Policy–enforcement pairings are dominated by prompt-based constraints, and policy themes differ in enforcement breadth¶

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

Subclaims¶

q2 i? Prompt-Based Constraint accounted for 184 of 423 observed policy–enforcement pair instances (43.5%), and a chi-square analysis showed policy themes and enforcement themes were not distributed independently (χ2(20) = 60.77, p < .001, Cramer's V = .190). → Minsun Kim 2026 q2 i? Policy categories differed in enforcement breadth: Demonstrating Understanding appeared with all 11 enforcement mechanisms, while Direct-Solution Prohibition was connected to 5 mechanisms. → Minsun Kim 2026 (2)

Evidence¶

Minsun Kim 2026¶

Minsun Kim, S. Moonwara A. Monisha, Zihan Wu, David H. Smith IV. (2026). Exploring the Design Space of LLM-Based Programming Support in CS Education: A Scoping Review through the Lens of Assistance Governance. https://arxiv.org/abs/2607.21257 doi:10.48550/arxiv.2607.21257

q2 · i? · review · r2

Chi-square analysis of the theme-level policy–enforcement matrix across 423 pair instances from the 90-system corpus. The authors interpret this as a descriptive pattern rather than a causal claim because the data are multi-label and pair instances from the same paper are not fully independent.

"A chi-square analysis of the theme-level matrix showed that policy themes and enforcement themes were not distributed independently, χ2(20) = 60.77,p < .001, Cramer’sV=.190."

Minsun Kim 2026 (2)¶

Minsun Kim, S. Moonwara A. Monisha, Zihan Wu, David H. Smith IV. (2026). Exploring the Design Space of LLM-Based Programming Support in CS Education: A Scoping Review through the Lens of Assistance Governance. https://arxiv.org/abs/2607.21257

q2 · i? · review · r2

Category-level policy–enforcement heatmap analysis (Fig. 3) of pair instances across the 12 policy and 11 enforcement categories. Direct-Solution Prohibition was connected to 5 enforcement mechanisms, mainly Prompt-Based Constraint and Rule-Based Gatekeeping.

"Demonstrat- ing Understanding ( P6 ) showed the broadest enforcement breadth, appearing with all 11 enforcement mechanisms."

Discussion¶

Related Claims¶

• Underexplored PEA configurations center on participation support, AI literacy, and content generation, with two theme-level combinations absent from the corpus — related
• Coded prompt logs show different prompting strategies foreground different pedagogical themes, with technical prompts the largest strategy total (398 instances) — related
• Prompt-Based Constraint is the dominant enforcement mechanism in LLM-based programming support systems — a narrower finding that bears on this claim
• Authority over assistance boundaries is highly centralized in system logic in LLM-based programming support systems — related