Goal: make a prior prediction grounded in experience. Proposed opportunities: AC9SFH01, AC9SFI01, AC9SFI02, AC9SFI04.
Prepare: approved sample not yet viewed against the puppet; no-cover baseline; setup and prediction log. Model option: keep one supplied notebook result covered until after prediction.
- Choose the job · 3 min. “Should this window show Pip's circle or hide it?” Child chooses. Check the uncovered target is available from the viewing place. If not, pause/adjust, change to an accessible property or use an attributed report; do not attribute nonvisibility to the cover without an available baseline.
- Predict · 4 min. Ask “What do you think you will see through this sample, and what makes you think that?” Record exact words before the first view. “Not sure” is welcome.
- Try one view · 7 min. Put the sample at the marked position. Child observes or receives a named adult report. In the model route reveal the specified row; do not present it as an experiment.
- Compare · 5 min. Read the prior words beside the later observation. Ask “Same as you expected, different, or unclear?” Avoid turning a later guess into an earlier prediction.
- Use the result · 4 min. Would this sample help the chosen show/hide job in this setup? Accept changing the choice after seeing the result.
- Keep and next · 2 min. Save both records and any interruption. Tomorrow we compare samples from the same viewing place.