During checks, save the first full response before modelling the answer. Accept large tokens, tactile access, pictures, printed/raised symbols, pointing, AAC, neutral reading/scribing or pass. A child need not justify a route choice. Keep invented data clearly separate from real evidence.
| Day | Say or show |
|---|---|
| 21 | “Which source field can you see? Which one is not shown?” |
| 22 | “What question are we using to make these groups?” |
| 23 | “Did the badge change, or did our question change?” |
| 24 | “Which is the physical token, which is its picture, and which is the agreed code?” |
| 25 | “Use the new source under both rules. Where can you not tell?” |
| 26 | “What does this made-up record actually say about R?” |
| 27 | “Which source row supports each group?” |
| 28 | “Could another reader tell what H, SL and ? mean?” |
| 29 | “Which rule did you choose? What happened when you used the real screen, if you did?” |
| 30 | “Which trial results are stated, which were not tested, and which empty data kind might belong to you?” |
Record exactly: “Child placed physical badge 5 under circle and unknown border” samples representation. “Child selected BORDER on a running local page with keyboard, placed badge 5 under CANNOT TELL, pressed CHECK and pointed to feedback to answer their question” may support system exploration if actually observed. A drawn device, adult click, fictional plant badge or fictional holder result cannot be promoted to actual device, Science or Design evidence.