HCD.19:11 - SoTA-Echoing

The practice question is how to distinguish useful instruction, help, and repair without mistaking immediate success for a stronger learning result. The selected line inspects the material–task relation, preserves assistance-sensitive observations, and tests meaningful continuation. Compared with a single post-reading score, it costs a preserved first attempt and a targeted follow-up but yields a repair and a narrower, defensible claim.

Practice choice	Adopt, adapt, or reject; effect on this pattern	Source role, limits, and reopen condition
Preserve help when interpreting performance	Adapt assistance-sensitive comparison in sections 4.2 and 4.6 and the M0/M1 case. A result obtained with the answer in the worked example is still useful, but cannot establish independent acquisition.	Bastani et al. (2025) provide failure evidence for inferring later independent performance from assisted success in a particular school-mathematics AI arrangement. Its safeguards were bundled; this does not identify a universal mechanism. Reopen when the promised equipped-work arrangement or later independent claim differs.
Follow the relation the reader must construct across parts	Adapt principle/example integration and self-explanation in sections 4.3–4.4. Local recall is an inadequate comparator when the promised work combines separated concepts.	McCrudden et al. (2022) and 2024 contribute bounded topic-specific integration evidence. Their prompts count as support; they establish no universal principles-first sequence. Reopen for a different audience, relation, or route that makes another order preferable.
Preserve purposeful preparation while repairing unsupported difficulty	Adapt problem solving followed by instruction in section 4.6; reject both unconditional simplification and unqualified “productive failure.”	Sinha and Kapur (2021) synthesize conditional effects, largely in mathematics and physics. The practical requirement is that later instruction resolves the attempted alternatives. Reopen when that connection is absent or a different learning task changes the comparison.
Choose a representation repair by its function	Adapt mechanism-specific multimedia design in section 4.3 and the local-route case. A signal, correspondence, or pacing repair can target a difficulty without adding decoration throughout the course.	Noetel et al. (2022) supply a synthesis of multimedia design research, not proof of whole-book continuity. Reopen if the repaired representation still leaves the required relation unavailable.
Keep structural diagnosis separate from experienced burden	Adopt appropriate respondent and timing conditions for a selected human-burden observation; section 5.4 first retains the human report and narrows the untested claim, then selects new observation only for its obtainable, worthwhile contribution.	Krieglstein et al. (2025) provide failure evidence about retrospective load timing in studied materials. They supply no conversion from narrative epiplexity to human burden. Reopen for a changed segment, timing, instrument, or respondent population.