Designing and Using Assessments
A practical lesson on aligning learning goals, assessment tasks, evidence, and decisions.
What Is and What It Is For
What Is and What It Is For
is the deliberate collection and interpretation of evidence to make a judgment or choose a next step. A score alone is not the whole process: the evidence should help someone understand learning and decide what, if anything, to do.
The links three elements. Cognition is the learning expected; observation is what students do or produce; and interpretation is how that evidence is understood. These must fit together. If the goal is to understand students’ reasoning, a task should reveal their reasoning, and the interpretation should address what that evidence can show.
The distinction between formative and summative is about the purpose and use of evidence. is ongoing: evidence is interpreted to adjust instruction, learning strategies, or feedback. summarizes learning at a particular point, often to report or certify achievement. A quick classroom question is not automatically formative; it becomes formative when its results guide what happens next. One task may serve either purpose—or both—depending on how its results are used.
For example, after teaching fractions, a teacher might ask students to explain why is greater than . If several explanations compare only the denominators, the teacher can address that misconception with a model and a follow-up task. Used this way, the responses are formative evidence. A broader at the end of the unit could ask students to compare, represent, and solve fraction problems, summarizing achievement against unit goals. The difference is not simply the question asked; it is what the evidence is used to do.
Design from the Learning Goal
Design from the Learning Goal
Design an by starting with the intended learning, rather than choosing a convenient task first. Specify what learners should know or be able to do, including the kind of thinking or performance expected. Then ask what observable evidence would support a reasonable conclusion about that learning.
Next, choose or create tasks that elicit that evidence, establish criteria for judging responses, and plan how the results will inform feedback or decisions. If the task, scoring, or interpretation does not fit the goal, revise the design. This keeps the expected learning, what students show, and what is inferred from it aligned.
An can make those connections visible by mapping goals to evidence and tasks. For a goal such as “explain how evidence supports a scientific claim,” a written explanation and an oral discussion could provide evidence of reasoning. A vocabulary quiz may show whether students know relevant terms, but by itself it does not establish whether they can use evidence to support a claim. The task and criteria should therefore match the intended reasoning, not merely a related piece of knowledge.
Evaluate Evidence and Decide What to Do Next
Evaluate Evidence and Decide What to Do Next
results are estimates based on a sample of what a student knows and can do. No single response or score captures all learning, so the strength of a conclusion should match the evidence and the ’s coverage.
Several checks help determine whether evidence is suitable for its intended use:
Alignment: Do the tasks represent the learning goals and the level of thinking expected? leaves important learning unassessed. Construct-irrelevant difficulty—such as unnecessarily complex wording—can obscure the skill the task is meant to assess.
: Do the evidence and reasoning support this particular interpretation and use? is not a permanent label attached to a test. A result that helps guide tomorrow’s lesson may not, on its own, justify a high-stakes placement decision. Consider task content, how students respond, scoring, and likely consequences for the proposed use.
: Are results consistent across occasions, versions, or scorers? Clear directions, defined scoring criteria, representative tasks, and scorer calibration can improve consistency. But a consistent result is not necessarily valid: a vocabulary test can reliably measure vocabulary and still be inadequate evidence of scientific reasoning.
Fairness: Do students have a fair opportunity to demonstrate the target learning? Review for cultural or linguistic assumptions unrelated to the goal, accessible administration, and appropriate accommodations. Consider whether a score reflects the intended skill or an unrelated barrier.
Practicality: Are the time, materials, and scoring effort manageable, and will results be available when they are useful? An elaborate may be less useful for a classroom decision if results arrive too late.
Gather, interpret, and act
Choose evidence-gathering methods to fit the goal. A performance goal may call for a demonstration; conceptual understanding may be revealed through explanation, comparison, or justification. Evidence can also come from targeted questions, exit tickets, quizzes, problem-solving tasks, projects, observations, portfolios, and self- or peer . Combining evidence from different tasks or occasions can give a fuller picture than relying on one result.
Interpret evidence against the learning goal and criteria, not only by comparing students with classmates. Look for what is secure, which errors or misconceptions recur, and what evidence is missing or ambiguous. For formative purposes, these patterns can guide whether to reteach, explain differently, provide targeted practice, extend learning, or gather more evidence. Feedback should identify what is working and a feasible next step. For summative decisions, report what the supports and avoid conclusions broader than its coverage allows.