The third time a student fails the same skill check, the room changes. She knows the number. I know the number. The rubric on the clipboard has not moved since the first attempt, and neither of us is able to pretend this is routine anymore.
When I built the dental assisting program at Beckfield College, the remediation structure I started from was the one nearly every clinical program uses: a counting policy. Three attempts at a competency, then the student does not pass the course. It is clean, it is defensible, it is administrable, and a program needs exactly that kind of record. What a count cannot do is teach. It tells you a student missed the standard three times. It is silent on why, and the why is the entire job.
A failed skill check is a diagnosis, not a verdict
A rubric captures an outcome. It cannot tell you which link in the chain broke, and the chain is longer than most of us admit when we are annotating a checklist and thinking about the next student waiting outside.
Consider a student who cannot pass an instrument processing competency. She might not know the sequence, which is a knowledge problem and the easiest kind to solve. She might know the sequence perfectly and still fumble the transfer, which is a motor problem that needs repetition rather than another explanation. She might perform it cleanly at the bench on Tuesday and come apart on Thursday with me standing behind her holding a clipboard, which is not a skills problem at all but performance anxiety wearing a skills costume. Or she might do every step correctly and take twice as long as the standard allows, which is a fluency problem, and the one most often mistaken for incompetence.
Four very different students. One identical mark in the gradebook. If your response to all four is "practice more," you will be right about one of them.
The fourth attempt has to be a different conversation
The instinct after a third failure is to schedule a fourth attempt. That is not remediation, that is repetition with higher stakes, and for an anxious student it makes things measurably worse.
What I learned to do instead was stop testing and go back to teaching, with the specific step isolated. Not the whole sequence. The step. I would ask the student to narrate while she worked, because narration exposes where the mental model is thin in a way that silent performance never does. I would move the practice to a time when the lab was empty and I was not visibly evaluating. And when I suspected anxiety rather than skill, I would let her run the sequence with someone else in the room and no paperwork in my hands, then compare that to what happened under observation. The gap between those two performances is the actual finding.
Most of what stops an adult student is not academic
The population I taught was career changers, first-generation students, and working adults holding a job and dependents together while retraining. Several had a bad history with school and arrived expecting to fail. For those students, a third failure is rarely a story about instrument processing.
More than once, the real answer surfaced only when I stopped asking about the competency and asked what her week had looked like. A car that stopped working. A childcare arrangement that collapsed. A shift schedule that changed without notice, so she had been practicing at midnight. None of that belongs in a rubric, and all of it determines whether the fourth attempt is worth scheduling this week or next.
There is a line, and I did have to hold it
Everything above can slide into an unwillingness to ever fail anyone, and that is its own failure. Infection control is not a competency a student passes eventually. Sterilization protects a patient who will never know her name and never get to evaluate her work. When someone could not reach the standard, my obligation ran to that patient.
The distinction I eventually learned to make was between "not yet" and "not this." Not yet means the trajectory is right and the timeline is wrong, and the honest response is another term, a different sequence, more supported practice. Not this means the trajectory is flat across every modality I have tried, and continuing costs her tuition and time she cannot afford. Telling a student that plainly is the hardest conversation in the job. Telling her late, after two more terms of debt, is worse.
What I would tell a new instructor
The counting policy exists for the institution. It gives you a record, and you need one. But the teaching happens outside it, in the part nobody documents, where you work out which of the four students in front of you you are actually looking at.
And it is worth saying that the student who fails three times and passes on the fourth is frequently the graduate I would hire first. She has been forced to understand exactly how she learns something difficult, which most students never have to figure out. She knows what she does when she is behind. In a clinical setting, that is not a consolation prize.