Keep this as a short curiosity hook — no loading yet. Point at one finished build from earlier in this project (do not load it yet). Ask for hands: look tougher, or holds more? Steer the class to the number held.

Link to the project: groups already have a full-size load-bearing build and success criteria from earlier sessions (for example hold four library books, hold a tablet without tipping, take wet coats). Keep those criteria on the wall or IWB in one line per group if useful.

Key question: Engineers show an improvement with a measurement, not an opinion. What number will we write down today?

Board stays short: one big idea (load test = write the number held) and the question. Bring in one change as the label when groups fix, and evidence when they compare before and after numbers. Use the full table below yourself; do not project it.

Worked model (keep inside five minutes): Pre-load the demo scrap span once before the lesson (or in the first 30 seconds) and write before: 6 on the board so the class does not wait through two full load-to-failure runs. Live, think aloud only the change and the re-test:

• I wonder whether one brace will raise how many coins this little span holds.
• I predict it will hold more after a triangle brace because triangles hold their shape and stop the span sagging sideways.
• Before (already on the board from the quick pre-load): 6.
• I change one thing — tape one folded triangle under the middle. Nothing else. (Tape only so the re-test is fair straight away.)
• I re-load the same way — coins in the same place, same speed. It holds 11. I write after: 11.
• I think the brace worked because the number went up by 5. If the number stayed the same, the change did not earn a gain. That rise is our evidence.

Misconception to head off: children may say the build is better because it looks neater. Keep asking what do the numbers say?

Ideas they may use for the one change (from earlier in this unit): triangle brace; arch or rolled tube for a weak middle or beam; taped triangular gusset instead of plain tape; stop a pivot join that should not move.

Before this step

Fallback if a full prototype is missing or damaged: that group uses their half-size test piece from earlier in the project, or a quick scrap span like the teacher demo, still with an agreed load and a clear failure rule. A low before number is still useful data.

Groups collect the build they made earlier in this project, their Design Brief success criteria, and the real load or agreed weights. Clear floor space so nobody stands under a failing build.

How to run it

• Roles are on the board: one loader, one counter, one safety spotter per group. Rotate if helpful.
• Same load type and same place on the build every time (fair test).
• Count out loud so the group agrees the number.
• If it reaches the target without failing, the before number is the target (note that it passed).
• On the PrototypeEval page: job, target, number held before, where it failed or sagged (shape or join).

What good looks like: a clear before number and a named weak point (for example middle of flat beam, tape peel at a join, tip at the front feet).

If a build fails immediately: record held = 0 or 1, name the fail point, and move to peer advice. The low number is still useful data.

Label in talk: this is the load test and the before number.

Differentiation: less confident groups use large easy-to-count loads (books). Confident groups may use smaller units (coins in a tub) for a finer number, still one agreed unit.

In-place peer visit (do not move full-size builds): prototypes stay on their desks or floor spots. Pair neighbouring groups. About 3 minutes looking at the host build, 2 minutes giving advice, then reverse so each group hosts and visits (about 7 minutes total including the reverse). Keep advice tied to ideas from earlier in this unit (triangles hold shape; arches and braced middles or beams; gussets and firm joins versus pivot joins).

Sentence stems to post: The weakest part we see is… One idea that might help is… because…

Your job while circulating: stop groups from listing five fixes. They may hear several ideas but must pick one change to make next. Write that one change on the PrototypeEval or ResultsAndConclusion page before they touch materials. Block any attempt to drag a coat rail or shelf across the room.

Fold the class in: after one pair of groups shares aloud, ask the room Do you agree that is the weak point? Which idea matches: triangle, stronger middle shape, or stronger join?

Rules that keep the second test fair

• One change only (shape or join fix, not both piles of extras).
• Same load units and placement as the before test.
• Same failure rule.
• Rebuild only what the change needs; do not secretly rebuild the whole model stronger everywhere.
• No wet glue. The timed change must use sticky tape, pre-cut card, split pins or lolly sticks so the join holds for an in-session re-test. Wet PVA will not set in time and is not fair evidence.

Timing (protect the re-test — say the stop aloud and keep it on the board): hard mid-step stop after about 6–7 minutes of fixing. Call a 2-minute warning, then every group re-loads even if the fix is simple. Roughly 6–7 minutes to make the change, 5–6 minutes to re-load and record. Journal order after the re-test: after row on PrototypeEval first, then before / one change / after / gain on ResultsAndConclusion. Pre-cut triangle and gusset scraps sit on the central tray so groups are not cutting from scratch. Tape-only fallback: a single taped triangle or taped overlap is enough for a valid after number.

Journal (both pages): PrototypeEval completes the after row (number held and where it failed or sagged). ResultsAndConclusion holds before number, the one change in plain words, after number, gain (after take away before, or stayed the same / went down), and a one-line conclusion (The change worked because… or The number did not rise, so next time we would try…). Label the rise (or flat or lower result) as evidence.

Unexpected results are data: if the number falls, that is still science. Ask what the change may have weakened. Pupils write went down / no gain, not a blank.

Safety: same as the first load. Adult help for difficult cuts. Spare tape, pre-cut card strips, split pins and gusset scraps on a central tray so groups are not hunting the cupboard.