Coded prompt logs show different prompting strategies foreground different pedagogical themes, with technical prompts the largest strategy total (398 instances)¶

Claim · All claims
Evidence · 2 studies · 2 causal r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? The theme–strategy matrix recorded 1,051 coded co-occurrences: technical prompts totaled 398, practical 310, exploratory 158, structured 105, narrative 80; by theme, instructional activities appeared most frequently (265), then learning objectives (212) and assessment design (189). → Bi T 2026 q2 i? Structured prompts made objective specification, constraint setting, assessment criteria, faculty verification, student disclosure, and risk control explicit; coding reliability was Cohen's κ = 0.82 for strategy type and 0.79 for theme presence. → Bi T 2026 (2)

Evidence¶

Bi T 2026¶

Bi T, Araily S, Lyu A and Xiu S (2026) Research on the optimization of the training system of university faculty development centers in the context of GenAI: a comparative analysis based on Chinese and Kazakhstani universities. Front. Psychol. 17:1890508. https://doi.org/10.3389/fpsyg.2026.1890508

q2 · i? · causal · r2

Rule-based qualitative coding of prompts from the structured training (30% double-coded subsample of 96 prompts; overall agreement 87.5%). The matrix recorded 1,051 coded co-occurrences, and technical prompts formed "thelargeststrategytotal,with398instances"; by theme, instructional activities appeared most frequently (265), then learning objectives (212) and assessment design (189).

"Technical prompts accounted forthelargeststrategytotal,with398instances,followed by practical prompts with 310, exploratory prompts with 158, structured prompts with 105, and narrative prompts with 80."

Bi T 2026 (2)¶

Bi T, Araily S, Lyu A and Xiu S (2026) Research on the optimization of the training system of university faculty development centers in the context of GenAI: a comparative analysis based on Chinese and Kazakhstani universities. Front. Psychol. 17:1890508. https://doi.org/10.3389/fpsyg.2026.1890508

q2 · i? · causal · r2

Reliability analysis of the 30% double-coded subsample of 96 prompts. The article reports "Cohen's κ was0.82forpromptstrategytypeand0.79forpedagogicaltheme presence", with binary-theme κ values from 0.74 to 0.84 and overall coding agreement 87.5%. The article also states structured prompts "made objective specification, constraint setting, assessment criteria, faculty verification, student disclosure, and risk controlexplicit".

"Cohen's κ was0.82forpromptstrategytypeand0.79forpedagogicaltheme presence."

Discussion¶

Related Claims¶

• ChatGPT research in programming education prioritizes classroom practice and learner interaction, with limited attention to assessment design and governance — related
• Student reflections identify speed and scaffolding as the dominant theme, with verification burden, overreliance worries, and prompt precision also prevalent — related
• Policy–enforcement pairings are dominated by prompt-based constraints, and policy themes differ in enforcement breadth — related
• Students showed bounded trust and uneven verification behavior: 23 of 837 student turns (about 2.7%) contained explicit risk-aware moves — related
• Low comment rewriting does not imply a smooth process: students describe the main effort as verifying generated code rather than revising prompts — related
• Instructional decisions form the hub of decision-type co-occurrence, most strongly with assessment decisions — related