Fast Gates: The Question Before Repair
Wrong answers are outcomes. Fast Gates help diagnose causes.
Sometimes a wrong answer looks more informative than it really is.
A pupil writes that a hockey puck would eventually stop moving if all forces were removed. The answer is wrong, but the wrongness is not the most important part. The important question is what kind of wrongness it is. Does the pupil think force is needed to keep motion going? Have they confused balanced forces with no forces? Are they repeating a half-remembered phrase about resultant force without a model of motion underneath?
That is the diagnostic problem teachers face all the time. A wrong answer appears on a mini-whiteboard, in an exam response, during questioning or in a half-formed sentence from the back of the room. It is tempting to treat the answer itself as the problem: they do not know it, they need it explained again, they need more practice, they need a scaffold.
Sometimes that is true. But often the answer is only the final visible point of failure. Underneath it, very different things may be happening. One pupil may not yet have the idea at all. Another may have the idea but lose precision when applying it. Another may know the words but not the relationship that makes those words useful. This is the problem at the centre of Concept Repair, the framework I am developing for diagnosing learning breakdowns before choosing how to respond.
The visible answer is not the diagnosis. It is the outcome.
That is why we need Fast Gates. A Fast Gate is a short diagnostic question, or pair of questions, used before the teacher decides how to respond. It is not a full assessment. It is not proof of mastery. It is not a teaching strategy by itself. It is a decision mechanism.
Hinge questions, in Dylan Wiliam’s sense, often help a teacher decide whether the class is ready to move on. A Fast Gate sits close to that family, but has a narrower purpose: it helps the teacher decide what kind of help is needed next.
Does the idea need rebuilding, or can the teacher build on what is already there?
Teachers already collect evidence constantly: mini-whiteboards, verbal answers, written responses, circulation, retrieval questions, exam answers, exit tickets and hinge questions. The problem is not that teachers lack evidence. The problem is that evidence does not interpret itself. A Fast Gate inserts a small diagnostic pause between error and response: is the model visible, can the pupil use it, or are they only holding fragments, phrases or surface features?
A good Fast Gate makes the model work
A good Fast Gate does not simply ask whether the pupil can repeat information. It asks whether the underlying model can do something. A pupil can remember a definition without understanding the structure beneath it. They can recognise a phrase, reproduce a sentence from the board, or give a correct answer in a familiar format while still lacking a usable model underneath.
That is why a Fast Gate has to be more than a keyword check.
Take diffusion. A weak check might ask:
What is diffusion?
A pupil may answer:
Movement from high to low concentration.
That answer may be correct, but it may not show enough. Does the pupil understand random particle movement? Do they understand net movement? Do they know that diffusion is not just any movement of particles? Do they understand what happens when concentrations become equal?
A stronger Fast Gate might ask:
Particles have moved from an area with fewer particles to an area with more particles. Is this diffusion? Explain your answer.
Now the pupil has to reject a non-example. They have to show that diffusion is net movement from higher to lower concentration. The definition has to do work. That is the difference between a recall check and a Fast Gate. A recall check asks whether the pupil can remember the words. A Fast Gate asks whether the words are attached to a working model.
What makes a bad Fast Gate?
A bad Fast Gate lets pupils appear secure without making the model visible. It asks only for a definition. It checks a keyword but not the relationship underneath it. It copies the original question too closely. It gives away the structure inside the question. Or it tests a broad topic rather than one tight mental model.
For example, if the issue is synaptic transmission, asking What is a synapse? might check vocabulary. It does not necessarily show whether the pupil understands the causal sequence.
A better Fast Gate would ask two small questions:
What travels across the synapse, and by what process?
Then:
If the receptor on the next neurone is blocked, what happens to the impulse?
The first question checks the essential mechanism: neurotransmitters diffuse across the synapse. The second question makes the model predict. If the receptor is blocked, the neurotransmitter cannot bind, so a new electrical impulse is not triggered. That is much more diagnostic than asking for a definition. It reveals whether the pupil understands the sequence, not just the label.
The simplest Fast Gate structure
In its simplest form, a Fast Gate often uses two questions.
The first question checks whether the essential structure is present. Can the pupil name, describe or identify the key relationship without which the model cannot operate?
The second question makes the model do something. Can the pupil predict, explain, reject a non-example or reason through a small variation?
The simplest Fast Gate structure is therefore:
Can the pupil name or describe the essential structure?
Can the pupil use that structure to predict, explain or reason in a changed situation?
If pupils cannot answer the first question, the essential structure is probably not yet available. If they can answer the first question but fail the second, they may have the parts without the relationship. If they can answer both, the teacher can probably move towards refinement: precision, independence, application, written expression or exam performance.
The details of designing that second question matter, and I will return to them in a later post. For now, the key point is simpler: a Fast Gate is not a longer assessment. It is a small diagnostic pause before repair.
Keep it small
Fast Gates are meant to be quick. They should not become mini-assessments, elaborate worksheets or tasks that require ten minutes of marking before the teacher can act. A good Fast Gate can often be answered on a mini-whiteboard in under a minute. The point is not to collect lots of data. The point is to create enough evidence to improve the next decision. The best Fast Gates are brief but not shallow, fair but not leading, familiar enough to be accessible and varied enough to make the model operate.
The answer should change what the teacher does next. If it does not change the repair, it is not yet doing enough diagnostic work. What frustrates me is not that pupils make mistakes. That is part of learning. What frustrates me is how easily we can give pupils the wrong kind of help and mistake our responsiveness for precision.
Fast Gates do not remove uncertainty from teaching. Classrooms are too complex for that. Pupils can guess. They can hesitate. They can give partial answers. Teachers can misread what they see.
But teaching does not need perfect certainty before acting. It needs a better next decision.
A Fast Gate creates a small diagnostic pause between the error and the response. It asks whether the underlying model is visible, and whether it can still operate when the surface conditions shift. This is the wider promise of Concept Repair: not simply better answers, but better decisions about what to do next.
Wrong answers are outcomes. Fast Gates help diagnose causes.