Public Education Decoded

A fourteen-year-old sits in a classroom. A bell rang to put her there. Another bell will release her in forty-seven minutes. Between those bells, she will sit in a room with twenty-eight other students who share her birth year but not her interests, abilities, or readiness for the material. The teacher will deliver the same content at the same pace to all of them, because the schedule permits no alternative. The student already understands today’s lesson—she grasped it two days ago—but she cannot move ahead, because advancement is time-based, not mastery-based. She will spend the period performing understanding she already possesses, practicing patience she should not need to practice, and learning the real lesson the system teaches most reliably: that compliance is what matters, and your own pace of understanding is irrelevant to the institution that claims to serve it. In the next room, a different student is lost, three lessons behind, unable to follow today’s instruction because it builds on concepts he never absorbed. He, too, will sit through the period. Both students are being processed. Neither is being educated in any meaningful sense. And the system, by its own metrics, is working fine.

The Six Functions

Every public school system in America claims that its primary purpose is learning. This is not a lie, but it is not the full truth, and the gap between the claim and the reality explains most of what confuses people about education.

Childcare is the function no one discusses at school board meetings and everyone relies on. Parents need somewhere for their children during working hours. The school day maps to the work day. Summer break, originally tied to agricultural schedules, persists because no one has solved the childcare problem it would create if eliminated. When schools close unexpectedly—as they did during the COVID-19 pandemic—the economic disruption reveals how central this function is. Schools are, first and most practically, the infrastructure that makes dual-income households possible.

Socialization is the second function: teaching children how to behave in groups, follow rules, navigate authority, manage time, and cooperate with peers. These are genuine and valuable skills. They are also skills that require compliance, and the system rewards compliance above almost everything else. A student who follows instructions, arrives on time, sits quietly, and completes assignments as directed will succeed in school regardless of whether deep learning is occurring. A student who is curious, creative, and intellectually restless but unwilling to conform to institutional rhythms will struggle.

Sorting is perhaps the most consequential function and the least acknowledged. Schools rank students through grades, test scores, tracks (honors, regular, remedial), and behavioral records. This ranking begins in elementary school and compounds through the system: early labels shape teacher expectations, expectations shape opportunities, and opportunities shape outcomes. By the time a student applies to college, twelve years of sorting have produced a ranked position that predicts life trajectory more reliably than any measure of actual competence. Employers, colleges, and graduate programs depend on this sorting. They do not need schools to produce learning. They need schools to produce rankings.

Credentialing is sorting’s formal output: the diploma, the GPA, the transcript. Bryan Caplan, the George Mason University economist whose book The Case Against Education argued that schooling is primarily about signaling rather than skill development, analyzed the “sheepskin effect”—the disproportionate earnings boost that comes from completing a degree versus dropping out with nearly identical coursework. If education were about learning, a student who completed three and a half years of college would earn nearly as much as a graduate. They do not. The credential, not the learning, is what the labor market rewards.

Civic formation—creating shared national identity, teaching democratic values, building a common cultural foundation—is the original justification for public education, dating back to Horace Mann’s campaigns in the 1840s. This function is real but diminished, and it creates ongoing political conflict over whose values, whose history, and whose cultural narrative the schools should transmit.

Learning—the actual acquisition of knowledge, skills, and the capacity for independent thought—also happens. But it happens despite the system’s structure, not because of it. In other words, learning is the stated mission and the residual output: what occurs in whatever space the other five functions leave behind.

The Factory That Persists

The structure of American schooling was designed in the late nineteenth and early twentieth centuries, during the height of industrialization. The organizational model is, quite literally, a factory. Students are grouped by age into cohorts (batch processing). They move through a standardized curriculum at a standardized pace (assembly line). Advancement is determined by time served rather than mastery achieved (seat-time requirements). Subjects are divided into isolated units taught by specialists (division of labor). The day is segmented into arbitrary periods announced by bells.

This model assumes that students are interchangeable units to be processed uniformly. It was designed for throughput efficiency, not learning effectiveness. And it conflicts with essentially everything cognitive science has discovered about how learning actually works.

Robert Bjork, the UCLA psychologist whose research on desirable difficulties has reshaped our understanding of memory and learning, has demonstrated that effective learning requires conditions the factory model systematically prevents. Learning is most durable when it is spaced over time rather than massed into single sessions. It is most transferable when topics are interleaved rather than blocked into isolated units. It is deepest when it involves productive struggle—encountering difficulty and working through it—rather than the kind of smooth, frictionless instruction that feels good but produces shallow retention.

Standard schooling does the opposite: it masses practice before tests (cramming), blocks subjects into isolated periods, minimizes productive struggle in favor of clear presentations, and replaces intrinsic motivation (genuine curiosity) with extrinsic motivation (grades, rewards, threats). The system knows better. The research is not obscure. But the factory model persists because it serves childcare beautifully, produces sorting efficiently, and generates credentials reliably. Learning is the one function it handles worst, and learning is the one function no one in the system is directly rewarded for improving.

The Incentive Landscape

Teachers are evaluated on classroom management, parent satisfaction, administrator approval, and increasingly on student test scores. They are not measured on whether students retain the material six months later, whether students can apply concepts in novel situations, or whether students develop genuine intellectual curiosity. The feedback loop between teaching and meaningful learning outcomes is too slow and too noisy to function as an incentive signal.

Administrators are rewarded for test scores, low incident reports, and budget management. They are not measured on the depth of learning in their buildings or on what their graduates can actually do five years after leaving. Their time horizon extends to the next evaluation cycle, the next budget year, the next board meeting. Long-term student development falls outside the window their incentives measure.

Districts optimize for enrollment (since funding follows students) and for test score rankings that affect property values and community reputation. They are not measured on the long-term outcomes of their graduates—whether they are literate, numerate, capable of critical thought, or prepared for the demands of adult life. No district tracks these outcomes in any systematic way.

Politicians cycle through education reform initiatives on electoral timescales. A governor who launches a literacy program needs visible results within a term or two. The actual effects of educational changes often take a decade or more to manifest. The incentive is to pursue initiatives that produce immediate, measurable, quotable improvements—which is to say, test score gains—regardless of whether those gains reflect genuine learning improvement.

In other words, no one in the entire chain from classroom to statehouse is directly rewarded for improving learning measured on meaningful timescales. Everyone is rewarded for compliance, metric optimization, and institutional stability. The system produces exactly what its incentives select for.

The Goodhart Trap

Education reform almost always means: raise test scores. When test scores become the target, Goodhart’s Law activates with predictable precision. Teaching narrows to tested material. Test preparation replaces broader instruction. Untested subjects—art, music, physical education, creative writing, the humanities—get cut or reduced. Schools focus resources on “bubble kids” near the proficiency threshold, where a small score improvement changes the school’s statistical profile, while neglecting students far above or below the threshold.

The test score may rise. The underlying capacity it was supposed to measure may not. A student drilled in reading comprehension test strategies may score higher without becoming a stronger reader. A student trained to solve standardized math problems may perform well on the test without understanding the mathematics. The measure improves while the thing being measured stagnates.

Daniel Koretz, the Harvard education researcher whose book The Testing Charade documented the phenomenon across decades, calls this “score inflation”—rising test scores that do not correspond to rising achievement. The cheating scandals in Atlanta, El Paso, and elsewhere are the extreme version, but the ordinary, legal forms of score inflation are more corrosive precisely because they are invisible. The system reports improvement. The improvement is an artifact of optimization against the metric.

Who the System Serves

The current system advantages wealthy families, who can supplement school with tutoring, test preparation, extracurriculars, and enrichment. It advantages compliant students, who are rewarded for following rules regardless of intellectual engagement. It advantages students with strong test-taking aptitude, whose scores may overrepresent their actual capabilities. It advantages students from stable home environments, because school assumes parental support, homework supervision, and a quiet place to study.

It disadvantages poor families, whose children attend under-resourced schools and receive no supplemental support. It disadvantages non-conforming students—the creative, the energetic, the differently wired—who experience the compliance system as oppressive rather than supportive. It disadvantages late developers, whose early sorting locks them into trajectories before their abilities have fully emerged. And it disadvantages non-traditional learners, for whom the one-size-fits-all approach is the wrong size entirely.

Public education is often described as the great equalizer. The structural reality is closer to the opposite: it amplifies existing advantages and compounds existing disadvantages. The sorting function ensures that the children who enter with the most resources leave with the best credentials, and the children who enter with the fewest resources leave with rankings that confirm their position.

What Follows From This

This is not a call for abolition. Public education serves real functions—childcare, socialization, civic formation—that society needs. The problem is the gap between the claimed function (learning) and the actual optimization target (everything else). Closing that gap would require restructuring incentives so that the people who run the system are rewarded for learning outcomes measured meaningfully and over appropriate timescales. It would require abandoning the factory model in favor of approaches aligned with what cognitive science tells us about how people actually learn. It would require separating the sorting function from the learning function, so that ranking students does not become the system’s primary output.

These changes are structurally difficult because they threaten every incumbent interest: the childcare infrastructure, the credentialing economy, the testing industry, the administrative apparatus, and the political dynamics that treat education reform as a messaging opportunity rather than a design problem. The system persists not because no one knows better, but because every actor within it is rewarded for maintaining it.

For individuals, the practical implication is direct: do not confuse school with learning. Credentials are not competence. Grades are not understanding. The system sorts. Whether you actually learn is, in the end, up to you.

How This Was Decoded

Synthesized from Bryan Caplan’s signaling model of education, Robert Bjork’s research on desirable difficulties and effective learning, Daniel Koretz’s documentation of score inflation and test-based accountability failures, historical analysis of the factory model’s industrial origins, and multi-level incentive mapping across teachers, administrators, districts, and political actors. Cross-verified against NAEP longitudinal data, international comparative studies, and the consistent failure of test-score-based reform initiatives to produce durable learning gains. The mechanism is structural: a system that rewards compliance, sorting, and metric optimization will produce compliance, sorting, and metric optimization—regardless of what its mission statement says about learning.