A curriculum demonstration ends with a row of finished projects. The photographs look impressive. The students appear engaged. Then a school leader asks a more revealing question: What can the students explain that they could not explain before?
Student work can help answer that question, provided we examine more than the finished product. For STEM curriculum evaluation, the useful evidence includes the decisions students made, the observations they used, the support they received and what they can do when a familiar task changes. Those details matter to school teams choosing instructional materials and publishers seeking an independent content review.
Start with the learning claim, then read the work
Suppose a unit promises that students will use measurements to explain how materials affect heat transfer. A colorful temperature graph establishes that someone recorded and displayed numbers. By itself, it does not establish that the student understands the comparison.
Before reading the samples, write down what would count as evidence: the student compares relevant measurements, explains the result using the intended science and recognizes conditions that make the comparison fair. This makes the review more precise than judging whether a response sounds sophisticated.
Assessment research treats learning as something inferred from performance, rather than directly visible on a page. A peer-reviewed primer on science assessment explains why the target concept, the task and the interpretation of the response need to fit together. Our practical application is to examine the assignment alongside its answers. If the prompt never asks students to explain a result, their silence tells us little about their capacity to explain it.
Three answers to the same STEM task
Illustrative example: The task, measurements and responses below are invented to demonstrate a review method. They are not classroom findings, actual student records or STEM.org award criteria.
Imagine an upper-elementary unit comparing two identical covered cups of warm water. The cups begin with equal water volumes at the same temperature, stand in the same room and use the same measurement procedure. One cup has an insulating wrap. After ten minutes, its water has cooled by 8°C; the unwrapped cup has cooled by 14°C. Students must recommend a cup arrangement for keeping water warm and justify their choice.
Sample A: A choice without an explanation
“Use the wrapped cup. We followed the steps and it worked best.”
The student identifies the intended choice. The response does not show which measurements matter or why the wrap might help. A reviewer should inspect whether the materials provide opportunities to develop that explanation, rather than assume that choosing correctly demonstrates understanding.
Sample B: Evidence attached to the choice
“Use the wrapped cup because its water cooled by 8 degrees, compared with 14 degrees in the other cup.”
This response connects the recommendation to the observed difference. It supports a narrower conclusion: the student can use the supplied measurements to justify a choice. We still need to hear what the student thinks caused that difference.
Sample C: Explanation with a boundary
“The wrap slowed heat transfer from the warmer water to the cooler room. That helps explain the smaller temperature drop. Both cups needed the same starting conditions. I would repeat the test before deciding how dependable the difference is.”
Here the student connects evidence, mechanism, a condition of the comparison and a reason for further testing. That gives a reviewer more to examine. It still does not prove independent understanding if the explanation was supplied by a teacher, teammate or digital tool.
These examples are not a ladder for ranking children. They show how different responses support different conclusions. The curriculum question is whether the materials help teachers notice those differences and decide what to teach next.
Keep the support visible
Now add a detail: Sample C was written after a teacher dictated the first two sentences. Sample B was produced independently. The interpretation changes immediately.
Support is part of teaching. Worked examples, discussion, vocabulary aids and accessible response formats can make demanding ideas approachable. During curriculum assessment, record which support was available and whether it supplied the reasoning being examined.
A sentence starter such as “The measurements suggest…” leaves the evidence choice to the learner. A completed statement naming the mechanism supplies a crucial part of the answer. Both may have an instructional purpose; they answer different assessment questions.
Ask publishers to show the student prompt, teacher guidance and sample response together. Look for later opportunities to explain without content-specific hints, while retaining necessary accommodations. An annotated drawing or oral explanation may reveal scientific reasoning that a polished paragraph conceals. The task should assess the intended science without unnecessarily making writing fluency the deciding factor.
Change the situation and listen again
The next question is whether students can use the idea beyond the demonstration. Research synthesized in How People Learn shows that using knowledge in a new context cannot be assumed from success in the original one. Learning conditions and the depth of understanding matter.
Offer a modest variation: the same cups now contain water colder than the room. Which arrangement should slow the water’s warming? Ask students to predict, explain and describe a fair comparison before testing.
A student who reasons that insulation slows heat transfer in either direction is using the underlying idea. A student who says “the wrap makes things hot” reveals an interpretation that the original cooling task might never uncover. Neither response should be reduced to a label about the student’s ability.
For curriculum buyers, inspect whether later tasks provide these opportunities and whether teacher notes help interpret them. Keep the variation within knowledge students have been taught. One altered task offers evidence about this particular application; it is not proof that students can transfer their learning to every unfamiliar problem.
Request samples that can inform an adoption decision
In a curriculum adoption meeting, ask for a small, interpretable evidence set rather than a gallery of exceptional work. Request anonymized examples showing different responses to the same task, the original instructions, scoring guidance, available support and any subsequent revision. Establish whether examples are real classroom work, publisher-created exemplars or composites.
Ask how the samples were selected. A collection chosen to demonstrate excellent work cannot tell you how commonly that performance occurs. Without an earlier comparable response, a final sample also cannot establish how much a learner improved during the unit.
Use the evidence to make a concrete decision. Perhaps the materials prompt strong explanations but need clearer teacher guidance. Perhaps a pilot should test whether the expected reasoning appears under your school’s normal conditions. Record that finding, its limit and the next action.
This focused examination complements our broader guide to evaluating STEM curriculum before adoption. It also reflects STEM.org’s emphasis on human agency: students should have opportunities to make and explain judgments that belong to them.
Bring the right evidence to a curriculum review
STEM.org Reviewed™ is the pathway for eligible published learning content, including curriculum, books and educational media. Its current scope addresses STEM alignment, educational value, accuracy and clarity, usability, and support for the publisher’s claims. Recognition applies to the evaluated title or content scope.
For publishers, a useful preparation step is to assemble the relevant edition, intended audience, learning claims, student tasks and teacher guidance. Where authorized samples are available, explain their origin and conditions. Follow the documentation requested for your submission; this article’s illustrative method is not a substitute for the applicable review requirements.
For buyers, verify the credential and awarded scope, then examine the evidence needed for your own implementation decision. A content review and a classroom effectiveness study answer different questions. Neither a seal nor a striking sample establishes that every learner will achieve the same outcome.
Planning an independent STEM curriculum review? Explore STEM.org Reviewed™ and its submission pathway. Start with the learning you claim, then make the work supporting that claim available for inspection.
