Why ‘grammar schools for all’ won’t work

A better, but overlong, title for this would be “Why grammar schools don’t work for all and why ‘grammar schools for all’ (probably) won’t work”.

At the birth of the comprehensive school movement, prime minster Harold Wilson made his well-known rallying cry, “Grammar schools for all’! Every child, no matter their background, or academic potential could go to a school which would share the values of the selective Grammar schools. It was a lovely idea and, as we all know, it failed to materialise. The reality, for very many children, became secondary moderns for all. Of course Wilson was well-intentioned; of course he can’t be blamed for the slide into child-centred ideology; of course he had no way of foreseeing the nightmare that was to come. But he was an optimist, and an unscrupulous one at that.

In particular, it would be unfair to blame Wilson for failing to foresee the formulation of group socialisation theory. Judith Rich Harris was an academic outsider who was able to say the unsayable and point out that parenting has practically no effect on how children turn out. For a long time we’ve known that pretty much all human characteristics are roughly 50% heritable. That is to say, biologically determined by our parents’ genes. That leaves another 50% to be accounted for by the environment and, until Harris, the environment was synonymous with the nurture provided by our parents. Her contribution to developmental psychology was to observe that culture is primarily transmitted between peers and that the missing 50% could, almost in its entirety, be attributed to what she dubbed group socialisation. I won’t go into the elegant trail of evidence she offered up here, but I would point you to her fantastically well-researched and very readable book, The Nurture Assumption.

What’s all this got to do with grammar schools? Well, group socialisation theory predicts that the most important variable for determining children’s educational success is the peer culture at their school. In a selective school, parents go to some trouble to make sure their children pass a demanding entrance exam and, although some people don’t want to admit it, grammar schools serve largely homogenous populations of parents with similar socioeconomic status. Parents’ values are handed down to children whilst they’re at home and continue to hold sway as long as these are values shared by a critical mass on the child’s peer group. If a small minority of students from different socioeconomic backgrounds attend the school, they will take on the values shared by their peers and abandon those of their parents. They’ll start to speak differently – at least whilst at school – and, because the school is academically selective, they’re likely to take on beliefs about the value of hard work and be increasingly motivated by academic achievement.

But, if the minority group is large enough – how large Harris admits she doesn’t know but speculates it could be as few as 3 or 4 – then children will identify with those most like themselves and be socialised in opposition to the dominant group. Differences between groups tend to exaggerate over time as they become increasingly distinct. Harris reckons that “no circle is more vicious than the one having to do with intelligence.” (247)  The heritability of intelligence actually increase over time from about 50% to as much as 80%; the cleverer you are, the better the choices you’re likely to make, the more likely you are to get cleverer still. Group contrast effects can have an indirect but profound effect on the heritability of IQ. If the group values hard work and good behaviour, individuals within the group will learn more; if the group thinks school is for geeks and trying hard is for losers, individuals within the group will learn less. “What starts as a different attitude to schoolwork might well end up as a difference in average IQ.” (248)

The stereotypes espoused by a group can have a lasting influence on group members. If the group values hard work then it becomes important to identify as a hard worker. If our group values mucking about and being defiant, then that’s how we’re likely to identify. When we find ourselves in situations where we’re torn between two sets of values, we experience what Claude Steele dubbed ‘stereotype threat‘. Steele found that all you had to do to make African-American students perform worse on a test was to give them a pre-test questionnaire which included a question about race. Simply being reminded of our group affiliations is enough to trigger the associated stereotypes about who we’re supposed to be.

Grammar schools work for most of their students because gaining status within their peer groups is about being academically successful. This might also explain why children from lower socioeconomic status tend to do worse in grammars; if there are sufficient students from a similar background they’re likely to band together around their own shared values and see themselves as distinct to the majority. Harris explains why it might not work to send a group of students from a shared background to an academically selective school: “They form a group of their own and retain the attitudes and behaviors they brought with them to the school.” (260)

I want to make absolutely clear at this point that this blog is in no way an argument for academic selection. Academic selection might well work for the majority of student who attend grammar schools, but such students would – and do – form like-minded groups with shared values within comprehensive schools. This isn’t really the problem though. The real issue is that the non-selective schools where the children who’ve failed the selection test end up will exaggerate heritable differences in IQ downwards. As Graham Nuthall pointed out, “Ability is the consequence not the cause of what happens in the classroom.” Whenever there is academic selection, children categorise themselves as clever or not-so-clever. Groupness makes us like those in our groups best and instead of feeling low self-esteem at being in the not-so-clever group, children gain self-esteem through gaining status within their group. If the group thinks school is for losers, then they’ll feel better about themselves if they muck about and don’t try. This is the Matthew Effect: “For unto every one that hath shall be given, and he shall have abundance: but from him that hath not shall be taken away even that which he hath.”

If we were to initiate a well-intentioned programme of ‘grammar schools for all’ staffed by high achieving teachers teaching a culturally rich academic curriculum our initiative might well fail for the same reason that sending large groups to selective schools might fail; because the idea of groupness hasn’t been addressed. This is not fate, but it is the prediction made by group socialisation theory. If you want schools or classes to be successful you need to address the peer culture. Teachers and school leaders have some real power in this regard. Leaders can do much to change the characteristics of the group and, thankfully, leaders do not have to be group members to be successful leader.

Harris points out three ways in which teachers can shape peer culture (245):

  1. By defining group norms. You don’t have to influence every member of a group, you just need to nudge enough of the most influential members. This then generates a kind of ‘herd immunity’ against poor choices and bad behaviour.
  2. By defining the boundaries of the group. We can, to a greater or lesser extent, control who is in and who is out, who is us and who is them. Of course it’s possible for subcultures to form within classes and schools but by engendering strong social norms we can make belonging to the in-group both inclusive and desirable.
  3. By defining the image or the stereotype students have of themselves. If we encourage students to value hard work and disciplined behaviour in each other, then we will have done our job; they will police the social norms themselves.

This is, I think, how successful schools in disadvantaged areas operate. They create an in-group where ‘we’ are different to everyone out there. ‘We’ feel privileged to be in the in-group and appalled at the idea of what it must be like to be a member of the out-group. ‘We’ notice everything that makes us different from ‘them’ and we revel in the differences.

The potentially fatal flaw in the idea of ‘grammar schools for all’ is that there has to be an out-group for this to work. If we look to the way strong positive social norms operate in the classrooms of many east Asian countries then we might find a model that could work across the system, but then again, maybe it would rely too heavily on cultural difference we just wouldn’t be able to replicate. I’m not sure if there’s a way around this problem but to ignore it is to be an unscrupulous optimist. Problems don’t go away just because we don’t think about them.

27 Responses to Why ‘grammar schools for all’ won’t work

  1. Let me check I’ve understood correctly. You’re arguing that with a heterogeneous intake, a school can only create a strong culture of achievement and discipline by defining themselves against the ‘losers’ out there?

    • David Didau says:

      No, I’m saying that’s one of the predictions of group socialism socialisation theory. It seems to fit the facts. But, as I suggest in final paragraph, East Asian schools might offer a more hopeful direction.

      • East Asian schools are situated in a much more traditional culture, where authority and the wisdom of the centuries are widely esteemed. Unless there is a radical reversal of cultural trends in England, traditional schools are going to continue being counter-cultural, as the culture at large does not respect authority or the past.

        The unscrupulous optimism of politicians such as Harold Wilson lay in thinking they could promote traditional education, while dismantling the culture on which it was based through their social policies.

        I often wonder whether those who support the teaching of knowledge and strict discipline realise just how radically they are swimming against the tide, culturally speaking. This is the reason they provoke such outrage.

        • David Didau says:

          Yes, I know, which is precisely why I hedged as to whether they could offer a solution to education systems in the west. I do think though that every school can work hard to promote an in-group with academic values and see themselves as being in competition with other excellent out-groups. This is, perhaps, how house systems work (when they work.)

    • Alison Honeybone says:

      That is exactly what the entirely admirable Katherine Birbalsingh can be heard doing at 1.08 in this video – unless I have completely misunderstood (entirely possible!) Her school is fantastic and her methods are very impressive – but here she does seem to be keen to enhance her students’ sense of themselves as part of something special and different.

  2. […] ‘Why grammar schools for all won’t work’ […]

  3. An Timire says:

    The problem is that adolescent culture has been dumbed down significantly in recent times, as discussed in the ‘Dumbest Generation’ by Mark Bauerlein. Rates of reading, theatre-going, museum attendance are all in decline among the young, and children hardly value intellectual pursuits at all, which is in contrast to previous generations who were more likely to be ashamed at their own ignorance. With the rise of the digital age, the influence of peer culture has been magnified and that of older generations diminished, which is a worrying trend when youth culture is so anti-intellectual.

    • David Didau says:

      Yes, that is a worrying trend. Schools are not powerless in the face of this though.

    • mc says:

      Many of today’s youth watch “Adventure Time”. Anyone who enjoys that show cannot be non-intellectual. I dare you to find youth fiction of any sort with deeper themes.

      And then there is the highly popular Minecraft, which while used for many things, is often used for artistic and programming purposes.

      There are of course all sorts of new ways intellectual pursuits can be pursued.

      Reading is of course often not an intellectual act, and theatre-going is generally worse.

      While I won’t say today’s youth are any more or less intellectual than in the past, passive use of specific media isn’t a good metric to gauge a person’s intellectualism by.

      • An Timire says:

        The fact that reading and theatre-going rates are in decline is a symptom of a broader withdrawal from the adult world among adolescents: Mark Bauerlein cites various studies and surveys indicating that young people are less involved in civic life and less politically aware than previous generations. While they may play video games that appear to be mentally stimulating, these games produce cognitive benefits that only serve to improve their game-playing skills but that do not lead to the habits of concentration and knowledge acquisition that are necessary to become a well-read and informed citizen. Thus while it is true that IQ has risen in recent decades as a result of a more technologically advanced society, levels of cultural knowledge and civic activity have decreased among the young because of their absorption in the digital world. Young people rarely use the internet to access cultural content: 9 out of the 10 most popular websites frequented by them are social media sites. Bauerlein’s point is that reading is fundamental for democracy. The knowledgeable the citizenry in the history and politics of their country, the more likely they are to engage in political and civic life and the better able they are to hold their government accountable.

  4. […] last post was written to explain why I thought ‘grammar schools for all’ was probably an […]

  5. Interesting, and probably correct. But faintly depressing nonetheless, although the suggestions for how teachers/schools can influence group culture are sound, I think. I remember reading somewhere that human character was 25% heredity, 25% parenting, 25% peer group and 25% who-the-hell-knows…or did I imagine that?

    Harold Wilson was undoubtedly unscrupulous at times…but aren’t all politicians? Indeed, aren’t we all…at times? And based upon my parents recollections, the comprehensive model, flawed as it was/is, is better that what went before?

    • David Didau says:

      You may well have heard that “human character was 25% heredity, 25% parenting, 25% peer group and 25% who-the-hell-knows” but that if that ever was a consensus view it certainly isn’t now. The 50/50 stat is common currency and obviously varies from trait to trait – some traits are more heritable than others, but even height is only considered to be 60-80% heritable. Harris’s finding that the 50% that wasn’t heritable comes almost completely from peer influences was a bombshell which is still being argued but has now become the consensus view.

      • It certainly is a bombshell for me! And for some reason, I find it rather depressing. I suppose it is to be expected, given that humans are social animals. I just wonder about how “peer group” can be defined. For example, would the parents and immediate family of home-schooled children count as their peer group? That’s probably an extreme case, but I have observed that some families seem to have more of an influence on their offspring than others. Are those statistical outliers on the consensus model?

        • David Didau says:

          No, peer groups are, as you’d expect, close in age. The mistake underpinning that attribution of environmental effects to parenting assumes that the goal of a child is to become a successful adult and therefore to imitate adult behaviour. In fact, the goal of children is to be successful children. If a home schooled child was prevented from forming a peer group – if they weren’t allowed any friends – then parental influence would dominate up to the point that the child was free to leave the family home.

  6. Grumpywearymathsteacher says:

    Don’t children often choose as their friends those who have similar values and interests, which they have generally got from their families? I don’t think the two things are independent.

  7. Pique Boo says:

    The JRH pecking order is genes, then peers, then parents. The difficulty is distinguishing between the influence of genes and the influence of parenting. Given e.g. a irksome competitive child, was it just their irksomely competitive parent who trained a child to be like that, or did the child end up with the same bunch of genes that influence a competitive streak?

    This is complicated. It’s been a while since I read that quite enjoyable book, but I took it less as peers affecting their personality full stop, and more as apparent personality being context sensitive. Some children do seem to be consistent everywhere, but for many I think it’s a case of wearing their at-school hat when they’re at school.

    My Y9 daughter is a perfect example. At-home girl is a fidgety chatterbox who is quite good at fillibusters and talking all over the climax of your favourite TV series. In-class girl is very quiet, somewhat enigmatic hands-down girl and this is routinely raised by school-side because 21st model child is an extrovert and there’s little else to put in the ‘to improve’ box. I’m not sure whether it was covered by JRH but I think neutrality is another way of not being the ‘proud nail that gets hammered down’ i.e. a way to fit in enough without having to adopt too much of the peer culture. Roll your skirt up just one turn to signal ‘one of us’, diss the common out-group (teachers) but for the most part keep your head down.

    With-close-school-friends girl behaves differently, and so does at-Scouts girl (the respective peer groups don’t overlap). So which one is my daughter? I reckon it’s the personality somewhere between at-home and with-close friends which have quite a lot of overlap. It’s what she does when she’s with people she knows well and trusts.

    I’m not sure how far school-side culture will stick outside school. It would be a bad mistake to assume they’re passive and not aware of adult agendas and manipulation. Daughter might play along with “We can beat Eton” etc. but she’s not daft enough to believe it. The difficulty with inter-school academic competion is what form that takes. We have lots of inter-school sport competition and the reality is that few care unless they’re on the team, and telling them they can beat anyone will wear-off when they don’t.

    • David Didau says:

      I think I was pretty clear that the equation is (roughly) 50% genes, 50% peers. Parenting obviously influences peer culture but the finding from adoption studies is that family background has 0 correlation with children’s eventual personalities and traits: if siblings share no genes then their parenting makes no difference to how they turn out. If one sibling is a natural child and shares 50% of parents’ DNA then they are similar to parents whilst the adoptive sibling is not at all similar. Therefore we can be pretty sure that heritability must account for the similarities.

      Harris is excellent on showing just how sticky peer culture is when she delves into the outcomes of immigrants’ children and deaf children born to hear parents. If children join a peer group that doesn’t share parents’ culture, the peer group’s values always win out. You’re right that parents’ values hold sway in the family home but the rather obvious point is that children grow up and leave the family home.

      I’m not sure how far peer culture as manipulated by a school will last either; it depends on the extent children buy into it. But, there’s excellent reasons to think from studying the research into in-group/out-group socialistation that it could be effective.

      • The implication you’re making is that parents can’t do anything to harm their children.

        • David Didau says:

          I’ve not implied that as it’s obviously wrong. What I’ve stated it is that the effects of parenting on adult characteristics are virtually nil. But, if a parent were to kill a child then that would clearly harm them. There’s also less extreme extremities: severe neglect or abuse before the age of 5 seems to cause permanent neurological harm. On a more sensible note, parents can do much to children’s childhoods nicer and can provide opportunities and choices which otherwise would not be available. But the point remains that parenting does not have lasting, direct effects on personality.

          • I think the key thing you said there was ‘lasting’ – which gives me hope for children I come across who seem to certainly be experiencing some personality modification due to difficult parental circumstances. However, how do we know? How can we ever say how that individual’s personality would have turned-out at the age of 45 if they hadn’t had that upbringing? My own father died suddenly when I was 6 – we found him dead – and that certainly soured things for a while – it didn’t seem great parenting on his part to me, and I was bitter, but I’m not sure that my personality is in anyway affected now. But again, how would I, or anyone else, know…?

          • David Didau says:

            How do we know? Decades 0f twin and adoption studies. As with all these types of research, we can make fairly robust empirical claims at the level of populations which will, of course, not hold true at the level of individuals.

          • Thank you David – that satisfies me. I’ve also realised that your link to The Nurture Assumption is a full pdf, so I’m going to make use of that instead of quibbling here anyway!

          • David Didau says:

            Sounds like a plan 🙂

  8. […] “This is, I think, how successful schools in disadvantaged areas operate. They create an in-group where ‘we’ are different to everyone out there. ‘We’ feel privileged to be in the in-group and appalled at the idea of what it must be like to be a member of the out-group. ‘We’ notice everything that makes us different from ‘them’ and we revel in the differences.” (From The Learning Spy) […]

Constructive feedback is always appreciated

%d bloggers like this: