PDF PAGE 1

Science and technology classroom starter pack

Prepared for arqiv.si | Three 40 minute lessons | Classroom draft 1.0

Students test a prediction, build a model and improve a design. This pack contains everything a teacher needs to run three separate lessons using paper and ordinary classroom supplies. Activities work offline and require no student accounts or online submissions.

Provisional audience: grades 6 to 8, approximately ages 11 to 14. The grade range is a starting assumption, not a curriculum alignment or classroom-tested recommendation. Review reading level, local safety rules and timing before use.

Choose a lesson

Print and prepare

• Print only the two student pages for the lesson you choose. One set per student is ideal; pairs may share and use extra writing paper.

• Keep teacher pages and answer keys out of student packets. For AI, hand out page 4 only after students finish page 3.

• Print in black and white at 100% on Letter paper. On A4, use Fit to printable area. Students measure real materials, so scaling the printed pages will not alter an activity.

• Read the teacher plan and try the physical setup once. Space needs a clear 3.2 m working length. Engineering needs four matching sheets per group and stable book supports.

Access and participation

Use pairs or groups of three. Offer reader, measurer, builder and recorder roles without assuming every student can grip, fold, see small print or move around the room. Students may explain aloud, dictate or draw instead of writing full sentences. Read table headings aloud and supply enlarged pages when needed.

AI | TEACHER PLAN

PDF PAGE 2

When a pattern fails

40 minutes | Grades 6 to 8 provisional | Pairs | Student pages 3 and 4

Students classify fictional parcels using examples. Two rules fit the training data perfectly; new cases reveal that only one matches the warehouse routing rule. This paper simulation introduces learning from patterns and checking predictions. It is not a trained AI system.

Learning goals

• Distinguish examples used to choose a rule from new examples used to test it.

• Use a count or percentage to describe performance on a named test set.

• Explain why a plausible prediction still needs a reliable check.

Materials and preparation

Per pair: pencils, pages 3 and 4, optional calculator. Keep page 4 and the answer key hidden. The warehouse specification on page 5 is the authority for the fictional labels. Do not reveal it until predictions are complete.

Run the lesson

0 to 5 minutes Ask: If a rule gets every practice example right, must it get the next example right? Take predictions without settling the question.

5 to 13 minutes Give page 3. Define feature, label and training set. Students check both proposed rules against all six training rows, then commit to what the evidence does and does not show.

13 to 22 minutes Give page 4. Students predict both routes for all eight new parcels. Require pen or a clear first draft; predictions must be visible before labels are revealed.

22 to 28 minutes Read the warehouse specification from page 5, then reveal the eight correct labels. Students mark errors and calculate each rule's test score.

28 to 36 minutes Compare training and test scores. Ask why the badge rule appeared convincing, which missing training examples would have helped, and what should happen with an unclear new parcel.

36 to 40 minutes Use the exit question on page 4. Listen for a specific check against an authoritative source, rather than asking another model to agree.

Support and assessment

Read rigid as holds its shape and flexible as bends easily. Use correct out of eight before introducing percentages. A secure response cites the 6/6 training tie, the 4/8 versus 8/8 test result, and the checked specification. Extend by asking students to design one training example that separates the two rules.

Background: NIST recommends realistic, representative tests and warns that performance can generalize poorly beyond training conditions. This activity illustrates that idea; its 50% result is not an error rate for AI systems in general. [1]

AI | STUDENT WORKSHEET 1 OF 2

PDF PAGE 3

Learn a rule from examples

Team __________________________ Date __________________________

A fictional warehouse sends parcels to Route A or Route B. You can see a parcel's structure and badge. The examples below already have checked routes. Your task is to find a rule that could predict the route of a new parcel.

A feature is an input you can describe, such as the badge. A label is the checked answer, such as Route A. A training set is a collection of examples used to choose or fit a rule.

1 Inspect the training set

Rigid means it holds its shape. Flexible means it bends easily. The badge is a printed symbol. All parcels and labels in this activity are invented.

2 Test these two candidate rules

Badge rule: triangle badge goes to A; round badge goes to B.

Structure rule: rigid parcel goes to A; flexible parcel goes to B.

Badge rule matches ______ of 6 checked routes.
Structure rule matches ______ of 6 checked routes.

3 Decide what the evidence tells you

Can these six examples prove which rule the warehouse uses? Explain with one example or comparison from the table.

Describe a new parcel that would make the two rules disagree.

Stop here. Keep your first answers. Your teacher will give you a new test set.

AI | STUDENT WORKSHEET 2 OF 2

PDF PAGE 4

Test and verify the rule

Team __________________________ Use this page after worksheet 1

4 Predict before checking

Use both rules from page 3. Predict A or B for each parcel before your teacher reveals the answers. Leave Checked route blank.

5 Check the predictions

Circle predictions that disagree with the checked route. For percent correct, divide the correct count by 8, then multiply by 100.

Badge rule: ______ correct out of 8 = ______ % correct
Structure rule: ______ correct out of 8 = ______ % correct

Which rule is better supported? Use the scores and the warehouse specification.

6 Decide when to ask for help

A semi-rigid parcel does not fit either structure category clearly. What should the warehouse do before routing it?

Exit question

An AI answer sounds confident. What would you check before treating it as a fact, and why?

AI | TEACHER ONLY

PDF PAGE 5

AI answer key and discussion

Keep this page out of student packets

Reveal after students record predictions

Warehouse specification: Route A accepts rigid parcels. Route B accepts flexible parcels. Badges do not determine routes. A parcel whose structure is unclear must be checked by the responsible warehouse staff.

Worksheet 1

Both rules match 6 of 6 training examples, or 100%. The training set alone cannot distinguish them: every rigid parcel has a triangle badge and every flexible parcel has a round badge. A rigid parcel with a round badge, or a flexible parcel with a triangle badge, would separate the rules.

Worksheet 2

Badge rule: 4/8 = 50%. Structure rule: 8/8 = 100%. The badge rule fails on T3, T4, T5 and T6. The structure rule fits both the new examples and the explicit specification.

For semi-rigid: accept “pause and ask the responsible person” or “check a more precise current specification.” Guessing from the badge does not resolve the missing definition.

What to listen for

A strong explanation says that a pattern can fit a limited dataset for the wrong reason. Here, the training data omitted both mismatched combinations. Changing the badge rule after seeing these labels is an improvement step, but a fair new evaluation would need further unseen cases.

For the exit question, accept a specific check such as opening an original source, checking the evidence and date, reproducing a calculation or consulting a qualified person for a consequential decision. Agreement between two generated answers is not independent verification.

Keep the analogy accurate

Students manually compare two rules. Many machine-learning systems instead fit numerical parameters using an algorithm. This exercise does not model all AI, explain how a chatbot generates text, or show that any system will be 100% accurate outside its test conditions. [1]

SPACE | TEACHER PLAN

PDF PAGE 6

How far apart are the planets

40 minutes | Grades 6 to 8 provisional | Groups of three | Student pages 7 and 8

Students calculate distances at a common scale and build a Sun-to-Neptune model. The aim is to replace evenly spaced planet pictures with measured distances, then explain what this one-dimensional model leaves out.

Learning goals

• Convert an astronomical-unit distance to centimetres using a scale factor.

• Measure every position from the same zero point and distinguish positions from gaps.

• State two limitations of a scale model.

Materials and preparation

Per group: 3.2 m of string or paper strip, ruler or tape measure, nine scrap-paper labels, tape for fixing labels, pencil, and optional calculator. Use joined desks, a clear wall or a teacher-led model. Mark the Sun at zero. Keep labels small enough to show the first four positions. Do not stretch the string.

Safety: keep string off walkways and away from necks. Secure loose ends. No student needs to climb, run or look at the Sun. A seated group can measure on a table while another person places labels.

Run the lesson

0 to 5 minutes Ask students to sketch where eight planets might lie on a line from the Sun. Save the sketch to compare with the measured model.

5 to 10 minutes Introduce AU, the shared zero point and the scale 1 AU = 10 cm. Work Earth's example. Explain that these are approximate orbital-distance values, not today's positions. [2, 3]

10 to 18 minutes Students complete page 7. Partners check multiplication and units before construction. A misplaced decimal should be caught here.

18 to 28 minutes Build the model. Students measure from the Sun every time, mark each position and attach its label. Check that Neptune lands at 300.6 cm.

28 to 36 minutes Use page 8 to calculate gaps and explain the model's limits. Compare the first sketch with the measured positions.

36 to 40 minutes Ask for an exit explanation: What does the model show well, and what might a viewer wrongly assume?

Support and assessment

Provide completed distances if arithmetic would block the science. Offer a shared model with students directing placement verbally. With only 1.6 m available, build at 1 AU = 5 cm by halving the completed 10 cm-scale positions; keep the worksheet calculations labelled with their original scale.

Check for units, a common origin and the statement that planets are neither fixed nor lined up this way. Planet-label sizes are not scaled. Full answers and misconceptions are on page 9.

SPACE | STUDENT WORKSHEET 1 OF 2

PDF PAGE 7

Calculate and build the model

Team __________________________ Date __________________________

An astronomical unit, or AU, is about 150 million kilometres, approximately Earth's average distance from the Sun. We will use 1 AU = 10 cm in a model. The table gives approximate orbital-distance values for the planets. [2, 3]

1 Convert the distances

Multiply the distance in AU by 10 to get the model distance in centimetres. Earth is worked for you. Keep one decimal place where needed.

Partner check: the closest planet is ______ cm from the Sun.
The farthest planet is ______ cm from the Sun.

2 Build on a string or paper strip

1. Lay out at least 3.2 m of string or paper. Mark one end Sun and 0 cm.
2. Starting at the Sun each time, measure and mark each planet's position.
3. Add labels beside the marks. A label's size does not represent planet size.
4. Check Earth at 10 cm and Neptune at 300.6 cm.
5. Compare the first four planets with the full length of the model.

Do not measure the next planet from the previous planet. Each table entry is a distance from the Sun, not the gap between neighbouring planets.

3 Record an observation

Describe one way your measured model differs from an evenly spaced picture of the planets. Include a measurement.

SPACE | STUDENT WORKSHEET 2 OF 2

PDF PAGE 8

Explain what the model shows

Team __________________________ Use the 1 AU = 10 cm scale

4 Calculate the gaps

To find a gap, subtract the nearer position from the farther position. Show your calculation and include units.

a. What is the gap between Earth and Mars?

b. What is the gap between Jupiter and Saturn?

c. Which gap is larger, and by how much?

5 Read the scale

How many times as far from the Sun is Jupiter as Earth in this model? Explain using the positions from your table.

A classroom has only 1.6 m of clear space. At 1 AU = 5 cm, how far from the Sun would Neptune be? Would it fit?

6 Challenge the picture

A visitor says, “Your model shows where all the planets are today, and your paper labels show their actual sizes at this scale.” Give two corrections.

Exit question

Finish both thoughts: This model helps me understand … It does not show …

SPACE | TEACHER ONLY

PDF PAGE 9

Space answer key and discussion

Reference scale 1 AU = 10 cm

The closest planet is Mercury at 3.9 cm. The farthest is Neptune at 300.6 cm, or 3.006 m. Every position is measured from the model Sun.

Worksheet 1 observation

Answers vary. A supported example: “The four inner planets fit within 15.2 cm of the Sun, while Neptune is 300.6 cm away.” Accept another accurate observation with a measurement. Avoid treating equal-size labels as equal-size planets.

Worksheet 2 calculations

4a. Earth to Mars: 15.2 – 10.0 = 5.2 cm.
4b. Jupiter to Saturn: 95.4 – 52.0 = 43.4 cm.
4c. Jupiter to Saturn is larger by 43.4 – 5.2 = 38.2 cm.
5. Jupiter is 52.0 / 10.0 = 5.2 times as far from the Sun as Earth.
At 1 AU = 5 cm, Neptune is 30.06 x 5 = 150.3 cm, or 1.503 m. It fits within 1.6 m.

Two essential corrections

• The line represents approximate orbital distances measured from a common origin. Planets move around the Sun and are not generally arranged on one straight line. It does not show their current positions.

• Paper labels are markers. Their widths do not represent scaled planet diameters. This activity scales distance only.

Scientific precision for the teacher

The AU values are reproduced from NASA/JPL's scale reference and are rounded educational values. They approximate the semi-major axes of the planetary orbits, commonly described in classroom materials as average distances from the Sun. Do not treat them as a measurement for a specific date. Values in other references may differ slightly through rounding. [2]

One AU has an exact defined value of 149,597,870,700 metres; “about 150 million kilometres” is sufficient here. The model does not show orbital eccentricity, inclination, motion, planet diameters or other solar-system objects. [3]

ENGINEERING | TEACHER PLAN

PDF PAGE 10

How folds change a bridge

40 minutes | Grades 6 to 8 provisional | Groups of three | Student pages 11 and 12

Students compare flat and folded paper bridges using repeated tests. They hold the material, span and loading method constant, then use the results to propose one design change. Success means a defensible test and explanation, not the highest coin count.

Learning goals

• Change one design feature while keeping test conditions consistent.

• Record repeated results, including a test that stops before failure.

• Use evidence to propose a revision and explain a limit of the evidence.

Materials and preparation

Per group: four fresh sheets of the same copy paper and size; two stable book stacks; ruler; one light paper cup; 20 identical coins; pencil; timer. No cutting, tape or glue. A fifth sheet is optional for testing a redesign. Teacher: try the setup first, use low supports and choose coins with a combined mass no greater than 200 g.

Safety: test on a clear table, keep feet and hands clear of falling coins, and keep small objects out of mouths. Stop at 20 coins even if the bridge holds. Do not stand on, climb on or load a real bridge.

Run the lesson

0 to 5 minutes Show one flat sheet and one lengthwise accordion-folded sheet. Ask whether shape can change resistance to bending while the amount of paper stays the same. [4]

5 to 10 minutes Set a 15 cm gap between book stacks, each 4 to 6 cm high. Demonstrate the same centred cup and one coin added every five seconds. Explain the stop conditions on page 11.

10 to 16 minutes Groups predict, identify controls and prepare two matching folded sheets. Give pre-creased sheets or a folding partner where needed.

16 to 29 minutes Run four trials with a fresh sheet each time: flat 1, folded 1, flat 2, folded 2. Students record the last load held for five seconds and what ended the test.

29 to 36 minutes Complete page 12. Compare repeats, explain differences and sketch one revision. Testing the revision is optional; the core lesson ends with an evidence-based plan.

36 to 40 minutes Ask: What can your results support, and what would require another test? Use the exit question as the assessment.

Support and troubleshooting

Use builder, loader and recorder roles; students may direct a partner. If the empty cup causes failure, record 0 and keep the same cup for all trials. If every design reaches the cap, report an inconclusive comparison at this limit. Do not increase the load to force a result. See page 13 for interpretation.

ENGINEERING | STUDENT WORKSHEET 1 OF 2

PDF PAGE 11

Build and run a fair test

Team __________________________ Date __________________________

Your question: Does an accordion fold let a paper bridge carry more coins than a flat sheet under the same test conditions?

1 Plan and predict

I predict __________________ will carry more because __________________

Keep these the same: paper type and size, 15 cm gap, support height, cup, coin type, cup position and loading pace. Change only the paper shape for this comparison. Use a fresh sheet for every trial.

2 Make two matching folded bridges

Divide the short edge into six roughly equal panels. Fold five alternating creases along the long direction. Choose a zigzag depth from 1 to 2 cm and match it on both folded sheets. Creases run from one book support to the other.

Diagram not to scale. For every trial, span the gap with the sheet's long direction, centred with equal overlap on both books. Do not fix it to the books.

3 Test in the same way each time

1. Put a fresh sheet over the supports. Place the empty cup at the centre.
2. Add one coin. Wait five seconds before adding another.
3. Stop if the bridge slips, collapses or touches the table, or the cup tips or touches it.
4. Record the last coin count held for five seconds before that happened.
5. At 20 coins held for five seconds, stop and record 20+; failure was not reached.

Record 0 if the empty cup fails. An accidental spill, knock or moved support invalidates the trial; repeat with fresh paper if possible. Never exceed 20 coins.

ENGINEERING | STUDENT WORKSHEET 2 OF 2

PDF PAGE 12

Use the evidence to improve a design

Team __________________________ Use your results from page 11

4 Compare the repeated tests

What loads did your flat bridges hold? __________ and __________ coins
What loads did your folded bridges hold? __________ and __________ coins

Which shape performed better, or is the comparison inconclusive? Cite both trials for each shape. If a result is 20+, say at least 20.

Did the two trials of the same shape agree? Give one setup or construction difference that might explain any variation.

5 Plan one improvement

Keep the same paper, span and loading method. Sketch one change to the folded bridge. Label what you will change and explain how it might help.

What will you keep the same so the new test is a fair comparison?

Exit question

Your bridge held 20 coins and testing stopped. A classmate claims, “It can hold exactly 20 coins and will fail at 21.” Is that supported? Explain.

ENGINEERING | TEACHER ONLY

PDF PAGE 13

Engineering answer key and discussion

There is no required winning design or target coin count

What counts as a fair test

Students change the sheet shape while holding paper type and size, span, support height, cup, coin type, load position and pace constant. Fresh sheets limit damage from previous loading. Repeated trials expose variation from folding, alignment, slippage and placement. Coin counts can be compared between groups only if the loading objects and other conditions match.

How to record results

• If a bridge holds 7 coins for five seconds but fails after the eighth is added, record 7. If the empty cup makes it fail, record 0.

• If it holds 20 coins for five seconds, record 20+ and the reason “test limit.” This means at least 20 under these conditions. Its failure load is unknown.

• A tipping cup ends a valid test. An accidental spill, knock or shifted support makes the trial invalid; repeat with a fresh sheet if possible. Do not quietly replace an inconvenient result.

An example of a supported conclusion

Illustrative data only: flat trials hold 2 and 3 coins; folded trials hold 12 and 14. A suitable conclusion is: “In our setup, both folded bridges held more coins than either flat bridge. Folding may have helped the bridge resist bending.” These numbers are invented examples, not expected classroom results.

If both shapes reach 20+, students cannot rank their maximum capacities from this test. If repeated ranges overlap or one bridge slips early, discuss uncertainty and the need for better-controlled repeats. A result that differs from a prediction is still useful evidence.

Explain the engineering

Folds change the cross-section and place parts of the paper farther from the middle of the section. This can increase bending stiffness, so the sheet may bend less under the same load. The paper material itself has not become inherently stronger. Shape, fold depth, local buckling, support contact and slippage all affect the tested bridge. [4]

This activity measures a practical load limit using specified stop conditions. It does not directly measure stiffness, identify every failure mechanism or certify a structure as safe.

Acceptable redesigns

Examples include changing the number or depth of folds or forming two longitudinal folded edges. Require a labelled change, a plausible reason and controlled conditions. Students should test the revision before claiming it works. Avoid rewarding decoration or unsupported claims that more folds must always be better.

Exit answer and quick assessment

The “exactly 20” claim is unsupported. The test showed that the bridge held at least 20 coins for five seconds; no test at 21 took place. Look for an evidence-based comparison, sensible controls and recognition that a stopping limit is not a measured failure load.

REFERENCE | KEEP WITH TEACHER COPY

PDF PAGE 14

Sources and teacher notes

These references support the factual background. The parcel dataset, prompts, explanations and worksheet layouts are original to this pack. The activities are classroom models, not new scientific findings. Source organizations have not reviewed or endorsed the pack.

1 AI testing and reliability

National Institute of Standards and Technology. AI Risk Management Framework 1.0, section 3.1, “Valid and Reliable.” Supports realistic test conditions, representative test sets and limits to generalization beyond training conditions.

https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/

2 Planetary orbital distances

NASA Jet Propulsion Laboratory. Solar System Sizes and Distances Reference Guide; Make a Scale Solar System. The reference supplies all eight AU values. The activity explains scale conversion and the distinction between scaled planet sizes and distances. Here, each model position is the AU value multiplied by 10.

https://www.jpl.nasa.gov/edu/pdfs/scaless_reference.pdf

https://www.jpl.nasa.gov/edu/resources/project/make-a-scale-solar-system/

3 The astronomical unit

International Astronomical Union. Resolution B2, 2012. Defines the astronomical unit of length as exactly 149,597,870,700 metres. The student text uses the rounded value of about 150 million kilometres.

https://iauarchive.eso.org/static/resolutions/IAU2012_English.pdf

4 Structural shape and bending

Australian National University. Paper Bridge Challenge. Supports the flat-versus-folded setup and lengthwise fold direction. Mark Schenk and Simon D. Guest, University of Cambridge, Origami Folding A Structural Engineering Approach. Supports direction-dependent changes in bending stiffness from folding.

https://science.anu.edu.au/engagement/science-lab-experiments-home-school/paper-bridge-challenge-science-strong-shapes

https://www3.eng.cam.ac.uk/~sdg13/preprint/5OSME.pdf

Before public use

• Confirm the intended age range, local curriculum, classroom language and accessibility needs. Grades 6 to 8 remain provisional.

• Pilot each lesson, especially the paper bridge. Check the real materials, table layout, timing and whether the load cap produces useful comparisons.

• Retain source links and these qualifications when adapting the worksheets. Keep teacher answer keys separate from student downloads.

Source check date 2 October 2026 | Version 1.0 | Teacher review pending