American education still certifies learning with two instruments. One counts the hours a student spent in a seat. The other counts the courses they finished.
Neither one asks what the student can do.
The first, the Carnegie Unit, was devised in 1906 to solve an administrative problem. Colleges needed a common way to compare high school transcripts. Nearly 120 years later, it is still the backbone of the American transcript.
Damian Creamer, founder and CEO of Primavera Online School and StrongMind, has spent 25 years building schools within that measurement regime and has reached a blunt conclusion about it.
“Seat time does not measure learning. It measures attendance,” Creamer says. “Course completion does not measure learning either. It measures that a transaction closed. We kept both because they are cheap to audit and they make scheduling work, which are administrative virtues and not educational ones.”
What the Metric Actually Rewards
The trouble with counting hours is that they punish students at both ends of the distribution.
“Think about who this serves badly,” Creamer says. “The student who understands it in twenty minutes still owes you the other forty, so what they learn is that school is mostly waiting. The student who needs ninety minutes gets forty, then a C, then the next unit built on top of the thing they did not finish. Neither one was measured. They were both just clocked.”
Course completion compounds the problem.
“A completed course is a container that closed. It tells you the student was present for the sequence and produced enough acceptable work along the way. It does not tell you what survived. Two students finish the same course with the same grade and one of them has the material a year later and one does not, and the transcript is identical.”
The compounding is what concerns him most. Gaps only expand throughout the year.
“By the time a grade lands on a report card, the information is stale and the student has already moved on to something that depends on it,” he says. “We are performing autopsies and calling it assessment.”
Mastery Is a Trajectory, Not an Event
Creamer’s alternative is not simply swapping hours for tests. He is skeptical of that trade, because a test is also a single moment.
“A test tells you what a student could produce on a Tuesday morning under time pressure,” he says. “That is one data point about performance, and performance and learning are different things. Real mastery is durable. It survives a week, a unit, a change of context. You cannot see durability in a snapshot. You can only see it over time.”
What he wants measured is the trajectory. Whether a student’s command of a concept is holding, improving or quietly eroding, and whether it transfers when the problem is dressed differently.
“The question is not whether a student passed the quadratics test,” Creamer says. “It is whether they still have quadratics in March, and whether they can use them when nobody tells them that is what this problem is about. That is harder to track. It is also the only version of the question worth asking.”
Why Nobody Did This Sooner
Creamer is careful not to frame a century of seat time as institutional laziness. It was a rational answer to a hard constraint.
Learners vary in preparation, pace, motivation and cognitive readiness. With one teacher and thirty students, continuous mastery tracking was arithmetic that did not work.
“Holding thirty trajectories in your head, across five subjects, updating them weekly, and adjusting instruction for each is not something a person can do,” he says. “So the system did the only available thing. It standardized the input, held the clock fixed, and let the outcomes vary. Then it graded students on the variance.”
Making Mastery Visible at Scale
Continuous tracking was unaffordable in teacher attention. It is not unaffordable in software.
“Measurement is the reason the intelligence sits underneath the platform instead of beside it,” Creamer says of the work at StrongMind.
“You cannot bolt continuous mastery tracking onto something built to deliver content on a schedule. The tracking has to be what the thing is, or you end up with a dashboard reporting on the same stale information faster.”
A system observing a learner continuously can separate a concept genuinely acquired from one performed adequately last Thursday.
It can surface the gap in week two instead of week nine. It can let time flex while the standard holds, which is the inversion he is after.
“Flexible time, fixed standard,” he says. “That is the whole reform in four words. We currently do the reverse and then express surprise at uneven outcomes.”
He is equally clear that better measurement is not more measurement. “This is not an argument for testing children more often. It is an argument for measuring the right thing continuously and quietly, inside the work, instead of stopping everything four times a year to discover what we should have known in real time.”
The Accountability Objection
The most serious pushback is about auditability. Seat time, whatever its flaws, can be verified by a state. Mastery is a judgment, and judgments invite inflation.
Creamer takes the concern seriously and argues it points the other way.
“If you are worried about rigor, seat time should alarm you,” he says. “It certifies presence. A student can accumulate every required hour, pass with a C built on three years of unresolved gaps, graduate, and then get taken apart by a college algebra class. We audited that student’s attendance perfectly. We simply never checked the thing the diploma was supposed to represent.”
Mastery-based measurement, in his framing, is harder to fake precisely because it demands evidence rather than a timestamp.
“Demonstrating you can do something, repeatedly, across contexts, is a higher bar than occupying a chair. It costs more to verify. It is also verifying something.”
What Changes If the Metric Changes
Creamer’s larger point is that measurement quietly determines everything upstream of it.
“You get the system your metric asks for,” he says. “Measure hours and you will build something optimized for scheduling. Measure courses closed and you will build something optimized for throughput. Measure mastery and you are forced to build something optimized for learning, because there is nowhere left to hide. Every structural choice follows from the number you decided to care about.”
Which is why he treats a 1906 accounting convention as more than a historical curiosity.
“We are having a very sophisticated argument about artificial intelligence in education while still certifying the output with a stopwatch,” Creamer says. “Fix the stopwatch and a surprising number of the other arguments resolve themselves.”
