Assess the inquiry—not just the model
A visually convincing 3D model can be satisfying, but appearance alone says little about what a learner understood.
An ASCAND activity can reveal learning through the decisions made around the reconstruction. Learners can frame a question, predict what the fixed camera will observe, record relevant conditions, inspect the result, distinguish observation from explanation, identify uncertainty and propose a better test.
The reconstructed model is evidence within that process. It is not an answer key. Missing, incomplete or unexpected geometry can be especially valuable when learners use it to reason about visibility, surface response, occlusion, capture conditions or the limits of the selected method.
Assessment can therefore focus on how well a learner builds and revises a claim from evidence. The objective is not to reward every imperfection, nor to ignore technical quality. It is to ask what the result supports, what it does not support and how the learner knows.
This page provides a reusable inquiry and formative-assessment framework. It is not a validated rubric, grading scheme, standards mapping or promise of educational effectiveness.
Use the curriculum framework
Review the classroom sequence
Look for evidence across the whole investigation
Useful evidence of learning can appear before, during and after capture.
Question framing
Can the learner state a bounded question that can be investigated with observable evidence? “Why did the scan fail?” may become more useful when reframed as “Which parts of this surface were visible and distinguishable from the fixed camera position?”
Prediction
Does the learner predict an outcome and give a reason? A prediction creates a traceable starting point. It also reduces the temptation to describe any result afterward as expected.
Conditions and control
Can the learner identify relevant conditions and distinguish what was held constant from what changed? These might include object orientation, framing, illumination or the selected capture method. The expected level of control depends on the activity.
Observation
Can the learner describe what is actually visible in the physical object, capture evidence or reconstructed result without immediately turning that description into a causal claim?
Interpretation
Can the learner connect a claim to specific evidence and consider another plausible explanation? A reconstruction difference may have more than one possible cause.
Uncertainty and limits
Can the learner state what remains unknown? This includes recognizing hidden geometry, ambiguous surface behavior, representation limits or a conclusion that requires direct measurement rather than model inspection.
Comparison and revision
Can the learner compare a prediction with a result, or one investigation with another, and revise the explanation in a way that follows from the evidence?
Communication
Can the learner present the question, conditions, evidence, claim, uncertainty and proposed next step clearly enough for someone else to inspect the reasoning?
Technical operation can also be assessed when it is part of the stated learning purpose. It should not silently replace conceptual understanding. A learner may follow a procedure successfully without explaining the evidence, while another may diagnose an imperfect result with strong reasoning.
Explore subject lenses
Plan an investigable question
Preserve the reasoning from prediction to revision
An inquiry record makes learning visible across time. It is most useful when learners complete it during the investigation rather than reconstructing the entire reasoning afterward.
1. State the question
Name the relationship being investigated and the evidence that could help answer it.
2. Record the prediction
Describe the expected outcome and the reason for expecting it. Drawings, annotations or short statements may be appropriate depending on the activity.
3. Document relevant conditions
Record the object, orientation, capture method and other variables that matter to the question. Do not collect detail merely because it is available.
4. Separate observation from interpretation
First record what can be seen or measured. Then state what that evidence may mean. Keeping these fields separate exposes unsupported leaps in reasoning.
5. Connect claims to evidence
Identify the particular feature, comparison or record that supports each claim. “The model looks wrong” is less useful than naming the region and the observed difference.
6. Name uncertainty and alternatives
State what the evidence cannot distinguish and what other explanation remains possible. Confidence is not the same as certainty.
7. Revise the claim or define the next test
Conclude with a more precise claim, a purposeful change to one variable, a different evidence source or a justified decision to stop.
The record may include photographs, screenshots, annotations, direct measurements or model-derived observations only when their use is appropriate and permitted. It should preserve the provenance of evidence so that physical measurements are not confused with values inferred from reconstructed geometry.
Find supporting resources
Review learner roles
Use feedback to improve the next decision
Formative feedback is most useful when it moves the inquiry forward.
Instead of supplying the explanation immediately, an educator can ask:
- Which observation supports that claim?
- Are you describing the result or explaining it?
- What changed between the two attempts?
- Which variables were kept sufficiently stable?
- What other cause could produce a similar result?
- What does the reconstruction not reveal?
- Which conclusion requires a physical measurement?
- What single change would make the next comparison more informative?
- What evidence would make you revise your explanation?
These prompts do not assume that every inquiry must end with another scan. A learner may need to inspect the source evidence, choose another object, compare a different representation or conclude that the question cannot be answered with the available method.
Feedback can also recognize good restraint. Identifying insufficient evidence is stronger reasoning than making a confident claim that the data do not support.
Inspect, explain and iterate
Understand educational boundaries
Make comparison part of the evidence
Comparison becomes informative when learners know what they are comparing and why.
A physical object and its reconstruction are not equivalent forms of evidence. The physical object contains surfaces and properties that may not be captured. The model is a computed representation derived from observations and processing choices. Differences between them can support inquiry, but they do not automatically identify a cause.
The same caution applies across two scans. If orientation, illumination, framing, object placement and processing choices all change, the results may differ without revealing which change mattered.
Ask learners to record:
- the purpose of the comparison;
- the variable intentionally changed;
- the conditions intended to remain stable;
- the observed difference;
- the interpretation of that difference;
- remaining uncertainty;
- the next decision.
Iteration is purposeful when it tests a claim. Repeating capture until a model “looks better” may improve the output, but it does not necessarily produce interpretable evidence.
A justified decision not to repeat is also valid. The learner may determine that the object is unsuitable, that hidden geometry requires another orientation, that critical dimensions need direct measurement or that the available evidence cannot answer the original question.
Review model evidence
Assess object suitability
Build criteria around the learning purpose
An assessment framework should reflect the intended learning, not every possible feature of the workflow.
Adaptable criterion families include:
| Criterion family | Evidence to examine |
|---|---|
| Question | The question is bounded, investigable and connected to observable evidence. |
| Prediction | The prediction is recorded before the result and includes reasoning. |
| Evidence | Relevant conditions and observations are documented with clear provenance. |
| Interpretation | Claims are connected to evidence and separated from description. |
| Uncertainty | Limits, alternatives and unsupported conclusions are recognized. |
| Revision | The claim or next test changes purposefully in response to evidence. |
| Communication | Another reader can follow the relationship between question, evidence and conclusion. |
These are design categories, not a published rubric. An individual lesson must define which criteria matter, what evidence learners will produce and how performance will be interpreted in that setting.
A scoring scale should not be added by copying generic labels onto these categories. Descriptors need review for the specific objective, learner group, subject, language and context. If several educators will score the same work, they may also need shared examples and moderation.
Choose a verified lesson
Review the lesson-plan record
Assessment claims require their own evidence
This framework does not establish that an activity is suitable for a particular grade, aligned with a named standard, valid for a grading decision or proven to improve learning.
Before publishing an assessment instrument or claim, verify:
- the specific learning objective and intended learners;
- the evidence the activity actually produces;
- the relationship between criteria and objective;
- subject, grade and standards claims;
- clarity, fairness and accessibility of prompts and descriptors;
- language and translation quality;
- scoring, moderation and interpretation procedures;
- privacy, safeguarding, storage and retention of learner work;
- local institutional policies and required accommodations;
- version, ownership and review status.
Reliability and validity are not achieved by calling a table a rubric. A visually neat instrument may still assess the wrong construct, reward writing fluency more than technical reasoning or disadvantage learners through inaccessible language or format.
Do not collect learner names, images, accounts or identifiable work unless the activity justifies them and the applicable procedures are established. This page does not define those procedures.
Assessment can remain useful without a score. Observation notes, learner explanations, annotated comparisons, peer questions and revised claims may support formative decisions. Whether they contribute to a formal grade is a local educational decision requiring its own rules.
Review lesson-plan boundaries
Review teacher-resource boundaries
Decide what evidence matters before capture begins
Begin with the learning purpose. Decide what learners should question, observe, compare, explain or revise. Then define the smallest set of evidence that would make that reasoning visible.
Choose a verified lesson when one fits. Select supporting resources whose purpose and version are clear. Prepare the classroom workflow so learner roles preserve evidence rather than concentrating all action around one device.
Assessment attached at the end often rewards the easiest artifact to see: the final model. Assessment planned before capture can reveal the richer learning sequence behind it.
Prepare the classroom workflowTeacher resources are the materials around an activity that help learners make decisions, record evidence and explain what they observe.