Component Test

Weight 20% · Capability: Isolate · Due 9 April 2027, end of Week 6

Isolate one structural variable and test it under at least three conditions, using a protocol that would survive being handed to someone else.

Task at a glance#

• Mode — individual.
• Format — a test report: question, declared variable, protocol, comparison table, annotated evidence, interpretation and limitations.
• Scale — one variable, three or more conditions, plus a control where the question supports one.
• Evidence — physical specimens, non-food analogues, or the supplied measurement packs.
• Route — one file, submitted to the SLOP2895 submission point before the deadline.

The brief#

The Field Log recorded assemblies as they were found, with their histories already written into them and mostly unreadable. This task removes that ambiguity by building the specimen instead of meeting it. You choose one variable — stiffness category, moisture placement, interface grip, compression, layer thickness — and you change only that, across conditions you declare in advance.

The declared problem is that an assembly has too many variables to argue about honestly. Any observed difference between two whole sandwiches has a dozen candidate causes, and a comparison that changes three things at once cannot attribute the result to any of them. Isolation is the move that makes a difference mean something: hold everything else as close to constant as you can, vary one thing deliberately, and the result belongs to that thing rather than to the assembly in general.

A protocol is the second half of that argument. A procedure that only its author can execute is not a protocol. Write it so that another student in this course could run it from your document alone, get a result, and compare it with yours — and then say plainly what your test could not hold constant, because that list is part of the finding and not a confession appended to it.

A test whose result contradicts what you expected is not a failed submission. A well-designed, well-documented test that returns an inconvenient result marks higher than a loose one that returns a tidy story.

Required submission#

1. A focused research question, narrow enough that three conditions can answer it.
2. One declared variable under investigation, with the property being varied defined in measurable terms.
3. A repeatable test protocol: specimen preparation as a structural specification, the fixed conditions, the loading or exposure applied, what is measured, and how many times.
4. A comparison table covering every condition, including the control where one is used.
5. Annotated evidence — photographs, diagrams or measurement records — keyed to the conditions in the table.
6. An interpretation and limitations section.

Recommended process#

1. Declare the question and the single variable before you touch a specimen, and write down what you expect to happen.
2. Classify the conditions you will test, and state which one is the control or why the question does not support one.
3. Document the protocol in full, including every quantity you intend to hold constant across conditions.
4. Test each condition in the same way, in a stated order, and record what you did differently whenever you were forced to deviate.
5. Measure the response at the same point and in the same units for every condition.
6. Compare the conditions against each other rather than against your expectation.
7. Revise the protocol once, run at least one condition again under the revised version, and report both runs.
8. Defend the interpretation by naming the result that would have overturned it.

Evidence standard#

Keep the layers of your report separable. Observation is the measurement as recorded, including the ones that disagree with the others. Inference is the claim you draw about the variable, and it must point to the specific conditions that support it. Uncertainty is the spread across repeats, the resolution of your measurement method, and the readings you are not confident in — recorded as uncertain rather than quietly dropped. Limitation is what the protocol could not control: specimens that were not identical to begin with, ambient conditions that drifted, a measurement taken by eye.

• Observation

  Drawn solid: present in the evidence.

• Inference

  Drawn dashed: reasoning laid over the evidence.

• Uncertainty

  Drawn as an open span: the evidence fixes no one value.

• Limitation

  Drawn as far as the method reached, and no further.

Three conditions with one honest repeat and a stated measurement resolution is stronger evidence than six conditions measured once each. Fabricated or reconstructed measurements are not evidence and are marked as absent.

Equivalent participation#

Three routes are available, and all three assess the same outcome:

• Physical specimen — test real materials, with no requirement to consume anything you test.
• Non-food analogue — sponge, foam, cardboard, clay or textile standing in for the layer whose property you are varying. This route is often the better engineering choice here: analogues are more uniform between specimens, which is exactly what an isolation test wants, and they survive repeated destructive testing.
• Supplied evidence — the course measurement packs, which carry enough conditions to support a comparison.

No route loses marks, and the marking criteria are identical across all three. See the participation policies.

Safety, cost and waste#

No expensive or specialist materials are required, and premium ingredients earn nothing. Your protocol must justify the quantity of material it consumes; use the smallest specimen that can answer the question, and prefer a non-food analogue where the test is destructive and repeated. Tested material must never be represented as safe to eat. Full detail is on the policies page.

Preparation#

Weeks 4 to 6 build the isolation capability, with the measurement habit set earlier.

• Week 2: Bread Is a Beam — load steps, deflection, and recording a reading you are unsure of as uncertain.
• Week 4: Moisture Is a Load — a progressive boundary compromise, and how to stage it as conditions.
• Week 5: Load-Bearing Fillings — separating layers that support from layers that only add load.
• Week 6: Sauce at the Interface — the point at which an interface layer turns from binder to lubricant.