FRAC and the combined intervention shifted revision toward deeper levels, with deep revision (L3+L4) rising from 25% in the regular AI group to 55% in the combined group¶

Claim · All claims
Evidence · 1 study · causal r2 · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? The share of deep revision (L3+L4) rose from 25% (regular AI) to 35% (APCA), 40% (FRAC), and 55% (combined). → Dai 2026

Evidence¶

Dai 2026¶

Dai (2026). Rethinking AI-assisted writing instruction: feedback literacy scripts, calibration training, and student writing development. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1829268

q3 · i? · causal · r2

Revision-depth distribution across the four groups (Figure 7), coded from initial-final draft comparisons into four levels by two trained coders (Cohen's κ = 0.82). The regular AI group's revision was concentrated at shallow levels (35% L1, 40% L2). No effect size is printed.

"When L3 and L4 were considered together, the share of deep revisionrosefrom25%intheregularAIgroupto35%intheAPCAgroup, 40% in the FRAC group, and 55% in the combined group."

Discussion¶

Related Claims¶

• Diagnostic Review revisions contained larger shares of meaning-change, organization, and content/argument work, while Control revisions remained more dependent on surface-level editing — the general claim this finding bears on, with the studies behind it
• In AI-assisted writing, the combined FRAC+APCA intervention produced the highest writing quality gains, with FRAC alone exceeding APCA alone, across T1–T3 — related
• All three intervention groups converted AI feedback into effective revisions more successfully than regular AI use, with the combined group highest in effective adoption rate — related