Wanted outcome: first reasoning preserved, then a review against a later source. Prepare: educator selects only E-FIRST and its own blank. E-LATER stays separate until first capture; no source values are displayed in the ordinary opening. 30 = 3 orient + 3 exact briefing + 10 first + 2 capture + 8 later + 4 close.
0–3: remind learners that the new record is fiction. 3–6: read exact task directions and offer access, without content hints. 6–16: Route 5A first. 16–18: capture first response. 18–26: issue only E-LATER and its own blank; learner reviews prediction, compares the new values and states a limit. 26–30: collect later response and choose targeted next teaching. Keep, revise or decline certainty can all be justified; no retrospective coaching of first evidence.
- Route 5A · Assigned clean check. Complete the separate first/later tasks under assessment administration; criterion coverage is specific to this case. No peer model or key enters the first response.
- Route 5B · Ordinary error clinic after capture, or instead of the check. With O2, repair a fictional note “38 − 26 is A’s growth”. Calculate both start-to-end increases and identify what 38 − 26 actually compares. Add nine-day period and a limit about the different positions.
- Route 5C · Ordinary missing-data design after capture, or instead. O2 lacks temperature and measured light. Choose which to record for a stated question, specify instrument/unit/time if a teacher could provide it, and explain why collecting a new variable would help but not repair the old record retrospectively. Mark outcome not observed.