-244. 10.1080/00220671.1986.10885684</bibtext> </blist> <blist> <bibtext> Selby, C. C, & Woollard, J. (2014). Computational thinking: The developing definition. Paper presented at the Special Interest Group on Computer Science Education, Canterbury.</bibtext> </blist> <blist> <bibtext> Shi N, Cui W, Zhang P, Sun X. Evaluating the effectiveness roles of variables in the novice programmers learning. Journal of Educational Computing Research. 2017; 56: 181-201. 10.1177/0735633117707312</bibtext> </blist> <blist> <bibtext> Shute VJ, Sun C, Asbell-Clarke J. Demystifying computational thinking. Educational Research Review. 2017; 22: 142-158. 10.1016/j.edurev.2017.09.003</bibtext> </blist> <blist> <bibtext> Statter D, Armoni M. Teaching abstraction in computer science to 7th grade students. ACM Transction on Computing Education. 2020; 20; 1: 1-37. 10.1145/3372143</bibtext> </blist> <blist> <bibtext> Strobel J, van Barneveld A. When is PBL more effective? A meta-synthesis of meta-analyses comparing PBL to conventional classrooms. Interdisciplinary Journal of Problem-Based Learning. 2009; 3; 1: 44-58. 10.7771/1541-5015.1046</bibtext> </blist> <blist> <bibtext> Sullivan FR, Heffernan J. Robotic construction kits as computational manipulatives for learning in the STEM disciplines. Journal of Research on Technology in Education. 2016; 48: 105-128. 10.1080/15391523.2016.1146563</bibtext> </blist> <blist> <bibtext> Swidan, A, Serebrenik, A, & Hermans, F. (2017). How do scratch programmers name variables and procedures? Paper presented at the 2017 IEEE 17th International Working Conference on Source Code Analysis and Manipulation (SCAM).</bibtext> </blist> <blist> <bibtext> Taylor K, Baek Y. Collaborative robotics, more than just working in groups. Journal of Educational Computing Research. 2018; 56: 979-1004. 10.1177/0735633117731382</bibtext> </blist> <blist> <bibtext> The White House. (2016). Fact sheet: New progress and momentum in support of president obama's computer science for all initiative. Retrieved from https://obamawhitehouse.archives.gov/the-press-office/2016/09/14/fact-sheet-new-progress-and-momentum-support-president-obamas-computer</bibtext> </blist> <blist> <bibtext> Vakil S. Ethics, identity, and political vision: Toward a justice-centered approach to equity in computer science education. Harvard Educational Review. 2018; 88; 1: 26-52. 10.17763/1943-5045-88.1.26</bibtext> </blist> <blist> <bibtext> Walker A, Leary H. A problem based learning meta analysis: Differences across problem types, implementation types, disciplines, and assessment levels. Interdisciplinary Journal of Problem-Based Learning. 2009. 10.7771/1541-5015.1061</bibtext> </blist> <blist> <bibtext> Webb NM. Peer interaction and learning in cooperative small groups. Journal of Educational Psychology. 1982; 74; 5: 642-655. 10.1037/0022-0663.74.5.642</bibtext> </blist> <blist> <bibtext> Weintrop D, Beheshti E, Horn M, Orton K, Jona K, Trouille L, Wilensky U. Defining computational thinking for mathematics and science classrooms. Journal of Science Education and Technology. 2016; 25: 127-147. 10.1007/s10956-015-9581-5</bibtext> </blist> <blist> <bibtext> Wing JM. Computational thinking. Communications of the ACM. 2006; 49; 3: 33-35. 10.1145/1118178.1118215</bibtext> </blist> <blist> <bibtext> Wirkala C, Kuhn D. Problem-based learning in K–12 education: Is it effective and how does it achieve its effects?. American Educational Research Journal. 2011; 48: 1157-1186. 10.3102/0002831211419491</bibtext> </blist> <blist> <bibtext> Zelazo PD, Carter A, Reznick JS, Frye D. Early development of executive function: A problem-solving framework. Review of General Psychology. 1997; 1: 198-226. 10.1037/1089-2680.1.2.198</bibtext> </blist> <blist> <bibtext> Zhong B, Wang Q, Chen J. The impact of social factors on pair programming in a primary school. Computers in Human Behavior. 2016; 64: 423-431. 10.1016/j.chb.2016.07.017</bibtext> </blist> </ref> <aug> <p>By Kyungbin Kwon; Anne T. Ottenbreit-Leftwich; Thomas A. Brush; Minji Jeon and Ge Yan</p> <p>Reported by Author; Author; Author; Author; Author</p> <p></p> <p>Kyungbin Kwon is an Associate Professor of Instructional Systems Technology at Indiana University – Bloomington. Dr. Kwon's research focuses on facilitating positive interactions among students in contexts of Computer-Supported Collaborative Learning (CSCL) and designing effective instructions for computational thinking (CT).</p> <p>Anne T. Ottenbreit-Leftwich is an Associate Professor of Instructional Systems Technology within the School of Education and an Adjunct Professor of Computer Science at Indiana University – Bloomington. Dr. Leftwich's expertise lies in the areas of the design of curriculum resources, the use of technology to support pre-service teacher training, and development/implementation of professional development for teachers and teacher educators.</p> <p>Thomas A. Brush is the Barbara B. Jacobs Chair in Education and Technology and a Professor of Instructional Systems Technology within the School of Education at Indiana University, Bloomington. Dr. Brush's research interests focus on developing methods and strategies to promote inquiry-oriented learning, particularly with more open-ended instruction.</p> <p>Minji Jeon is a doctoral student in Instructional Systems Technology at Indiana University. Her research interest lies in integrating computational thinking and AI education in K-12 as well as enhancing collaborative argumentations through Computer Supported-Collaborative Learning (CSCL).</p> <p>Ge Yan is a doctorate student in the School of Education, Indiana University. His research interests focus on computer science education for novice learners. He holds other positions, including adjunct instructor and director of IT systems at Kelley School of Business. Ge has a BS and an MS degree in Computer Science from Temple University and an MBA from the University of Minnesota.</p> </aug> <nolink nlid="nl1" bibid="bib30" firstref="ref3"></nolink> <nolink nlid="nl2" bibid="bib67" firstref="ref4"></nolink> <nolink nlid="nl3" bibid="bib49" firstref="ref5"></nolink> <nolink nlid="nl4" bibid="bib54" firstref="ref6"></nolink> <nolink nlid="nl5" bibid="bib28" firstref="ref7"></nolink> <nolink nlid="nl6" bibid="bib68" firstref="ref8"></nolink> <nolink nlid="nl7" bibid="bib35" firstref="ref9"></nolink> <nolink nlid="nl8" bibid="bib18" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib71" firstref="ref12"></nolink> <nolink nlid="nl10" bibid="bib72" firstref="ref13"></nolink> <nolink nlid="nl11" bibid="bib14" firstref="ref15"></nolink> <nolink nlid="nl12" bibid="bib62" firstref="ref17"></nolink> <nolink nlid="nl13" bibid="bib33" firstref="ref18"></nolink> <nolink nlid="nl14" bibid="bib40" firstref="ref19"></nolink> <nolink nlid="nl15" bibid="bib45" firstref="ref21"></nolink> <nolink nlid="nl16" bibid="bib55" firstref="ref22"></nolink> <nolink nlid="nl17" bibid="bib11" firstref="ref24"></nolink> <nolink nlid="nl18" bibid="bib59" firstref="ref25"></nolink> <nolink nlid="nl19" bibid="bib12" firstref="ref31"></nolink> <nolink nlid="nl20" bibid="bib74" firstref="ref32"></nolink> <nolink nlid="nl21" bibid="bib13" firstref="ref34"></nolink> <nolink nlid="nl22" bibid="bib10" firstref="ref36"></nolink> <nolink nlid="nl23" bibid="bib39" firstref="ref37"></nolink> <nolink nlid="nl24" bibid="bib60" firstref="ref38"></nolink> <nolink nlid="nl25" bibid="bib43" firstref="ref40"></nolink> <nolink nlid="nl26" bibid="bib61" firstref="ref41"></nolink> <nolink nlid="nl27" bibid="bib37" firstref="ref42"></nolink> <nolink nlid="nl28" bibid="bib53" firstref="ref45"></nolink> <nolink nlid="nl29" bibid="bib44" firstref="ref46"></nolink> <nolink nlid="nl30" bibid="bib63" firstref="ref48"></nolink> <nolink nlid="nl31" bibid="bib57" firstref="ref49"></nolink> <nolink nlid="nl32" bibid="bib41" firstref="ref52"></nolink> <nolink nlid="nl33" bibid="bib56" firstref="ref53"></nolink> <nolink nlid="nl34" bibid="bib17" firstref="ref54"></nolink> <nolink nlid="nl35" bibid="bib27" firstref="ref57"></nolink> <nolink nlid="nl36" bibid="bib22" firstref="ref58"></nolink> <nolink nlid="nl37" bibid="bib64" firstref="ref60"></nolink> <nolink nlid="nl38" bibid="bib26" firstref="ref61"></nolink> <nolink nlid="nl39" bibid="bib32" firstref="ref62"></nolink> <nolink nlid="nl40" bibid="bib46" firstref="ref63"></nolink> <nolink nlid="nl41" bibid="bib21" firstref="ref64"></nolink> <nolink nlid="nl42" bibid="bib66" firstref="ref65"></nolink> <nolink nlid="nl43" bibid="bib75" firstref="ref66"></nolink> <nolink nlid="nl44" bibid="bib19" firstref="ref67"></nolink> <nolink nlid="nl45" bibid="bib31" firstref="ref71"></nolink> <nolink nlid="nl46" bibid="bib34" firstref="ref80"></nolink> <nolink nlid="nl47" bibid="bib16" firstref="ref86"></nolink> <nolink nlid="nl48" bibid="bib51" firstref="ref87"></nolink> <nolink nlid="nl49" bibid="bib24" firstref="ref115"></nolink> <nolink nlid="nl50" bibid="bib48" firstref="ref119"></nolink> <nolink nlid="nl51" bibid="bib15" firstref="ref120"></nolink> <nolink nlid="nl52" bibid="bib38" firstref="ref121"></nolink> <nolink nlid="nl53" bibid="bib65" firstref="ref123"></nolink> <nolink nlid="nl54" bibid="bib50" firstref="ref124"></nolink> <nolink nlid="nl55" bibid="bib25" firstref="ref125"></nolink> <nolink nlid="nl56" bibid="bib42" firstref="ref126"></nolink> <nolink nlid="nl57" bibid="bib52" firstref="ref127"></nolink> <nolink nlid="nl58" bibid="bib69" firstref="ref129"></nolink> <nolink nlid="nl59" bibid="bib73" firstref="ref130"></nolink> <nolink nlid="nl60" bibid="bib29" firstref="ref132"></nolink> <nolink nlid="nl61" bibid="bib70" firstref="ref134"></nolink> <nolink nlid="nl62" bibid="bib47" firstref="ref135"></nolink> <nolink nlid="nl63" bibid="bib58" firstref="ref136"></nolink> <nolink nlid="nl64" bibid="bib20" firstref="ref137"></nolink> <nolink nlid="nl65" bibid="bib23" firstref="ref138"></nolink> <nolink nlid="nl66" bibid="bib36" firstref="ref139"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1316649
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Integration of Problem-Based Learning in Elementary Computer Science Education: Effects on Computational Thinking and Attitudes
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Kwon%2C+Kyungbin%22">Kwon, Kyungbin</searchLink> (ORCID <externalLink term="http://orcid.org/0000-0001-8646-0144">0000-0001-8646-0144</externalLink>)<br /><searchLink fieldCode="AR" term="%22Ottenbreit-Leftwich%2C+Anne+T%2E%22">Ottenbreit-Leftwich, Anne T.</searchLink><br /><searchLink fieldCode="AR" term="%22Brush%2C+Thomas+A%2E%22">Brush, Thomas A.</searchLink><br /><searchLink fieldCode="AR" term="%22Jeon%2C+Minji%22">Jeon, Minji</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-0301-2221">0000-0002-0301-2221</externalLink>)<br /><searchLink fieldCode="AR" term="%22Yan%2C+Ge%22">Yan, Ge</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Educational+Technology+Research+and+Development%22"><i>Educational Technology Research and Development</i></searchLink>. Oct 2021 69(5):2761-2787.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 27
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2021
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research<br />Tests/Questionnaires
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Elementary+Education%22">Elementary Education</searchLink><br /><searchLink fieldCode="EL" term="%22Grade+6%22">Grade 6</searchLink><br /><searchLink fieldCode="EL" term="%22Intermediate+Grades%22">Intermediate Grades</searchLink><br /><searchLink fieldCode="EL" term="%22Middle+Schools%22">Middle Schools</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Problem+Based+Learning%22">Problem Based Learning</searchLink><br /><searchLink fieldCode="DE" term="%22Elementary+School+Curriculum%22">Elementary School Curriculum</searchLink><br /><searchLink fieldCode="DE" term="%22Elementary+School+Students%22">Elementary School Students</searchLink><br /><searchLink fieldCode="DE" term="%22Student+Attitudes%22">Student Attitudes</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+Science+Education%22">Computer Science Education</searchLink><br /><searchLink fieldCode="DE" term="%22Programming%22">Programming</searchLink><br /><searchLink fieldCode="DE" term="%22Thinking+Skills%22">Thinking Skills</searchLink><br /><searchLink fieldCode="DE" term="%22Instructional+Effectiveness%22">Instructional Effectiveness</searchLink><br /><searchLink fieldCode="DE" term="%22Concept+Formation%22">Concept Formation</searchLink><br /><searchLink fieldCode="DE" term="%22Retention+%28Psychology%29%22">Retention (Psychology)</searchLink><br /><searchLink fieldCode="DE" term="%22Knowledge+Level%22">Knowledge Level</searchLink><br /><searchLink fieldCode="DE" term="%22Prior+Learning%22">Prior Learning</searchLink><br /><searchLink fieldCode="DE" term="%22Grade+6%22">Grade 6</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1007/s11423-021-10034-3
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 1042-1629
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: This study investigated how a computer science (CS) problem-based curriculum impacted elementary students' CS learning and attitudes. Four sixth-grade teachers and 200 of their students participated in the study. Researchers developed a CS curriculum in collaboration with the teachers, which consisted of two main units: (1) an introduction to block-based coding and (2) a problem-based learning (PBL) applied coding project. Overall, students significantly improved their knowledge of CT concepts after the introductory block-based coding lessons and retained that knowledge after completing the PBL activities approximately three months later. Results suggest that "Event" and "Parallelism" were challenging concepts for most of the students, whereas "Loop" and "Sequence" were easily grasped by most of the students. Further analysis based on prior knowledge levels revealed that the high-prior knowledge (HK) group outperformed the low-prior knowledge (LK) group on every measure. However, LK narrowed the gap of CT concepts after the introductory block-based coding lessons. Students also communicated relatively positive attitudes towards CS at the conclusion of the PBL unit. These results provide support for further exploring the integration of inquiry-oriented instructional strategies such as PBL to support CS instruction.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2021
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1316649
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1316649
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s11423-021-10034-3
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 27
        StartPage: 2761
    Subjects:
      – SubjectFull: Problem Based Learning
        Type: general
      – SubjectFull: Elementary School Curriculum
        Type: general
      – SubjectFull: Elementary School Students
        Type: general
      – SubjectFull: Student Attitudes
        Type: general
      – SubjectFull: Computer Science Education
        Type: general
      – SubjectFull: Programming
        Type: general
      – SubjectFull: Thinking Skills
        Type: general
      – SubjectFull: Instructional Effectiveness
        Type: general
      – SubjectFull: Concept Formation
        Type: general
      – SubjectFull: Retention (Psychology)
        Type: general
      – SubjectFull: Knowledge Level
        Type: general
      – SubjectFull: Prior Learning
        Type: general
      – SubjectFull: Grade 6
        Type: general
    Titles:
      – TitleFull: Integration of Problem-Based Learning in Elementary Computer Science Education: Effects on Computational Thinking and Attitudes
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Kwon, Kyungbin
      – PersonEntity:
          Name:
            NameFull: Ottenbreit-Leftwich, Anne T.
      – PersonEntity:
          Name:
            NameFull: Brush, Thomas A.
      – PersonEntity:
          Name:
            NameFull: Jeon, Minji
      – PersonEntity:
          Name:
            NameFull: Yan, Ge
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 10
              Type: published
              Y: 2021
          Identifiers:
            – Type: issn-print
              Value: 1042-1629
          Numbering:
            – Type: volume
              Value: 69
            – Type: issue
              Value: 5
          Titles:
            – TitleFull: Educational Technology Research and Development
              Type: main
ResultId 1