What Schooling Is For

A child in a rich country spends something like fifteen thousand hours in school between five and sixteen — more waking time than with any adult except a parent, and in many cases more than that.

Ask what all those hours are for and you get four answers, each held seriously, each with evidence, and each implying a different account of what a society is doing when it educates its children.

They are not mutually exclusive, and the most common error in this area is treating them as though they were.

One Tuesday, timed.

Follow a thirteen-year-old through a school day, and record what actually happens rather than what the timetable says.

. She arrives, and waits. 08
— registration: she is counted, and being late has consequences. 09
— assembly: three hundred children sit in rows, in silence, and are addressed by an adult about behaviour and about a forthcoming inspection.

to 10
— mathematics. She is taught for perhaps thirty-five minutes of it. The rest is settling, distributing materials, an interruption, dealing with two students, and packing away.

— a bell. Movement is by corridor, in one direction, with staff at intervals. 10
— break, which is unstructured and is where most of the day's social learning happens.

to 12
— two more subjects, unrelated to each other and to the first, switching on a bell whether or not she has finished thinking about anything.

— lunch. She queues. She sits with the same eight people she has sat with since September, and the composition of that group is one of the most consequential facts in her year.

to 15
— three more periods. In one, work is returned with a grade. She compares hers to her neighbour's before reading the comments.

— she is dismissed.

Now audit the day.

Time spent on subject content: perhaps three hours out of seven.

Time spent on everything else: queueing, waiting, being counted, moving in permitted directions at permitted times, being evaluated by strangers against a public standard, working alongside others while being forbidden to help them, and switching attention on a signal regardless of interest or completion.

Every one of those is a lesson, delivered daily for eleven years, and none of them is in the curriculum.

And here is the point of the story. Each of the four theories in this lesson looks at that audit and says: that is what school is for.

The functionalist sees a child learning to accept impersonal rules from strangers and to be judged by universal standards — the transition from family to society. The human capital theorist sees three hours of skill formation, which is what she is paid for later. The signalling theorist sees an eleven-year endurance test that reveals who she already was. And the reproduction theorist sees a workplace — punctuality, deference, extrinsic motivation, individual competition, and attention fragmented on command.

All four are describing the same Tuesday.

Why the manifest answer is not enough.

The official answer is that school teaches knowledge and skills. It does, and 9.1.2 shows how much.

But the official answer cannot explain several things about how schools are organised, and this is what the other theories exist for.

Why is the curriculum divided into forty-minute blocks that end regardless of where anyone has got to?

Why is helping another student sometimes cooperation and sometimes cheating, with the line falling exactly where assessment does?

Why do employers reward credentials whose content they never enquire about, in fields where the content is irrelevant to the job?

Why has the amount of schooling required for a given occupation risen continuously, without a corresponding change in what the occupation involves?

And why does the relative advantage of the well-off persist through every expansion of access (see 8.3.1)?

None of those is answered by "school teaches things". Each is answered by one of the theories below — which is the reason for taking all four seriously rather than picking one.

The four theories

One — the functionalist answer: school makes members.

Durkheim's version is the foundational one and it is more interesting than its reputation. The family, he argued, cannot produce a member of a society, because family attachment is particular — the child is loved for being this child. Society requires attachment to something impersonal.

The school is where that transition happens. It is the first institution in which a child is subject to rules that apply regardless of who they are, administered by an adult with no personal attachment to them, in a group where they are one among many.

So the discipline is not a means to the teaching. For Durkheim it is a large part of the teaching: the child learns to accept restraint that is not arbitrary and not personal, which is what living in a society requires (see 4.2.1).

Parsons developed the argument into the terms the classroom actually operates in, and the formulation is precise.

In the family, a child is evaluated on ascribed grounds (who they are), by particularistic standards (different for each child), on a diffuse basis (the whole person).

In the classroom they are evaluated on achieved grounds, by universalistic standards, on a specific basis — this piece of work, against the same criterion applied to everyone.

And the shift is the socialising function. The classroom is where a child first meets the principle that performance rather than identity determines reward — which is exactly what a labour market claims to run on.

Parsons added a second function: allocation. The school sorts children into the positions the occupational structure requires, and — crucially — it legitimates the sorting, because the criteria are public, the same for everyone, and applied over years. A child who fails has, on the face of it, been fairly beaten.

Where this account is strong: it explains the organisation of the day, the impersonality, the grading, and the enormous emphasis on rules that appear to have no educational content.

Where it is weak: it assumes the standards actually are universalistic and the sorting actually does track talent. Both are empirical claims, and Part 8 has already made trouble for them.

Two — human capital: school produces productive capacity.

The claim: education is an investment. Time and money go in; skills, knowledge and capabilities come out; and those raise productivity, which raises earnings.

This is the best-evidenced of the four, and it should be said plainly.

Estimates of the return to a year of schooling cluster in the region of six to ten per cent in earnings, across many countries and periods. And the estimates survive the obvious objection — that people who get more schooling would have earned more anyway (see 7.6.2) — because of natural experiments.

Compulsory schooling laws are the standard instrument (see 7.4.3): when a country raises its school leaving age, a cohort born a few months apart receives different amounts of education for reasons entirely unrelated to their ability or their families. Comparing them gives a causal estimate, and those estimates are generally similar to or somewhat above the ordinary ones.

So schooling does raise earnings, causally, and by a substantial amount. Any theory that denies this is contradicted by good evidence.

What the theory does not explain: why the credential matters more than the content in many hiring decisions; why requirements rise without job content changing; and why expansion has not reduced relative inequality of outcomes.

Three — signalling: school reveals rather than creates.

The alternative account. Employers cannot observe a person's productivity before hiring. Education is costly, and it is more costly for less able people — harder, longer, more likely to fail. So completing it is a credible signal of pre-existing characteristics, and employers pay for the signal rather than for anything the education did.

The strong version is deliberately provocative: the education could teach nothing at all and still be worth doing, both for the individual and for the employer, as long as it sorts.

The evidence usually offered is the sheepskin effect. If education produces skill continuously, earnings should rise smoothly with each year completed. If it signals, there should be a jump at the point where a certificate is awarded — the final year should be worth much more than the previous ones.

Discontinuities of that kind have been found repeatedly in earnings data, with completed credentials worth more than the years alone would predict.

And the evidence is genuinely contested, which this course must report. Studies using sharper designs — comparing students just above and just below the threshold for passing a school-leaving examination, which is a regression discontinuity (see 7.4.3) — have in some cases found little or no effect of the certificate itself, which is difficult for the strong signalling account.

The proportionate reading, and it is the one most economists now hold: both. Education builds capacity and sorts, in proportions that differ by level, field and country. The interesting question is not which, but how much of each, where — and that is answerable.

Four — reproduction: school prepares people for their place.

The correspondence argument holds that the social relations of schooling mirror the social relations of work, and that this is what schooling principally transmits.

Stated as a list, it is uncomfortably recognisable from the story. Hierarchical authority. Extrinsic motivation — work done for a grade rather than for itself. Fragmentation of the day into unrelated units. Competition between individuals for scarce rewards. And differentiation by level: schools serving different populations exhibit different regimes, with more autonomy and more discussion where students are expected to end up managing, and more rule-following and closer supervision where they are not.

The empirical claim underneath it is the interesting part. What schools reward, on this account, is only partly cognitive. Grades are substantially predicted by behavioural characteristics — perseverance, dependability, consistency, deference — and it is those characteristics, rather than measured cognitive ability alone, that predict later earnings.

And that claim has held up remarkably well from a completely different direction. Research on so-called non-cognitive skills has established that persistence, conscientiousness and self-regulation predict earnings and attainment substantially, independently of test scores.

The most striking single piece of evidence involves a school equivalency certificate. People who leave school early and later obtain a qualification by examination have, on average, cognitive test scores comparable to graduates — and labour market outcomes closer to those who left without anything. They have the knowledge and not the record of having stayed. Whatever completing school certifies, on this evidence, is not principally what is in the examination.

The criticisms of the correspondence account are serious.

It is functionalist in form — it identifies what an arrangement does for an economy and treats that as an explanation (see 5.1.1, 8.4.3), without a mechanism showing how schools come to serve it.

It is too deterministic. Students are portrayed as receiving a formation rather than as doing anything.

And the classic response to that came from inside the tradition. An ethnography of working-class boys in an English school found them actively rejecting the school's terms — mocking conformity, valuing manual work and physical toughness, treating academic effort as effeminate and pointless. They were not passively reproduced; they resisted.

And the resistance delivered them into exactly the jobs the school would have allocated them to anyway. By refusing the credential on the grounds that it was a con — which, for people in their position, it substantially was — they guaranteed their own subordination. That is a much better account than correspondence alone, because it has agency in it and still reaches the same destination.

And the Bourdieusian version supplies the mechanism the others lack (see 5.8.1): the school evaluates on dispositions — vocabulary, manner, ease, relation to authority, familiarity with what counts as knowledge — which it presupposes rather than teaches, and which are unevenly distributed on arrival. It then certifies the result as ability, converting an inherited advantage into an earned credential. The word for that is misrecognition, and it is why the sorting appears fair to almost everyone involved.

And a fifth, which is Weberian and explains something the other four cannot.

The credentialist account: education is not principally about skills or about signalling productivity. It is a device by which occupational groups close access (see 8.2.1 on social closure).

A profession that requires a credential controls entry to itself. The credential's content matters less than its function as a gate, and the level required rises over time not because the work changes but because more people obtain the previous level — so the gate is moved.

The prediction is credential inflation, and it is what has happened: jobs that required no qualification now require a school certificate; jobs that required a school certificate now require a degree; and the work has frequently not changed.

This is the account that best explains why educational expansion has not delivered what its advocates expected, and it is the same finding as 8.3.3's effectively maintained inequality, arriving from Weber rather than from stratification research.

Because "school is just X" is wrong for all five values of X, and the fifth question is the useful one.

What the evidence supports.

Schooling causally raises earnings — this is solid, and the natural experiments carry it.

It also sorts, and the sorting has value to employers independently of what was learned — the credential discontinuities and the equivalency-certificate finding both point here.

What it rewards is substantially behavioural, not only cognitive, and those behaviours are unevenly distributed on arrival.

And it certifies the result as merit, which is what makes the outcome acceptable to those who lose by it — the function Parsons named approvingly and the reproduction theorists named critically, describing the same thing.

Three questions to carry into the rest of this Topic.

When a school outcome is reported, ask what is being measured — knowledge, behaviour, or the disposition to persist in an institution.

When a credential requirement rises, ask what changed — the job, or the number of people holding the previous credential.

And when a system is described as meritocratic, ask what merit is operationalised as (see 7.2.2). If it turns out to be the dispositions that advantaged families supply, the system is not measuring merit; it is measuring origin, and calling it merit is the mechanism rather than a mistake.

One closing observation about the Tuesday. Nothing in that day was designed by anyone with any of these theories in mind. The bells, the corridors, the rows, the grading and the seating plan are administrative solutions to the problem of processing three hundred children with thirty adults. That they also constitute a formation in punctuality, deference, extrinsic motivation and individual competition is not a conspiracy. It is what happens when an institution's practical requirements have consequences nobody intended — which is 4.2.1's argument about social facts, and the reason a sociology of education is worth having.

Fifteen thousand hours, of which perhaps three in seven are subject content. The rest — queueing, being counted, moving on a bell, being evaluated by strangers against a public standard, working beside others while forbidden to help them — is delivered daily for eleven years and is not in the curriculum.

Four theories look at that audit and each says "that is the point".

Functionalism: the school is where a child moves from particularistic, ascribed, diffuse evaluation in the family to universalistic, achieved, specific evaluation in an institution — Durkheim's moral discipline and Parsons's socialisation, plus allocation which the school also legitimates, so that a child who fails appears to have been fairly beaten.

Human capital: education raises productive capacity, and this is the best-evidenced of the four. Returns of roughly six to ten per cent per year of schooling, confirmed causally by compulsory schooling natural experiments.

Signalling: education reveals rather than creates, being cheaper to complete for the more able. Credential discontinuities support it; sharper regression-discontinuity designs have sometimes found little diploma effect. The supported position is both, in proportions that vary.

Reproduction: the social relations of schooling mirror those of work — hierarchy, extrinsic motivation, fragmentation, individual competition, and differentiation by level. Its empirical core is that grades and earnings are substantially predicted by behavioural rather than cognitive characteristics, which converging research on non-cognitive skills supports — most strikingly in the finding that holders of an examination-based equivalency certificate have graduate-level test scores and dropout-level outcomes.

Its criticisms: functionalist in form, and too deterministic — answered from within by the ethnography of boys who actively rejected the school's terms and thereby delivered themselves into the jobs it would have allocated them. And the Bourdieusian version supplies the missing mechanism: the school evaluates dispositions it presupposes rather than teaches, then certifies the result as ability.

And the credentialist fifth account explains what the others cannot — rising requirements without changing job content, because the credential is a gate and the gate moves as more people pass the previous one.

Nothing in the Tuesday was designed with any of this in mind, which is the point: the formation is a by-product of administrative solutions, not a conspiracy.

Hidden curriculum — what is learned from the organisation of schooling rather than from its content.

Particularistic / universalistic standards — evaluation differing by who the person is; evaluation by criteria applied identically.

Ascribed / achieved; diffuse / specific — judged for who you are or for what you do; as a whole person or on a defined performance.

Allocation and legitimation — sorting into positions, and making the sorting appear fair.

Human capital — productive capacity created by education; measurable returns.

Signalling / screening — education as a credible indicator of pre-existing characteristics rather than a producer of them.

Sheepskin effect — an earnings discontinuity at credential completion, taken as evidence for signalling.

Correspondence principle — the mirroring of workplace social relations in the organisation of schooling.

Non-cognitive skills — persistence, conscientiousness and self-regulation, which predict earnings independently of test scores.

Counter-school culture — active rejection of the school's terms, which can deliver the outcome the school would have produced anyway.

Credentialism / credential inflation — education as a closure device, with requirements rising as more people obtain the previous level.

Misrecognition — an inherited advantage certified as an earned ability.

One — audit a day. Take a school timetable you can get hold of and estimate what proportion of the day is subject content. Then list what is being learned in the remainder.

Two — find the closure. Identify an occupation whose entry requirement has risen in your lifetime. Ask whether the work changed, and what the previous requirement was.

Three — test the signalling intuition. Ask three people what they actually use from their highest qualification. Then ask whether they would give up the certificate and keep the knowledge.

Four — locate the behavioural reward. Think of a time you were graded well or badly. Work out how much of the difference was knowledge and how much was submission on time, presentation, and doing it the expected way.

Five — apply the merit test. Write down how "merit" is operationalised in one selection process you know — a school, a university, an employer. Then ask which families are best placed to supply it.