Different ways to show what you know

We care too much about what students know to confuse assessment with sameness. Assessment should produce useful evidence of understanding and capability. That does not mean every meaningful question has one appropriate response format, and uniformity does not magically make a measurement fair.

A written analysis may be essential when the goal is disciplined academic writing. In another context, a model, oral defense, laboratory investigation, annotated design, or performance may reveal learning more clearly and more quickly. The method should serve the academic purpose rather than the convenience of grading.

Before setting any task, we insist on a deceptively simple question: what would count as evidence that a student understands this? Once that is answered, the appropriate format usually becomes obvious—and it is often not the format that habit would have selected.

Every assessment is an inference. A student's performance on a task is the observable thing; understanding is what we infer from it. That inference is only as good as the alignment between the task and the construct it claims to measure, and it is threatened from two directions: the task may miss part of what matters, or it may capture something irrelevant alongside it.

Timed written exams, for example, measure recall, organization under pressure, and writing fluency at the same time. For some purposes that combination is exactly right. For others it means a student's grasp of chemistry is partly a report on their handwriting speed. Neither observation argues for abolishing exams; both argue for knowing what a given instrument is actually reading.

This is also the honest case against over-interpreting any single score. A mark is a compressed summary of one performance on one day under one set of conditions. It is information. It is not a description of a person's capability, and a school that treats it as one will make predictable mistakes about its students.

Offering varied ways to practice and demonstrate learning is not the same as lowering expectations. Students can be asked to explain their decisions, use evidence, revise their work, respond to critique, and transfer ideas into unfamiliar contexts in every medium. An oral defense in which a student must answer unanticipated questions is, by most measures, a harder test of understanding than an essay written over a week.

Variation also helps teachers locate a barrier accurately. A student may understand a concept thoroughly yet struggle with the format used to express it. Separating those two questions leads to better feedback: one student needs more instruction in the discipline, another needs instruction in the form, and giving the second student more content review is wasted effort for everyone.

There is a design discipline that keeps variety from becoming incoherence. The criteria should stay constant even when the format changes. If clarity of reasoning, quality of evidence, and responsiveness to counterargument are what the course values, they should be assessable in a paper, a presentation, and a build alike.

Assessment changes learning mainly through what happens after it. Research on formative practice consistently finds that the value lies in feedback that tells a student what to do next, delivered while the work is still live enough to act on. A grade alone rarely does this; a grade accompanied by comments the student never uses does not do it either.

The practical implications are modest but demanding. Feedback needs to arrive quickly. It should name a small number of specific, actionable moves rather than cataloguing everything imperfect. And the course structure has to include an occasion to apply it—a revision, a second attempt, a related task—or the advice becomes commentary on a closed episode.

We defend revision passionately because it is too often mistaken for leniency. In every field where work matters, it is drafted, reviewed, and improved. A school that grades only first attempts is teaching students that the version they produce under initial conditions is the version that counts, which is true almost nowhere outside school.

We refuse to let one task carry the entire story of a student's progress. A balanced picture comes from work accumulated over time: discussion, practice, projects, formal assessments, reflection, and revision. Each source has characteristic blind spots, and a combination covers more of them than any one instrument can.

A portfolio view has a further advantage for students and families: it makes trajectory visible. Two students with the same end-of-term grade may have arrived very differently, and the difference matters for deciding what comes next. Where a record shows only outcomes, that information is lost.

When students understand what quality looks like and receive more than one opportunity to make their thinking visible, assessment becomes part of learning rather than only a verdict at the end.

Grades are the most familiar signal parents receive and among the least informative on their own. When you review a school's assessment practice, ask what proportion of a course grade rests on a single sitting, whether students ever revise graded work, and how quickly feedback is returned. Ask what a teacher would be able to tell you about your child beyond a number.

If your child has historically performed differently across formats—strong in discussion and weaker on timed tests, or the reverse—say so early. That pattern is diagnostic information, not an excuse, and it changes how a teacher should read early results.

We do not publish outcome statistics, rankings, or claims about where our students place, because we will not advertise figures we cannot substantiate. What we can do is discuss a specific student's record in detail. Admissions can arrange that conversation.