Mastery vs seat-time.
The Carnegie unit — the unit that runs the entire American high-school clock — came out of 1906 as a way to standardize faculty workload for pension eligibility. Whether or not a student had actually learned anything was not what it was designed to measure. That assumption is still running the school day. It does not run our cohort.
Every American teacher inherits one accidental decision from 1906. That year, the Carnegie Foundation needed a way to standardize how much teaching counted as a full-time faculty load — the unit they ultimately used to decide which professors were eligible for a pension under the foundation’s new retirement fund (the one that later became TIAA). What they came up with is what we now call the Carnegie unit: 120 hours of class contact time, roughly one course over one school year.1
That metric was never intended to measure learning. It measured seat-time: was the kid in the room, was the teacher in the room, was the clock running. A century later, almost the entire structure of American high-school transcripts and college admissions is still built on it. A class is a class is a class, and the clock is what makes it official.
“A pension formula from 1906 is currently running the biology classroom.”
What the research actually says
The case against fixed-time, “everyone advances at the same rate” instruction is not a hot take. It’s a fairly well-trodden corner of education research:
- Bloom’s mastery learning work. In the 1960s and 70s, Benjamin Bloom and his students documented something striking: when students were allowed to keep working on a unit until they actually demonstrated mastery (typically ~80% on a formative assessment), then advance, average outcomes improved by roughly one standard deviation, and the spread between weak and strong students narrowed. Bloom’s famous “two sigma problem” paper extended this to one-on-one tutoring with mastery, which produced about two standard deviations of improvement — the largest single intervention in education research.2
- Spaced and deliberate practice. Cognitive-science work from Anders Ericsson and others on deliberate practice, and from the “testing effect” literature, all point in the same direction: time-on-task is not the same as mastery-on-task, and the difference is large. The student who is shoved past a skill they haven’t yet built is not saving time. They are accumulating debt.3
- The Carnegie Foundation itself has tried to walk it back. The same foundation that created the unit has, in the last two decades, repeatedly published papers acknowledging that seat-time is a poor proxy for learning and urging movement toward competency-based credit. The K–12 system has, mostly, not moved.4
The science is fairly clear. The institutional inertia is fairly stronger.
Average student outcome (percentile rank) under three instructional regimes. Adapted from Bloom (1984).
A cohort of eight is the practical compromise — closer to the +2σ end of this chart than a 1-to-30 classroom can ever be.
What this looks like on a Saturday
A Saturday cohort of eight is small enough to actually run on mastery instead of seat-time. That sounds dramatic. In practice it’s a hundred small choices that all push the same way:
1. The skill, not the schedule, is the gate.
When the syllabus says “week 1: focus the microscope and identify four tissue types,” that’s the skill. If the student walks in week two and still can’t crisply focus on a stained slide and tell me which one is epithelial, we don’t skip ahead to keep up with the calendar. We re-do the bench checkpoint. The clock waits.
2. Re-doing isn’t a punishment, it’s the work.
One of the quiet psychological costs of seat-time school is that doing something twice feels like failure. In a mastery framing, doing something twice is just — doing it twice. That’s how musicians, athletes, and surgeons learn. Repetition with feedback is the entire technology.
3. The cohort moves; the individuals stretch.
Here’s the honest constraint. Eight students share a calendar, and we have to keep the cohort roughly together so that the dissection day still has a class around it. Pure self-paced mastery is not what we offer; we’re not a tutoring service. What we offer is a cohort calendar with built-in slack — specific checkpoints where individual students can be held back to redo a skill, or pulled forward to a deeper version — without the rest of the schedule collapsing.
We don’t pretend to abolish the schedule. We just don’t pretend the schedule is teaching anyone anything by itself.
Why we can do this and most schools can’t
I want to be honest about scale. Mastery-based teaching is dramatically harder at thirty students than at eight, and dramatically harder at eight hundred than at thirty. With a full classroom, formative re-checks, individualized feedback loops, and individual stretch problems are operationally brutal. Most public-school teachers I know would love to teach this way and do not have the staffing or the schedule to do so.
A small Saturday cohort is one of the few formats where you can. Eight students is a number where the instructor can actually see who has the skill and who doesn’t on a given day, and adjust. It is the format the research calls for, in roughly the size the research calls for. That’s not by accident. It’s the design.
What the kid actually walks away with
The visible artifact is the bound notebook and the capstone defense. The invisible one is harder to point at but more valuable: the muscle memory of having earned competence rather than been promoted past it. Once a student has had the experience, even once, of not being moved on until they could actually do the thing — and then doing it — the standard quietly recalibrates for everything else.
Students who’ve been through that loop tend to ask, in the next class they take, the only question that matters: do I actually know this, or am I just on schedule? That habit is more durable than any single unit of biology. It’s the habit that keeps a college student from quietly drowning in week six of organic chemistry, and a working professional from coasting on what was true ten years ago.
We move on when the work is right, not when the bell rings. That’s the entire pedagogy in one sentence. The bell is welcome to keep ringing on its own.
Sources & further reading
- The Carnegie unit was defined in 1906 by the Carnegie Foundation for the Advancement of Teaching as an administrative measure tied to a faculty pension fund (the same fund that became TIAA), not to student learning. See Carnegie Unit and Student Hour for the documented origin and the foundation’s subsequent century of administrative use of the unit.
- Bloom, B. S. (1984). “The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring.” Educational Researcher, 13(6), 4–16. doi:10.3102/0013189X013006004. The classic mastery-learning effect-size paper; tutored mastery groups outperformed conventional instruction by roughly two standard deviations. Background: Bloom’s 2 sigma problem.
- On retrieval practice / the “testing effect”: Roediger, H. L. & Karpicke, J. D. (2006). “Test-enhanced learning: Taking memory tests improves long-term retention.” Psychological Science, 17(3), 249–255. doi:10.1111/j.1467-9280.2006.01693.x. On deliberate practice as a distinct skill-acquisition mechanism: Ericsson, K. A. & Pool, R. (2016). Peak: Secrets from the New Science of Expertise.
- Carnegie Foundation for the Advancement of Teaching (2015), The Carnegie Unit: A Century-Old Standard in a Changing Education Landscape — the foundation’s own re-evaluation of the seat-time unit it originally designed. Coverage and summary: Paul Fain, “Carnegie Says Credit Hour, Although Flawed, Too Important to Discard,” Inside Higher Ed, Jan. 29, 2015.