CoChat
Sign Up
Cited by 60Teaching and Learning ProgrammingEducational Robotics and EngineeringSpreadsheets and End-User Computing

ChatScratch: An AI-Augmented System Toward Autonomous Visual Programming Learning for Children Aged 6-12

Open access
Chen, Liuqing, Xiao, Shuhong, Chen, Yunnong, Song, Yaxuan, Wu, Ruoyu, Sun, Lingyun·2024·Proceedings of the CHI Conference on Human Factors in Computing Systems
DOI 10.1145/3613904.3642229
Open this paper in CoChat — free accountOpen PDF

Abstract

As Computational Thinking (CT) continues to permeate younger age groups in K-12 education, established CT platforms such as Scratch face challenges in catering to these younger learners, particularly those in the elementary school (ages 6-12). Through formative investigation with Scratch experts, we uncover three key obstacles to children’s autonomous Scratch learning: artist’s block in project planning, bounded creativity in asset creation, and inadequate coding guidance during implementation. To address these barriers, we introduce ChatScratch, an AI-augmented system to facilitate autonomous programming learning for young children. ChatScratch employs structured interactive storyboards and visual cues to overcome artist’s block, integrates digital drawing and advanced image generation technologies to elevate creativity, and leverages Scratch-specialized Large Language Models (LLMs) for professional coding guidance. Our study shows that, compared to Scratch, ChatScratch efficiently fosters autonomous programming learning, and contributes to the creation of high-quality, personally meaningful Scratch projects for children.

AI snapshot

Written from the paper’s full text.

TL;DR

ChatScratch is an AI-augmented system for autonomous Scratch learning in children aged 6–12, combining structured storyboards/visual cues, drawing-based asset creation with Stable Diffusion/ControlNet, and a Scratch-specialized LLM code assistant. In a within-subjects study with 24 children, it improved project richness, expert-rated asset creativity, CT rubric scores, and code retention/expansion compared with Scratch.

Key findings

  • Visual element count was higher with ChatScratch (M = 4.81, SD = 2.34) than Scratch (M = 3.06, SD = 1.34), t(23) = 5.40, p < .01. Expert ratings were significantly higher for originality (4.29 vs. 2.17), consistency (4.04 vs. 2.79), and creativity (3.96 vs. 2.67), while quality was comparable (4.13 vs. 4.08, n.s.).
  • Dr. Scratch total scores were 13.21 (SD = 2.40) for ChatScratch vs. 8.13 (SD = 2.69) for Scratch, with significant differences in abstraction, parallelism, logical, synchronization, interactivity, and data, but not flow control. Code retention was M = 0.74 (SD = 0.15) and code expansion was M = 0.54 (SD = 0.17), indicating children retained about three-quarters of suggested code and implemented more than half of final code themselves.
  • Creativity Support Index ratings were significantly higher for ChatScratch in collaboration, exploration, expressiveness, immersion, and results worth effort; enjoyment was not significantly different.
  • Qualitative analysis found increased iterative improvement and personalization: 16/24 participants mentioned visual cues, 21 mentioned image polish, and 15 mentioned code templates; average use was 3.12 times for visual cues, 7.41 for image polish, and 4.78 for coding assistance.

Methods

  • Formative investigation with six experienced Scratch educators (3 female, 3 male; age 26–39, M = 31, SD = 4.44; teaching experience 4–12 years, M = 6.67, SD = 2.96), including classroom observations and semi-structured interviews analyzed with reflexive thematic analysis.
  • System design: ChatScratch includes an interactive storyboard with visual cues, drawing-based asset creation polished via Stable Diffusion with ControlNet, and a Scratch-specialized LLM code assistant using in-context learning and chain-of-thought prompting with GPT-4.
  • Evaluation: within-subjects study with 24 children aged 6–12 (10 female, 14 male; M = 8.96, SD = 2.12), two 70-minute sessions 48 hours apart, counterbalanced across tools and themes, with a 30-minute autonomous programming task; MANOVA followed by paired t-tests with Bonferroni correction, plus video and artifact-based interviews.

Limitations

  • The paper states that voice-based interaction lacks visual input feedback, so children may incorrectly assume interaction problems are their own mistakes rather than system limitations; future work includes rephrasing/refining inputs or asking clarifying questions.
  • The paper states that the structured three-act project plan limits generalizability: it suits stories with strong conflicts/twists but may not fit linear descriptive stories, and the system does not cover other work types such as music and games.
  • The paper states that the current system lacks mechanisms for assessing and providing feedback on the autonomous learning process. Evident from the text, though not framed as limitations: participants were native Mandarin speakers, Scratch beginners with less than 1 year of experience, and tasks were short (30 minutes), with no long-term retention or transfer measures reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1145/3613904.3642229

Cites

65 of 65 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 70 cited sources checked · 0 supported · 0 flagged

  • Findingcitation
    § ABSTRACT · p. 1

    Children experience bounded creativity in asset creation during autonomous Scratch learning.

    “Through formative investigation with Scratch experts, we uncover three key obstacles to children's autonomous Scratch learning: artist's block in project planning, bounded creativity in asset creation, and inadequate coding guidance during implementation.”
    • Also stated in: § 1 INTRODUCTION · p. 2 · plain
      “the existing options for assets creation within Scratch constrain children's creative expression, leading to a form of bounded creativity [56]”
    • Also stated in: § 3.2 Findings and Design Implications · p. 5 · hedged
      “This scenario illustrates the concept of bounded creativity [56], where predefined resources might inadvertently limit a child's imaginative expression.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [56] [56] Alexandre Anderson Romeiro and Thomaz Wood Jr. 2015. Bounded creativity: understanding the restrictions on creative work in advertising agencies. BARBrazilian Administration Review 12 (2015), 1-21. - source

      Not checked in the free run.

  • Findingown data
    § 1 INTRODUCTION · p. 2

    ChatScratch produced marked improvements in the creative aspects of children's programming, evidenced by an increase in the count of visual elements, a higher creativity support index, and expert ratings.

    “Quantitatively, our study showed marked improvements in the creative aspects of children's programming, evidenced by an increase in the count of visual elements, a higher creativity support index and expert ratings.”
    Quote in the paperSubstantial claim
  • Findingown data
    § ABSTRACT · p. 1

    ChatScratch contributes to the creation of high-quality, personally meaningful Scratch projects for children.

    “Our study shows that, compared to Scratch, ChatScratch efficiently fosters autonomous programming learning, and contributes to the creation of high-quality, personally meaningful Scratch projects for children.”
    Quote in the paperSubstantial claim
  • Findingown data
    § ABSTRACT · p. 1

    ChatScratch efficiently fosters autonomous programming learning compared to Scratch.

    “Our study shows that, compared to Scratch, ChatScratch efficiently fosters autonomous programming learning, and contributes to the creation of high-quality, personally meaningful Scratch projects for children.”
    Quote in the paperSubstantial claim
  • Findingcitation
    § ABSTRACT · p. 1

    Children experience artist's block in project planning during autonomous Scratch learning.

    “Through formative investigation with Scratch experts, we uncover three key obstacles to children's autonomous Scratch learning: artist's block in project planning, bounded creativity in asset creation, and inadequate coding guidance during implementation.”
    • Also stated in: § 1 INTRODUCTION · p. 2 · plain
      “children frequently get stuck suddenly, encountering an artist's block-a temporary creative impasse where they struggle to progress with ideas [32] during the project planning stage”
    • Also stated in: § 3.2 Findings and Design Implications · p. 5 · plain
      “Having described the main character and the waterfall, the girl found herself at an impasse and hit what is commonly known as an artist's block [32].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [32] [32] Zhuying Li, Yan Wang, Wei Wang, Stefan Greuter, and Florian 'Floyd' Mueller. 2020. Empowering a Creative City: Engage Citizens in Creating Street Art through Human-AI Collaboration. In Extended Abstracts of the 2020 CHI Conference on Human Factors in Computing Systems (Honolulu, HI, USA) (CHI EA '20) . Association for Computing Machinery, New York, NY, USA, 1-8. https: //doi.org/10.1145/3334480.3382976 - source

      Not checked in the free run.

  • May combine two claims. 59
  • Classified as probably not a substantial claim. 60
  • The quote only partly states this claim. 5

  • May combine two claims.

    Findingnone given
    § 1 INTRODUCTION · p. 3

    ChatScratch empowers children to craft personally meaningful projects and concurrently fosters their computational thinking skills.

    “We presented ChatScratch, an AI-augmented system designed for autonomous programming learning among primary school children aged 6-12. Leveraging visual cues, image polish, and a code assistance, ChatScratch empowers children to craft personally meaningful projects, concurrently fostering their computational thinking skills.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Findingnone given
    § ABSTRACT · p. 1

    Established CT platforms such as Scratch face challenges in catering to younger learners, particularly elementary school children aged 6-12.

    “As Computational Thinking (CT) continues to permeate younger age groups in K-12 education, established CT platforms such as Scratch face challenges in catering to these younger learners, particularly those in the elementary school (ages 6-12).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 1 INTRODUCTION · p. 3

    Barriers hinder children's autonomous Scratch learning and there is a heavy reliance on professional guidance.

    “Through a systematic formative investigation with expert Scratch educators, we identified barriers hindering children's autonomous Scratch learning and reveal a heavy reliance on professional guidance.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: In real-world teaching, educators approach personally meaningful projects with caution.
    • Rests on an assumption: Personally meaningful projects demand deeper programming understanding that can be daunting for children, and additional effort and time for teachers to provide individualized guidance and assessment.
    • Rests on an assumption: Standardized prescriptive projects often neglect to foster students' exploration and identification of personal interests.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Findingnone given
    § 3.2 Findings and Design Implications · p. 5

    Children naturally tend to either bypass the planning stage or integrate it directly into the coding process.

    “While children naturally tend to either bypass this planning stage or integrate it directly into the coding process”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Findingown data
    § 3.2 Findings and Design Implications · p. 6

    Preserving the original Scratch workflows avoids introducing additional cognitive burdens.

    “First, they underscored the significance of preserving the original Scratch workflows to avoid introducing additional cognitive burdens.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning
    § 3.2 Findings and Design Implications · p. 6

    Foundational code guidance provides a tangible starting point and instills a greater sense of confidence for children.

    “Such a foundational code guidance provides not only a tangible starting point but also instills a greater sense of confidence for children.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 5

    Scratch offers minimal imaginative support, hindering children's ability to discover appropriate details to enhance the narrative.

    “However, Scratch offered minimal imaginative support, hindering her ability to discover the appropriate details to enhance the narrative.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    The representative dataset contains 15 code examples.

    “Statistically, we have collected 15 code examples as our representative dataset.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 6

    Some children view coding as a means to animate their assets, with desirable assets having the potential to inspire their enthusiasm for coding.

    “Some children view coding as a means to animate their assets, with the desirable assets having the potential to inspire their enthusiasm for coding.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingreasoning
    § 3.2 Findings and Design Implications · p. 6

    Through drawing, children can more effectively achieve their ideas and personality.

    “The advantage lies in the fact that through drawing, children can more effectively achieve their ideas and personality.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 5

    Four interviewees mentioned employing a simple storyboard to conceptualize their stories and seeking inspiration to stimulate creativity.

    “To past this hurdle, four interviewees mentioned employing a simple storyboard to conceptualize their stories and seeking inspiration to stimulate creativity.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 6

    Many kids use the Scratch Cat as their character, not by choice, but due to the lack of appropriate assets, and such limitations frequently lead to frustration as their original creative intent is not truly represented.

    “I have seen many kids use the Scratch Cat as their character, not by choice, but due to the lack of appropriate assets. Such limitations frequently lead to their frustration as their original creative intent is not truly represented.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 6

    Fresh programmers anticipate guidance from their teachers to get started; otherwise they often engage in protracted deliberation.

    “They anticipate guidance from their teachers to get started; otherwise, they often engage in protracted deliberation.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingown data
    § 3.2 Findings and Design Implications · p. 6

    A transparent and structured workflow facilitates autonomous learning.

    “Second, our interviewees also emphasized the imperative nature of a transparent and structured workflow for facilitating autonomous learning.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 6

    Drawing requires more classroom time, and quality assurance is not guaranteed.

    “However, weaknesses are also evident. Drawing requires more classroom time, and quality assurance is not guaranteed, as reported by one interviewee.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 3.2 Findings and Design Implications · p. 5

    Leveraging Scratch's built-in assets library is the second and more prevalent method for asset creation.

    “The second and more prevalent method is leveraging Scratch's built-in assets library.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data
    § 7.3 Limitations and Future Work · p. 17

    Music and games are currently not covered in the ChatScratch system.

    “These areas, however, are currently not covered in our system.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Findingreasoning
    § 7.2 Cultivating Creativity and Programming Skills in Children through ChatScratch · p. 17

    In such cases, children are more immersed, willing to share, and feel a greater sense of participation, reflecting their investment and enthusiasm.

    “In such cases, children are more immersed, willing to share, and feel a greater sense of participation, reflecting their investment and enthusiasm.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Findingnone given
    § 8 CONCLUSION · p. 17

    A Scratch-specialized large language model assists children in their coding journey and improves their computational thinking skills.

    “Moreover, a Scratch-specialized large language model assists children in their coding journey and improves their computational thinking skills.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 6.3 RQ3: How does ChatScratch Support Children to Autonomously Create Personally Meaningful Projects · p. 15

    P13 navigated the gap between the actual effects of given code solutions and his expectation, and by identifying new requirements based on this disparity, synthesized multiple coding solutions to craft a more expressive code.

    “For programming phase, an illustrative example is the experience of P13, who navigated the gap between the actual effects of given codes solution and his expectation. By identifying new requirements based on this disparity, P13 synthesized multiple coding solutions to craft a more expressive code.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data
    § 7.3 Limitations and Future Work · p. 17

    Some participants mentioned in interviews their desire to use ChatScratch for creating other types of works, such as music and games.

    “In addition, some participants mentioned in interviews their desire to use ChatScratch for creating other types of works, such as music and games.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data
    § 6.2 RQ2: How does ChatScratch Support High-quality Code Implementation in Scratch Programming · p. 13

    Children were more comfortable with loops than with conditionals.

    “Based on our interviews, we learned that children were more comfortable with loops.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 6.2 RQ2: How does ChatScratch Support High-quality Code Implementation in Scratch Programming · p. 13

    Children were more hesitant with conditional statements such as 'if-else' and mentioned the challenge of determining the exact conditions needed to achieve desired program outcomes.

    “In contrast, when it came to using conditional statements such as 'if-else', children were more hesitant. They specifically mentioned the challenge of determining the exact conditions needed in 'if-else' statements to achieve their desired program outcomes.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data
    § 6.1 RQ1: How does ChatScratch Support Creative Tasks in Scratch Programming · p. 13

    P5 used ChatScratch to create a story of an unlucky hen that turned into roast chicken in a planetary collision; the story contained rich and exquisite assets and interesting plot design, and amused all the experts.

    “P5 used ChatScratch to create the story of the unlucky hen 8 , who turned into delicious roast chicken meat in the explosion of a planetary collision. This story contains rich and exquisite assets and interesting plot design, amused all the experts.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data
    § 6.3 RQ3: How does ChatScratch Support Children to Autonomously Create Personally Meaningful Projects · p. 16

    Three children (P2, P7, P14) all chose to develop their animal stories based on the classic fable of The Tortoise and the Hare, but each introduced a distinct twist.

    “We noticed a fascinating trend among three children (P2, P7, P14) who all chose to develop their animal stories based on the classic fable of The Tortoise and the Hare . However, each child introduced a distinct twist, infusing their unique interests into the narrative.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data
    § 7.1 The Benefit of Generative Artificial Intelligence in ChatScratch · p. 17

    The code expansion metric in the evaluation demonstrates that the children excelled in modifying, populating, and enriching the skeleton code.

    “The code expansion metric in our evaluation demonstrates that the children excelled in this aspect.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 6.1 RQ1: How does ChatScratch Support Creative Tasks in Scratch Programming · p. 13

    Children had a positive view of both systems, and ChatScratch received higher evaluations across all six dimensions.

    “From Figure 7, we can see that children have a positive view of both systems, and ChatScratch received higher evaluations across all six dimensions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 6 RESULTS · p. 13

    MANOVA showed a significant effect of tool (F(4,41)=21.59, p<0.01) but no significant effect of order (F(4,41)=0.46, p=0.76) or theme (F(4,41)=0.19, p=0.94).

    “As our MANOVA demonstrates, there is a significant effect on tool 𝐹 ( 4 , 41 ) = 21 . 59 , 𝑝 &lt; 0 . 01 ∗∗ , while no significance on order 𝐹 ( 4 , 41 ) = 0 . 46 , 𝑝 = 0 . 76 and theme 𝐹 ( 4 , 41 ) = 0 . 19 , 𝑝 = 0 . 94.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data
    § 6.3 RQ3: How does ChatScratch Support Children to Autonomously Create Personally Meaningful Projects · p. 15

    On average, the children used visual cues 3.12 times (SD=1.74), image polish 7.41 times (SD=2.33), and coding assistance 4.78 times (SD=1.78).

    “On average, the children used visual cues 3.12 times (SD = 1.74), image polish 7.41 times (SD = 2.33), and coding assistance 4.78 times (SD = 1.78).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given
    § 7.3 Limitations and Future Work · p. 17

    The current ChatScratch system lacks mechanisms for assessing and providing feedback on the autonomous learning process.

    “On the other hand, the current system lacks mechanisms for assessing and providing feedback on the autonomous learning process.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Findingown data
    § 7.1 The Benefit of Generative Artificial Intelligence in ChatScratch · p. 16

    The key necessity identified in the formative study is to provide a simple and easy starting point, thereby reducing the learning curve for the task.

    “In achieving this objective, the key necessity, as identified in our formative study, is to provide a simple and easy starting point, thereby reducing the learning curve for the task.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data
    § 6.1 RQ1: How does ChatScratch Support Creative Tasks in Scratch Programming · p. 13

    There was no significant difference in expert-rated quality between ChatScratch-produced assets and Scratch assets (M=4.13 vs 4.08, t=0.20, p=1).

    “While for quality, we observed closely matched means (4.13 for ChatScratch and 4.08 for Scratch) and standard deviations (0.74 for ChatScratch and 0.72 for Scratch) which confirms that ChatScratch-produced assets are on a quality level comparable to those from Scratch's built-in library.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 1 INTRODUCTION · p. 1

    Scratch remains the most commonly used platform in children's programming education.

    “While there are numerous alternatives, Scratch remains the most commonly used platform in children's programming education, primarily because of its inherent tinkerable nature, profound meaningfulness, and vibrant social community [51].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] [51] Mitchel Resnick, John Maloney, Andrés Monroy-Hernández, Natalie Rusk, Evelyn Eastmond, Karen Brennan, Amon Millner, Eric Rosenbaum, Jay Silver, Brian Silverman, and Yasmin Kafai. 2009. Scratch: Programming for All. Commun. ACM 52, 11 (nov 2009), 60-67. https://doi.org/10.1145/1592761.1592779 - source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 1 INTRODUCTION · p. 1

    Young learners often lack the requisite literacy, arithmetic capabilities, fine-motor skills, and hand-eye coordination to effectively engage with CT platforms.

    “Young learners often lack the requisite literacy, arithmetic capabilities, fine-motor skills, and hand-eye coordination to effectively engage with these CT platforms [64].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [64] [64] Jiahong Su and Weipeng Yang. 2023. A systematic review of integrating computational thinking in early childhood education. Computers and Education Open 4 (2023), 100122. https://doi.org/10.1016/j.caeo.2023.100122 - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 1 INTRODUCTION · p. 1

    Scratch is the most popular CT educational tool, with over 100 million registered users from more than 150 countries.

    “Scratch, due to its ability to provide teenagers with a more accessible and less abstract programming experience, has become the most popular CT educational tool, boasting over 100 million registered users from more than 150 countries [65].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [65] [65] Scratch Team. 2023. Community statistics at a glance. https://scratch.mit.edu/ statistics/ Accessed: 2023-09-02. -

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation
    § 3.1 Method · p. 4

    Educators hold an irreplaceable role in programming education.

    “Our decision to engage with professional educators instead of children and parents was driven by the understanding that educators still hold an irreplaceable role in programming education [22].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [22] [22] Michael S. Horn and Robert J. K. Jacob. 2007. Designing Tangible Programming Languages for Classroom Use. In Proceedings of the 1st International Conference on Tangible and Embedded Interaction (Baton Rouge, Louisiana) (TEI '07) . Association for Computing Machinery, New York, NY, USA, 159-162. https://doi.org/10. 1145/1226969.1227003 - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation
    § 2.3 Supporting Children's Personally Meaningful Projects through Child-AI Collaboration · p. 4

    Voice interfaces assist children in maintaining attention and dismantle literacy barriers by eliminating the need for written text.

    “They not only assist children in maintaining attention [45] but also dismantle literacy barriers by eliminating the need for written text [14].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [45] [45] Luiza Superti Pantoja, Kyle Diederich, Liam Crawford, and Juan Pablo Hourcade. 2019. Voice Agents Supporting High-Quality Social Play. In Proceedings of the 18th ACM International Conference on Interaction Design and Children (Boise, ID, USA) (IDC '19) . Association for Computing Machinery, New York, NY, USA, 314-325. https://doi.org/10.1145/3311927.3323151 - source

      Not checked in the free run.

    • cites [14] [14] Griffin Dietz, Nadin Tamer, Carina Ly, Jimmy K Le, and James A. Landay. 2023. Visual StoryCoder: A Multimodal Programming Environment for Children's Creation of Stories. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI '23) . Association for Computing Machinery, New York, NY, USA, Article 96, 16 pages. https://doi.org/10.1145/ 3544548.3580981 - source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation
    § 2.3 Supporting Children's Personally Meaningful Projects through Child-AI Collaboration · p. 4

    Strong adaptive capabilities and affective qualities are essential for the efficacy of AI assistants.

    “Furthermore, strong adaptive capabilities and affective qualities are also considered essential for the efficacy of AI assistants [33].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [33] [33] Chaolan Lin, Karl F MacDorman, Selma Šabanović, Andrew D Miller, and Erin Brady. 2020. Parental expectations, concerns, and acceptance of storytelling robots for children. In Companion of the 2020 ACM/IEEE International Conference on Human-Robot Interaction . Association for Computing Machinery, New York, NY, USA, 346-348. - source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 2.3 Supporting Children's Personally Meaningful Projects through Child-AI Collaboration · p. 4

    Some previous AI assistant systems rely on digital sensors but are limited to rudimentary interactions; other approaches rely on predefined rules to generate dialogue, often leading to logical inconsistencies.

    “Some systems [78] rely on digital sensors to perceive real world information but are limited to rudimentary interactions, and other approaches have relied on predefined rules to generate dialogue, which often led to logical inconsistencies [84].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [78] [78] Wenjie Xu, Jiayi Ma, Jiayu Yao, Weijia Lin, Chao Zhang, Xuanhe Xia, Nan Zhuang, Shitong Weng, Xiaoqian Xie, Shuyue Feng, Fangtian Ying, Preben Hansen, and Cheng Yao. 2023. MathKingdom: Teaching Children Mathematical Language Through Speaking at Home via a Voice-Guided Game. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI '23) . Association for Computing Machinery, New York, NY, USA, Article 93, 14 pages. https://doi.org/10.1145/3544548.3581043 - source

      Not checked in the free run.

    • cites [84] [84] Zheng Zhang, Ying Xu, Yanhao Wang, Bingsheng Yao, Daniel Ritchie, Tongshuang Wu, Mo Yu, Dakuo Wang, and Toby Jia-Jun Li. 2022. Storybuddy: A human-ai collaborative chatbot for parent-child interactive storytelling with flexible parental involvement. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems . Association for Computing Machinery, New York, NY, USA, 1-21. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 2.2 Enhancing Creativity in Scratch Learning Environments · p. 3

    Scratch integrates coding into the everyday creative activities of children, incorporating storytelling, drawing, and gaming as fundamental forms of expression.

    “In discussing Scratch, it extends beyond the traditional view of programming as merely a coding skill. Instead, Scratch integrates coding into the everyday creative activities of children, incorporating storytelling, drawing, and gaming as fundamental forms of expression [51].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] [51] Mitchel Resnick, John Maloney, Andrés Monroy-Hernández, Natalie Rusk, Evelyn Eastmond, Karen Brennan, Amon Millner, Eric Rosenbaum, Jay Silver, Brian Silverman, and Yasmin Kafai. 2009. Scratch: Programming for All. Commun. ACM 52, 11 (nov 2009), 60-67. https://doi.org/10.1145/1592761.1592779 - source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 1 INTRODUCTION · p. 1

    Current Scratch educational practices are mainly tailored for classroom environments, guided by professional educators, and a structured curriculum.

    “Additionally, current Scratch educational practices are mainly tailored for classroom environments, guided by professional educators, and a structured curriculum [22].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [22] [22] Michael S. Horn and Robert J. K. Jacob. 2007. Designing Tangible Programming Languages for Classroom Use. In Proceedings of the 1st International Conference on Tangible and Embedded Interaction (Baton Rouge, Louisiana) (TEI '07) . Association for Computing Machinery, New York, NY, USA, 159-162. https://doi.org/10. 1145/1226969.1227003 - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 1 INTRODUCTION · p. 1

    Early exposure to CT education is beneficial for children's future career and life prospects.

    “While early exposure to CT education has been recognized as beneficial for children's future career and life prospects [34], it is crucial to create effective learning method for young learners, especially those without professional guidance or in home-based educational settings.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [34] [34] Kirsti Lonka, Juho Makkonen, Minna Berg, Markus Talvio, Erika Maksniemi, Milla Kruskopf, Heidi Lammassaari, Lauri Hietajärvi, and Suvi Krista Westling. 2018. Phenomenal learning from Finland . Edita, Finland. -

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 1 INTRODUCTION · p. 1

    Project-based learning emerges as the most promising approach.

    “Concurrently, another line of research, has bypassed tool-specific discussions to focus instead on the topic of learning strategies, where project-based learning emerges as the most promising approaches [24, 64].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [24] [24] Ting-Chia Hsu, Shao-Chen Chang, and Yu-Ting Hung. 2018. How to learn and how to teach computational thinking: Suggestions based on a review of the literature. Computers &amp; Education 126 (2018), 296-310. - source

      Not checked in the free run.

    • cites [64] [64] Jiahong Su and Weipeng Yang. 2023. A systematic review of integrating computational thinking in early childhood education. Computers and Education Open 4 (2023), 100122. https://doi.org/10.1016/j.caeo.2023.100122 - source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation
    § 2.2 Enhancing Creativity in Scratch Learning Environments · p. 4

    Computer-assisted drawing tools lead children to demonstrate more creative theme selections compared to traditional paper-based drawing methods.

    “For instance, Kucirkova et al. [30] observed that, when provided with computer-assisted drawing tools, children demonstrated more creative theme selections compared to traditional paper-based drawing methods.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [30] [30] Natalia Kucirkova and Mona Sakr. 2015. Child-father creative text-making at home with crayons, iPad collage &amp; PC. Thinking Skills and Creativity 17 (2015), 59-73. - source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation
    § 3.1 Method · p. 4

    Educators' expertise in consolidating various educational materials, tools, and practices is pivotal for effectively nurturing CT skills in young children.

    “Their expertise in consolidating various educational materials, tools, and practices is pivotal for effectively nurturing CT skills in young children [38].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [38] [38] Andrew Manches and Lydia Plowman. 2017. Computing education in children's early years: A call for debate. British Journal of Educational Technology 48, 1 (2017), 191-201. https://doi.org/10.1111/bjet.12355 - source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation
    § 4.3 Goal 2: Drawing-based Assets Creation with Advanced Image Generation · p. 7

    FID gauges similarity in style and content between two image sets; a lower FID score signifies greater resemblance.

    “FID is crucial for gauging the similarity in style and content between two image sets, where a lower FID score signifies greater resemblance.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [11] [11] Min Jin Chong and David Forsyth. 2020. Effectively unbiased fid and inception score and where to find them. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . IEEE Computer Society, Washington, DC, USA, 6070-6079. - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    Beginners may not be able to find appropriate code blocks with only voice prompts.

    “For beginners, they may still not be able to find the appropriate code blocks with only voice prompts.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation
    § 3.2 Findings and Design Implications · p. 6

    Personally meaningful projects significantly enhance children's engagement, encouraging creative behavior, deepening understanding, and amplifying intrinsic motivation to explore and innovate.

    “First, personally meaningful projects significantly enhance children's engagement, thereby encouraging creative behavior, deepening their understanding, and amplifying their intrinsic motivation to explore and innovate [49].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [49] [49] AnnaReidandIanSolomonides. 2007. Design students' experience of engagement and creativity. Art, Design &amp; Communication in Higher Education 6, 1 (2007), 27-39. - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    Computing Levenshtein distance between LLM text responses and actual code block IDs prevents LLM illusions and improves quality of generated Scratch blocks.

    “This can prevent illusions caused by LLMs and improve the quality of the generated Scratch blocks.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationnone given
    § 4.3 Goal 2: Drawing-based Assets Creation with Advanced Image Generation · p. 7

    LLM-generated image prompts enhance the quality of the polished asset.

    “The image prompts generated by LLMs served to enhance the quality of the polished asset.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given
    § 4.1 System Overview · p. 7

    The code assistant can improve children's CT skills and engagement.

    “Leveraging a Scratch-specialized LLM, the assistant can not only assist children by generating step-by-step code guidance but also improve their CT skills and engagement.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 3.2 Findings and Design Implications · p. 6

    A foundational code solution is a classic embodiment of the scaffolding strategy.

    “In fact, a foundational code solution can be viewed as a classic embodiment of the scaffolding strategy [70].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [70] [70] Ge Wang, Jun Zhao, Max Van Kleek, and Nigel Shadbolt. 2023. 12 Ways to Empower: Designing for Children's Digital Autonomy. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI '23) . Association for Computing Machinery, New York, NY, USA, Article 91, 27 pages. https://doi.org/10.1145/3544548.3580935 - source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductown data
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    Educators independently wrote chain-of-thought code guidance for each code example, and mapping to Scratch blocks was discussed with educators.

    “To employ this paradigm, we asked the educators to independently write the code guidance with the chain-of-thought reasoning for each code example in the representative dataset. Then we discussed with educators how to map the text generated by LLMs onto specific Scratch blocks to achieve visual code generation.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation
    § 5.2 Procedure · p. 11

    A counterbalanced design alternating tool and theme combinations ensures unbiased results.

    “Additionally, we incorporated a counterbalanced design, alternating between combinations of tools (ChatScratch or Scratch) and themes (A or B) to ensure unbiased results [69].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [69] [69] Alina A von Davier, Paul W Holland, and Dorothy T Thayer. 2004. The Counterbalanced Design . Springer New York, New York, NY, 131-154. https: //doi.org/10.1007/0-387-21719-3\_9 -

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Interpretationreasoning
    § 4.2 Goal 1: Interactive Structured Storyboard and Visual Cues · p. 7

    Visual cues as stimuli help children detail and enrich the project.

    “To address this challenge, the interactive storyboard (Figure 3) employed visual cues as stimuli to help children detail and enrich the project.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    A representative example helps LLM elicit specific knowledge and abstractions necessary to accomplish the task.

    “A representative example helps LLM elicit specific knowledge and abstractions necessary to accomplish the task [80].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [80] [80] Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2023. ReAct: Synergizing Reasoning and Acting in Language Models. arXiv:2210.03629 [cs.CL] -

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data
    § 5.1 Participants · p. 11

    All participants were Scratch beginners with less than 1 year of experience with Scratch.

    “All participants were Scratch beginners and had less than 1 year of experience with Scratch.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation
    § 4.3 Goal 2: Drawing-based Assets Creation with Advanced Image Generation · p. 7

    SWD measures textural and structural parallels between image sets; a lower score indicates better similarity.

    “Similarly, SWD measures the textural and structural parallels between image sets, with a lower score indicating better similarity.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [29] [29] Soheil Kolouri, Kimia Nadjahi, Umut Simsekli, Roland Badeau, and Gustavo Rohde. 2019. Generalized sliced wasserstein distances . Curran Associates Inc., Red Hook, NY, USA, 261-272. -

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation
    § 3.2 Findings and Design Implications · p. 5

    Providing stimuli that can ignite creative thinking in children aids them in detailing and refining their stories and helps them overcome artist's block.

    “Second, it is crucial to provide stimuli that can ignite creative thinking in children, aiding them in the detailing and refinement of their stories; such stimuli can help children overcome artist's block [44, 71].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [44] [44] Hiroyuki Osone, Jun-Li Lu, and Yoichi Ochiai. 2021. BunCho: AI Supported Story Co-Creation via Unsupervised Multitask Learning to Increase Writers' Creativity in Japanese. In Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems (Yokohama, Japan) (CHI EA '21) . Association for Computing Machinery, New York, NY, USA, Article 19, 10 pages. https://doi.org/10.1145/ 3411763.3450391 - source

      Not checked in the free run.

    • cites [71] [71] Sitong Wang, Savvas Petridis, Taeahn Kwon, Xiaojuan Ma, and Lydia B Chilton. 2023. PopBlends: Strategies for Conceptual Blending with Large Language Models. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI '23) . Association for Computing Machinery, New York, NY, USA, Article 435, 19 pages. https://doi.org/10.1145/3544548.3580948 - source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 3.2 Findings and Design Implications · p. 6

    Children resonate more deeply with the sentiment 'This is something I created', fostering a stronger sense of ownership and achievement.

    “They resonate more deeply with the sentiment, ' This is something I created ', fostering a stronger sense of ownership and achievement [14].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [14] [14] Griffin Dietz, Nadin Tamer, Carina Ly, Jimmy K Le, and James A. Landay. 2023. Visual StoryCoder: A Multimodal Programming Environment for Children's Creation of Stories. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (Hamburg, Germany) (CHI '23) . Association for Computing Machinery, New York, NY, USA, Article 96, 16 pages. https://doi.org/10.1145/ 3544548.3580981 - source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    In-context learning proves efficacy for basic tasks with a couple of exemplars but faces hurdles with complex tasks demanding logical reasoning and multi-step resolutions.

    “Though in-context learning proves its efficacy for basic tasks involving a couple of exemplars, it faces hurdles with complex tasks that demand logical reasoning and multi-step resolutions, such as arithmetic or common-sense reasoning problems [28, 72].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [28] [28] Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. Advances in neural information processing systems 35 (2022), 22199-22213. -

      Not checked in the free run.

    • cites [72] [72] Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, and Denny Zhou. 2022. Rationale-Augmented Ensembles in Language Models. arXiv:2207.00747 [cs.CL] -

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning
    § 4.2 Goal 1: Interactive Structured Storyboard and Visual Cues · p. 7

    The interactive storyboard design helps children articulate ideas more clearly and fosters abstract thinking and attention to details.

    “This design not only helps children articulate their ideas more clearly but also fosters the development of their abstract thinking and attention to details.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    Generated Scratch blocks are a rough starting point; children still need to set loops, delay time, and interaction strategies.

    “The generated Scratch blocks (Figure 4) are just a rough starting point. Children still need to set the number of loops, change the delay time, and program interaction strategies.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning
    § 4.3 Goal 2: Drawing-based Assets Creation with Advanced Image Generation · p. 7

    Iterative polishing provides children with greater control over the final output, reinforcing ownership and engagement.

    “This approach provides children with greater control over the final output, thereby further reinforcing their sense of ownership and engagement with the project.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning
    § 3.2 Findings and Design Implications · p. 6

    Foundational code guidance offers necessary support to overcome initial hesitations and encourages proactive exploration and expansion of coding.

    “It offers them the necessary support to overcome initial hesitations about where to start and encourages more proactive exploration and expansion of their coding endeavors.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given
    § 5.1 Participants · p. 11

    All participants were native Mandarin speakers.

    “All participants were native Mandarin speakers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    Chain-of-thought reasoning integrates rationales as intermediate steps with examples.

    “To leverage the reasoning capabilities of LLMs, researchers have introduced the chain-of-thought (CoT) reasoning approach [73], where rationales are integrated as intermediate steps with examples, as depicted in Table 1.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [73] [73] Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems 35 (2022), 24824-24837. -

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    A visual programming practice can be expressed as the understanding of a single block and the effects that block combinations can produce.

    “As discussed in [53], a visual programming practice can be expressed as the understanding of a single block and the effects that block combinations can produce.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [53] [53] José Antonio Rodríguez-Martínez, José Antonio González-Calero, and José Manuel Sáez-López. 2020. Computational thinking and mathematics using Scratch: an experiment with sixth-grade students. Interactive Learning Environments 28, 3 (2020), 316-327. - source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation
    § 3.2 Findings and Design Implications · p. 5

    Most young children are at the formal operational stage of cognitive development, which often limits their abstract reasoning and executive functions, essential for constructing a cohesive project plan.

    “Yet, there is a hurdle: most young children are at the formal operational stage of cognitive development, which often limits their abstract reasoning and executive functions, essential for constructing a cohesive project plan [5].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [5] [5] Pierre Barrouillet. 2015. Theories of cognitive development: From Piaget to today. , 12 pages. - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Study design and conductnone given
    § 5.2 Procedure · p. 11

    A 48-hour interval between sessions avoids potential time-related and carryover effects.

    “To avoid potential time-related effects and carryover effects, we introduced a 48-hour interval between sessions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    The Scratch-specialized LLM is capable of recognizing and applying computational thinking concepts in classic Scratch programming cases.

    “Specifically, we proposed a LLM specialized for Scratch, which learns knowledge from a large-scale training corpus and is capable of recognizing and applying computational thinking concepts in classic Scratch programming cases.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given
    § 5.1 Participants · p. 11

    The experiment was approved by the university's ethics committee.

    “The participants were recruited via social media, and the experiment was approved by the university's ethics committee.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning
    § 4.3 Goal 2: Drawing-based Assets Creation with Advanced Image Generation · p. 8

    The role-playing strategy tailors LLM outputs to be educational and child-friendly.

    “For the LLM used to generate image prompts, we implemented a role-playing strategy [59], explicitly informing it to assume the roles of an educator and a child assistant. This approach is intended to tailor its outputs to be educational and child-friendly.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [59] [59] Murray Shanahan, Kyle McDonell, and Laria Reynolds. 2023. Role play with large language models. Nature 623 (2023), 493-498. https://doi.org/10.1038/s41586023-06647-8 - source

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims. The quote only partly states this claim.

    Study design and conductown data
    § 4.4 Goal 3: Code Assistant supported by Scratch-specialized Large Language Model · p. 9

    Three educators with more than four years of Scratch teaching experience independently reviewed all Scratch cards and selected representative examples, then reached consensus with two researchers.

    “three educators with more than four years of Scratch teaching experience were hired to work with the research team. Each educator was asked to independently review all Scratch cards and selected samples with educational significance.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning
    § 7.3 Limitations and Future Work · p. 17

    When children seek to craft linear, descriptive stories, like narrating a typical day in their lives, the structured three-act design might not be as suitable.

    “When children seek to craft linear, descriptive stories, like narrating a typical day in their lives, the structured three-act design might not be as suitable.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 6.3 RQ3: How does ChatScratch Support Children to Autonomously Create Personally Meaningful Projects · p. 15

    By incorporating visual cues, image polish, and code templates in ChatScratch, children receive ongoing intermediate feedback on their creative endeavors, allowing a gradual refinement process.

    “By incorporating visual cues, image polish, and code templates in ChatScratch, we provide children with the opportunity to receive ongoing, intermediate feedback on their creative endeavors. This allows for a gradual refinement process, ensuring that their projects progressively evolve to match their vision.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data
    § 6.1 RQ1: How does ChatScratch Support Creative Tasks in Scratch Programming · p. 13

    ChatScratch's collaborative support, visual cues, and image polish improve children's creative process by better representing efforts, supporting exploration, and addressing asset acquisition, allowing focus on creation.

    “In summary, the data sufficiently reflects our expectations for ChatScratch. With the collaborative support provided by ChatScratch (collaboration), children's efforts in creative tasks are better represented (results worth the effort). Visual cues support children in exploring the richness and details of projects (exploration), while image polish addresses the inconvenience of original asset acqu”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning
    § 5.4 Data Analysis · p. 12

    A Bonferroni correction was applied to mitigate the risk of Type I errors associated with multiple comparisons.

    “To mitigate the risk of Type I errors associated with multiple comparisons, a Bonferroni correction is applied.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 6.3 RQ3: How does ChatScratch Support Children to Autonomously Create Personally Meaningful Projects · p. 16

    Their stories, while rooted in the same origin, diverged in ways that mirrored their unique perspectives and preferences, showcasing the powerful impact of personal interests in shaping creative expressions.

    “Their stories, while rooted in a same origin, diverged in ways that mirrored their unique perspectives and preferences, showcasing the powerful impact of personal interests in shaping creative expressions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data
    § 7.1 The Benefit of Generative Artificial Intelligence in ChatScratch · p. 16

    The created assets meet children's expectations, fulfilling their vision for specific themes or images, thereby achieving their sense of accomplishment and engagement.

    “Third, the created assets need to meet children's expectations, fulfilling their vision for specific themes or images, thereby achieving their sense of accomplishment and engagement.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The contents generated by ChatScratch are non-toxic and safe, suitable for children.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given
    § 7.2 Cultivating Creativity and Programming Skills in Children through ChatScratch · p. 17

    ChatScratch's asset creation feature, which provides customizable characters and scenes, enables children to develop a strong attachment to their characters.

    “Furthermore, ChatScratch's asset creation feature, which provides customizable characters and scenes, enables children to develop a strong attachment to their characters.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning
    § 7.3 Limitations and Future Work · p. 17

    This might not have been a significant problem in previous systems, where the tasks handled were generally more straightforward and followed a more predictable pattern.

    “This might not have been a significant problem in previous system, where the tasks handled were generally more straightforward and followed a more predictable pattern.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § Notes: · p. 14

    The high retention rate indicates the usefulness of the suggested code and validates the accuracy of the coding solutions generated by the code assistant.

    “This high retention rate indicates the usefulness of the suggested code, thereby validating the accuracy of the coding solutions generated by our code assistant.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 7.1 The Benefit of Generative Artificial Intelligence in ChatScratch · p. 17

    The skeleton approach encourages children to modify, populate, and enrich it, ultimately leading them to craft their own unique implementation.

    “This approach encourages children to modify, populate, and enrich it, ultimately leading them to craft their own unique implementation.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data
    § 7.1 The Benefit of Generative Artificial Intelligence in ChatScratch · p. 16

    Generative artificial intelligence plays a crucial role in assisting children with project planning, supporting their creative endeavors, and providing guidance in programming.

    “In our ChatScratch, the generative artificial intelligence (GAI) plays a crucial role in assisting children with project planning, supporting their creative endeavors, and providing guidance in programming.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Generative artificial intelligence possesses the ability to understand, reason, and think creatively.
    • Rests on an assumption: The skills of understanding, reasoning, and creative thinking are crucial for guiding and enriching the creative process in children, especially in an autonomous learning environment where outside resources may be limited.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given
    § 7.2 Cultivating Creativity and Programming Skills in Children through ChatScratch · p. 17

    Combining support for both creativity and programming leads to higher completion rates and more engaging personally meaningful projects.

    “Combining support for both creativity and programming leads to higher completion rates and more engaging personally meaningful projects.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 6 RESULTS · p. 13

    Order and theme were effectively controlled, allowing direct reporting of the subsequent paired t-tests in RQ1 and RQ2.

    “This confirms that both order and theme were effectively controlled, allowing us to directly report the subsequent paired t-tests in RQ1 and RQ2.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given
    § 7.3 Limitations and Future Work · p. 17

    The structured project plan framework in ChatScratch, segmenting projects into three main acts with emphasis on characters, scenes, and events, is beneficial for children as it reduces their cognitive load during the conceptualization phase.

    “Overall, such a design is beneficial for children as it reduces their cognitive load during the conceptualization phase.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given
    § 5.2 Procedure · p. 11

    Researchers were only allowed to assist children when unexpected technical difficulties arose.

    “During the tasks, researchers were only allowed to assist children when unexpected technical difficulties arose.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given
    § 7.3 Limitations and Future Work · p. 17

    The three-act structure of 'Setup - Development - Climax' or 'Introduction - Conflict - Resolution' is more suited to stories with strong conflicts and twists, such as adventure tales often favored by children.

    “Specifically, the threeact structure of 'Setup - Development - Climax' or 'Introduction Conflict - Resolution' is more suited to stories with strong conflicts and twists, such as adventure tales that are often favored by children.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given
    § 5.3 Collected Data and Metrics · p. 11

    Any data requiring subjective evaluations were independently assessed by multiple evaluators, with consistency among them documented.

    “Any data requiring subjective evaluations were independently assessed by multiple evaluators, with consistency among them documented.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 7 DISCUSSION · p. 16

    ChatScratch effectively leverages generative artificial intelligence to overcome challenges identified in the formative study, providing substantial support for creativity (RQ1) and programming skill (RQ2) in children's visual programming activities.

    “Firstly, we demonstrate how ChatScratch effectively leverages the advantages of generative artificial intelligence to overcome the challenges identified in our formative study, thereby providing substantial support for creativity (RQ1) and programming skill (RQ2) in children's visual programming activities.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Generative artificial intelligence possesses the ability to understand, reason, and think creatively.
    • Rests on an assumption: The skills of understanding, reasoning, and creative thinking are crucial for guiding and enriching the creative process in children, especially in an autonomous learning environment where outside resources may be limited.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationreasoning
    § 7.2 Cultivating Creativity and Programming Skills in Children through ChatScratch · p. 17

    This heightened engagement and satisfaction are crucial in sustaining interest in ChatScratch, thereby encouraging its consistent use and practice.

    “This heightened engagement and satisfaction are crucial in sustaining interest in ChatScratch, thereby encouraging its consistent use and practice.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation
    § 6.2 RQ2: How does ChatScratch Support High-quality Code Implementation in Scratch Programming · p. 13

    High scores in the data dimension require children to master the use of complex data structures like lists.

    “According to Dr. Scratch's rubrics, high scores in the data dimension require children to master the use of complex data structures like lists.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [40] [40] Jesús Moreno-León and Gregorio Robles. 2015. Dr. Scratch: A web tool to automatically evaluate Scratch projects. In Proceedings of the workshop in primary and secondary computing education . Association for Computing Machinery, New York, NY, USA, 132-133. - source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims. The quote only partly states this claim.

    Interpretationreasoning
    § 7.2 Cultivating Creativity and Programming Skills in Children through ChatScratch · p. 17

    The project planning feature in ChatScratch provides a structured method for story development, prompting children to think logically and generate diverse ideas for expression through coding.

    “The project planning feature in Scratch provides a structured method for story development, prompting children to think logically and generate diverse ideas for expression through coding.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning
    § 7.2 Cultivating Creativity and Programming Skills in Children through ChatScratch · p. 17

    A robust creative idea instills resilience in children, making them less likely to give up and more likely to persevere through intricate coding challenges, learning from errors and continuously refining their projects.

    “Moreover, a robust creative idea instills resilience. Driven by the aspiration to realize their vision as imagined, children are less likely to give up. They persevere through intricate coding challenges, learning from errors and continuously refining their projects.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data
    § 6.1 RQ1: How does ChatScratch Support Creative Tasks in Scratch Programming · p. 13

    With the support of visual cues and image polish, children are able to create more freely (originality) and produce assets that align closely with their expectations (consistency), leading to an overall enhanced creative expression (creativity).

    “These results suggest that with the support of visual cues and image polish, children are able to create more freely (originality) and produce assets that align closely with their expectations (consistency), leading to an overall enhanced creative expression (creativity).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning
    § 7.3 Limitations and Future Work · p. 17

    In systems incorporating LLMs like ChatScratch, the situation is different: these systems are expected to handle more complex tasks, and the system itself is more flexible and sensitive in processing inputs.

    “In systems incorporating LLMs like ChatScratch, the situation is indeed different. These systems are expected to handle more complex tasks, and the system itself is more flexible and sensitive in processing inputs.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: LLM-based systems like ChatScratch are expected to handle more complex tasks than previous systems.
    • Rests on an assumption: LLM-based systems like ChatScratch are more flexible and sensitive in processing inputs than previous systems.
98 of 98

  • The paper relies on this without support.

    The classroom-based educational model for Scratch is often inaccessible to students in under-resourced communities or developing countries.

    “Such an educational model is often inaccessible to students in under-resourced communities or developing countries.”

  • The paper relies on this without support.

    In real-world teaching, educators approach personally meaningful projects with caution.

    “Yet, when it comes to real-world teaching, educators approach these projects with caution.”

  • The paper relies on this without support.

    Personally meaningful projects demand deeper programming understanding that can be daunting for children, and additional effort and time for teachers to provide individualized guidance and assessment.

    “The primary reasons being the deeper programming understanding they demand, which can be daunting for children, and also the additional effort and time teachers need to invest for individualized guidance and assessment.”

  • The paper relies on this without support.

    Standardized prescriptive projects often neglect to foster students' exploration and identification of personal interests.

    “While they ensure uniform learning benchmarks, they often neglect to foster students' exploration and identification of personal interests.”

  • The paper relies on this without support.

    Previous AI assistants designed for children have often fallen short.

    “Despite these ideals, previous AI assistants designed for children have often fallen short.”

  • The paper relies on this without support.

    Dr. Scratch has emerged as the most popular assessment tool within the community due to its intuitive interface and ease of accessibility.

    “Due to its intuitive interface and ease of accessibility, Dr. Scratch has emerged as the most popular assessment tool within the community.”

  • The paper relies on this without support.

    Children's drawing abilities are limited.

    “considering the limitations in children's drawing abilities and the aspiration for high-quality, tailored assets, we incorporated Stable Diffusion with ControlNet [55, 83].”

  • The paper relies on this without support.

    Personally meaningful projects are more likely to reflect the child's individual interests, experiences, or aspirations than standardized projects.

    “Unlike standardized projects that focus primarily on imparting specific knowledge or programming concepts, personally meaningful projects are more likely to reflect the child's individual interests, experiences, or aspirations.”

  • The paper relies on this without support.

    Scratch cards are quintessential introductory materials for programming.

    “In the Scratch community, Scratch cards 1 stand out as quintessential introductory materials for programming.”

  • The paper relies on this without support.

    The compactness and portability of Scratch cards make them an optimal resource for both newcomers and classroom courses.

    “The compactness and portability of Scratch cards render them an optimal resource for both newcomers and classroom courses.”

  • The paper relies on this without support.

    Children's literacy level requires voice chats for conversations.

    “Considering children's literacy level, such conversations will be presented as voice chats.”

  • The paper relies on this without support.

    Generative artificial intelligence possesses the ability to understand, reason, and think creatively.

    “This capability is precisely where GAI excels-it possesses the ability to understand, reason, and think creatively.”

  • The paper relies on this without support.

    The contents generated by ChatScratch are non-toxic and safe, suitable for children.

    “Second, the contents generated by our system must be non-toxic and safe, suitable for children.”

  • The paper relies on this without support.

    The skills of understanding, reasoning, and creative thinking are crucial for guiding and enriching the creative process in children, especially in an autonomous learning environment where outside resources may be limited.

    “These skills are crucial for guiding and enriching the creative process in children, especially in an autonomous learning environment where outside resources may be limited.”

  • The paper relies on this without support.

    Continual interaction is vital for developing and mastering visual programming skills, particularly in the context of autonomous learning.

    “Such continual interaction is vital for developing and mastering visual programming skills, particularly in the context of autonomous learning.”

  • The paper relies on this without support.

    Creativity serves as a key enhancer in the programming process.

    “Creativity serves as a key enhancer in the programming process.”

  • The paper relies on this without support.

    LLM-based systems like ChatScratch are expected to handle more complex tasks than previous systems.

    “These systems are expected to handle more complex tasks, and the system itself is more flexible and sensitive in processing inputs.”

  • The paper relies on this without support.

    LLM-based systems like ChatScratch are more flexible and sensitive in processing inputs than previous systems.

    “These systems are expected to handle more complex tasks, and the system itself is more flexible and sensitive in processing inputs.”
18 of 18

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • A second major objective is to situate the development of computational thinking and programming competence within authentic practices, using AI-powered programming environments and intelligent feedback mechanisms to lower entry barriers, provide personalized guidance, and ensure timely feedback for novice learners.

    “A second major objective is to situate the development of computational thinking and programming competence within authentic practices. Here, the target is to support students in refining their algorithmic reasoning, problem decomposition, and coding practices. The integration of AI-powered programming environments and intelligent feedback mechanisms serves to meet these goals, particularly by low”

    From A Systematic Review of Generative AI in K–12: Mapping Goals, Activities, Roles, and Outcomes via the 3P Model (2025) · cites this paper as [96]

    Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

    Experimental — accuracy not yet measured

  • Within the AI-empowered framework, students actively set prompts and control content generation, while AI, functioning as an autonomous collaborator, continuously analyzes learner preferences, behaviors, and project progress to adaptively optimize its models for individual needs.

    “Within this framework, students actively set prompts and control content generation, while AI, functioning as an autonomous collaborator, continuously analyzes learner preferences, behaviors, and project progress to adaptively optimize its models for individual needs [70,96,108,132,144].”

    From A Systematic Review of Generative AI in K–12: Mapping Goals, Activities, Roles, and Outcomes via the 3P Model (2025) · cites this paper as [96]

    Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

    Experimental — accuracy not yet measured

  • AI fosters generative practices: it supports creative writing, guides user-centered design thinking, enables more autonomous programming, and helps generate arts and history content, encouraging learners to move beyond 'the right answer' toward more original exploration and construction.

    “Second, AI fosters generative practices: it supports creative writing [135], guides user-centered design thinking [117], enables more autonomous programming [96], and helps generate arts and history content [72,113], encouraging learners to move beyond 'the right answer' toward more original exploration and construction.”

    From A Systematic Review of Generative AI in K–12: Mapping Goals, Activities, Roles, and Outcomes via the 3P Model (2025) · cites this paper as [96]

  • Several reviewed studies involved multiple AI roles.

    “with several studies [4,11,59,62,63] involving multiple roles”

    From Child-AI Co-Creation: A Review of the Current Research Landscape and a Proposal for Six Design Considerations (2025) · cites this paper as [11]

  • Other reviewed systems adopt an enhancer approach where children first generate ideas or artifacts and then receive alternative suggestions or elaborations from AI.

    “Others adopt an enhancer approach [11,35,59,60,[62][63][64], where children are first encouraged to generate ideas or artifacts on their own, and only afterward receive alternative suggestions or elaborations from the AI.”

    From Child-AI Co-Creation: A Review of the Current Research Landscape and a Proposal for Six Design Considerations (2025) · cites this paper as [11]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Animal-story participant changes from P14 to P11

    Minor

    The text identifies P2, P7, and P14 as the three children who adapted the fable, but its third example names P11 instead of P14.

  • Children's contribution to final code is reported as both 54% and 57%

    Minor

    The results report a mean code expansion of 0.54, described as the share of final code implemented by children. The discussion gives that share as 57%.

  • Results Worth Effort has conflicting significance markers

    Minor

    The prose marks this result with ∗∗, but Table 4 marks the same result with ∗ despite reporting p = 0.006. The table's note defines ∗∗ as p < 0.01.

3 of 3

Cites

Flagged first, then by citations

Research works printed but not linked: 13

  1. Teresa M Amabile. 1982. Social psychology of creativity: A consensual assessment technique.Journal of personality and social psychology 43, 5 (1982), 997.
  2. Virginia Braun and Victoria Clarke. 2006. Using thematic analysis in psychology. Qualitative research in psychology 3, 2 (2006), 77–101.
  3. Virginia Braun and Victoria Clarke. 2021. One size fits all? What counts as quality practice in (reflexive) thematic analysis?Qualitative research in psychology 18, 3 (2021), 328–352.
  4. Karen Brennan and Mitchel Resnick. 2012. New frameworks for studying and assessing the development of computational thinking. In Proceedings of the 2012 annual meeting of the American educational research association, Vancouver, Canada, Vol. 1. American Educational Research Association, Washington, DC, USA, 25.
  5. Laura Delgaty. 2009. Curriculum mapping: are you thinking what I’m thinking? A visual comparison of standardized, prescriptive programmes. Annual Review of Education, Communication & Language Sciences 6 (2009), 35–58.
  6. Jill Denner, Linda Werner, Shannon Campe, and Eloy Ortiz. 2014. Pair programming: Under what conditions is it advantageous for middle school students?Journal of Research on Technology in Education 46, 3 (2014), 277–296.
  7. Google. 2023. Teachable Machine. https://teachablemachine.withgoogle.com/ [Online; accessed 13-November-2023].
  8. Shuchi Grover, Roy Pea, and Stephen Cooper. 2015. Designing for deeper learning in a blended computer science course for middle school students. Computer science education 25, 2 (2015), 199–237.
  9. Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, and Yusuke Iwasawa. 2022. Large language models are zero-shot reasoners. Advances in neural information processing systems 35 (2022), 22199–22213.
  10. Stable Diffusion Online. 2023. Stable Diffusion: A Latent Text-to-Image Diffusion Model. https://stablediffusionweb.com/ Accessed: 2023-09-08.
  11. Hae Won Park, Mirko Gelsomini, Jin Joo Lee, and Cynthia Breazeal. 2017. Telling Stories to Robots: The Effect of Backchanneling on a Child’s Storytelling. In 2017 12th ACM/IEEE International Conference on Human-Robot Interaction (HRI). Association for Computing Machinery, New York, NY, USA, 100–108.
  12. James V Wertsch. 1984. The zone of proximal development: Some conceptual issues.New directions for child development 1984, 23 (1984), 7–18.
  13. Randi Williams. 2018. PopBots: leveraging social robots to aid preschool children’s artificial intelligence education. Ph. D. Dissertation. Massachusetts Institute of Technology.

Cited by

Sorted by citations

Related

Authors
Chen, L., Xiao, S., Chen, Y., Song, Y., Wu, R., & Sun, L.
Affiliations
  • Zhejiang University (CN)
  • Beijing Normal University (CN)
ORCID
Venue
Proceedings of the CHI Conference on Human Factors in Computing Systems
Publisher
ACM
Published
2024-05-11
Pages
1-19
DOI
10.1145/3613904.3642229
OpenAlex
W4391673398
Semantic Scholar
dcdc899fdbfe882a941528c0512bd840b68e8645
Cited by
60
References
59
Citations by year
2026: 24 · 2025: 35 · 2024: 1
FWCI
26.32
Type
conference paper
Language
en
Open access
Licence
www.acm.org/publications/policies/copyright_policy#Background (vor)
Locations
  • CHI Conference on Human Factors in Computing Systems (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • The Ng Teng Fong Charitable Foundation in the form of ZJU-SUTD IDEA Grant (188170-11102)
  • National Natural Science Foundation of China (62207023)
Topics
  • Physical Sciences › Computer Science › Computer Science Applications › Teaching and Learning Programming
  • Physical Sciences › Computer Science › Artificial Intelligence › Educational Robotics and Engineering
  • Physical Sciences › Computer Science › Software › Spreadsheets and End-User Computing
Keywords
Scratch, visual programming, computational thinking, self-directed learning, children, large language models, project planning, digital drawing, image generation, elementary school students, AI-augmented learning

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review