Teaching to the test is only a problem when the test is bad.
I think we make this issue more mysterious than it needs to be. Assessment always changes behavior. Students look at what counts. Teachers look at what will be rewarded. Classes bend toward the thing that will be measured. That is not a bug in education. That is education doing what systems do.
The real question is not whether assessment creates washback. It does. The question is whether the washback is pointing everyone in the right direction.
Good washback is not something to fear.
It is what happens when assessment, instruction, and learning outcomes are pulling in the same direction.
The Test Is Already Teaching
Every test teaches students what the course really values.
If the final exam is a list of disconnected grammar items, students learn that accuracy in isolation is the thing that matters. If the assessment asks students to explain an idea, respond to a partner, revise after feedback, and use language for a purpose, students learn something else. They learn that language is not just something to know. It is something to do.
This is why I do not find the phrase "teaching to the test" automatically damning. Of course teachers teach toward the test. Students deserve to know what game they are playing. The ethical problem appears when the game is badly designed.
Bad Washback Is Usually Bad Design
When people complain about washback, they are often complaining about misalignment. A course claims to value communication, but the test rewards memorized forms. A syllabus says critical thinking matters, but the assessment rewards copying the teacher's answer. A language program says fluency matters, but the final grade comes from silent paper tests.
That kind of assessment produces bad classroom behavior because it gives teachers and students bad instructions. It tells them to spend time on the wrong things. Then we blame the washback, as if washback itself were the villain.
The villain is the mismatch.
The key question: if students spent a month preparing for this assessment, would they become better at the thing the course claims to teach?
Design the Washback You Want
I would rather start with the behavior I want and design backward.
If I want students to speak more clearly, the assessment should make clear speaking valuable. If I want students to revise, the assessment should include revision. If I want students to use feedback, the assessment should reward what they do after receiving feedback. If I want students to communicate under real pressure, the assessment should contain some version of that pressure.
Then, when teachers teach toward the assessment, they are also teaching toward the learning outcomes. That is the whole point.
Make It Formative
The best version of washback is formative. It does not just tell students where they rank. It tells them what to do next.
A formative assessment gives students usable information while there is still time to act on it. It helps teachers see where the class is actually struggling. It turns assessment from a final judgment into part of the learning process.
This is especially important in language education because language growth is not a single event. Students need cycles: attempt, feedback, revision, another attempt. A well-designed assessment can create those cycles. A poorly designed assessment kills them.
The Better Version of Teaching to the Test
Here is the version I want: build an assessment so aligned with the course goals that preparing for it is indistinguishable from meaningful learning.
- If the goal is speaking, preparation should involve speaking.
- If the goal is writing, preparation should involve drafting, feedback, and revision.
- If the goal is communication, preparation should involve audience, purpose, and response.
- If the goal is growth, preparation should involve reflection and another try.
In that kind of course, teaching to the test is not a compromise. It is a sign that the test is doing its job.
TL;DR
Washback is desirable when the assessment is formative and aligned with the intended learning outcomes. "Teaching to the test" becomes a problem only when the test itself is poorly designed. The answer is not to pretend assessment will not shape learning. The answer is to build assessments worth teaching toward.