Step-by-step explanations revealed one at a time, plus removed code auto-insertion and copy buttons, encouraged deeper engagement with generated explanations¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2–r3 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? In the lab study, 8/9 users progressed through step-by-step explanations, clicking the 'next step' button between 3-12 times (M=6.4, SD=3.9), and four participants said they appreciated the breakdown compared to ChatGPT. → Tseng 2026 q2 i? Participants reported that removing the copy button and auto-insertion made them more intentional, reading generated code line by line rather than copy-pasting without understanding. → Tseng 2026 (2)

Evidence¶

Tseng 2026¶

Tseng, T., Seoror, L.H., Adda, J., Factor, M., Darabi, R., Matschke, K.R., Fu, T., Lin, A., Maram, A., Sinha, A. (2026). Flowcode: An AI-Powered Programming Environment for Scaffolding Iteration in Creative Computing Education. https://arxiv.org/abs/2607.06721 doi:10.48550/arxiv.2607.06721

q2 · i? · design · r3

Log analysis from the lab study (n=9) using the updated Flowcode. The article states revealing one step at a time "can help users focus on individual parts and reduce cognitive load," and quotes P7 preferring steps over "walls of text."

"Our logs revealed that 8/9 users progressed through step-by-step explanations, with users clicking on the 'next step' button between 3-12 times (M=6.4,SD=3.9)."

Tseng 2026 (2)¶

Tseng, T., Seoror, L.H., Adda, J., Factor, M., Darabi, R., Matschke, K.R., Fu, T., Lin, A., Maram, A., Sinha, A. (2026). Flowcode: An AI-Powered Programming Environment for Scaffolding Iteration in Creative Computing Education. https://arxiv.org/abs/2607.06721. doi:10.48550/arxiv.2607.06721

q2 · i? · design · r2

Participant interview quote from the lab study (P6) on the effect of removing the copy-to-clipboard button. The article frames this as friction that helps users "be more intentional with generated code."

"but for Flowcode, it does not have the copy button so subconsciously, I would go through the code, read through it line by line, and try to understand what it means"

Discussion¶

Related Claims¶

• Automatically updating the flowchart and injecting generated code into editors drew participants away from explanations and toward direct code editing — related
• Participants used the Flowcode LLM across all stages of their design process, most frequently asking for code explanations — related
• Letting learners pace instruction helped on complex material in one experiment, while learner control in general showed almost no effect across 18 studies — a broader claim this one bears on
• Novice creative coders exhibited three distinct initial interaction patterns with Flowcode: flowchart-first, code-first, and prompt-first — related
• In an exploratory pilot, a soft barrier shifted participants from direct copying toward editing, understanding, and iterative resubmission without preventing task completion — related
• An LLM-generated flowchart linking semantic feature descriptions to code across HTML, CSS, and JS files supports novice creative coders' orientation, navigation, and code understanding — related