Diversity As A Constraint On Social Capital Formation: A Study Of English School Children

Diversity as a constraint on social capital formation: a study of English school children

Irene Bruegel

London South Bank University

A. Introduction

While diversity is officially celebrated in Britain, an alternative discourse is gaining ground; that too much ethnic or cultural diversity is a threat to social cohesion. As Goodhart 2004 put it, ‘one of the central dilemmas of political life in developed societies: sharing and solidarity can conflict with diversity’. Goodhart’s worries underlie Blunkett’s ‘citizenship as integration’ agenda but they also carry with them a far darker agenda, of ‘tipping points’ beyond which a good society cannot be maintained. Much of the argument derives from pop evolutionary psychology, but it also draws on social research which suggests that local diversity hinders the development of social capital, sometimes measured as trusting neighbours and other times as participation in associations.

In this paper we argue that such research tends to treat both diversity and social capital as fixed entities, ignoring the processes by which recognition, trust and social solidarity are developed or thwarted. We draw on our on-going research on children’s friendship patterns in 12 Inner City and other primary schools in England.

This is not to deny the reality of racism, or the corrosive effects of racist attitudes, or still less to ignore the general tendency for people to choose friends from people they regard as ‘like themselves’ or to trust those who they identify as ‘looking like themselves’ (de Bruine 2005), but to put these into context.

The Home Office Citizenship Survey 2003 shows (Home Office 2004) that ethnic mixing is quite common in the UK especially amongst younger people and those who live in areas of high diversity. There is then both a cohort and a learning effect, and no inherent conflict between ‘diversity’ and social capital.

Our analysis of the British Household Panel Survey shows too that thinking yourself as ‘like your neighbours’ does not necessarily generate trust [Figure1]. Within poor neighbourhoods, 90% of those who thought their neighbours were very like themselves didn’t think they could be trusted, compared to about 78% of those who thought their neighbours were neither like, or unlike themselves. To that point at least, the more people in poor and mixed neighbourhoods identify their neighbours as like themselves, the less they trust them. Indeed this analysis shows that whether people saw their neighbourhoods as ‘poor’ or not, is key the trust they have of neighbours. This suggests that what is important to social capital and trust are people’s subjective evaluations of their neighbourhoods and their feelings about how they end up living where they do.

The disjuncture between sameness and trust is especially evident when gender diversity is considered. Gender is the most clear-cut cleavage in school friendship patterns with around 90% of children selecting friends of the same gender. And yet to our knowledge no one has argued that gender diversity – which carries over into voluntary activity and work association in adulthood - is any constraint on social cohesion, quite the opposite.

Over and above this we want to argue that whether a person ‘sees’ someone else as ‘like’ themselves, or even as ‘looking like’ themselves is not self-evident, but contextual (Ali 2003). Hence we aim to explore the way children (and their parents) conceive of what it is for someone to be ‘like themselves’, given the many dimensions of difference and similarity.

If 'beliefs, attitudes and values [are] shaped in formative years of primary school’ (Grugeon and Woods 1990), understanding how ethnicity is negotiated by young people would seem to be significant for understanding the ‘progressives dilemma’ that haunts Goodhart and others. Since children are often a catalyst to local friendships, including cross ethnic ones, children’s friendships can often have wider significance. On this basis alone there is reason to look at friendships amongst young people. But we also want to argue that the research in schools can overcome some of the conceptual and methodological problems of the developing research on diversity and social capital, without, however, solving them all.

In this paper we look briefly at criticism that can be made of existing studies under four themes: the meaning of given to diversity; how it is measured; the concept of proximity or neighbourhood; and the concept of social capital or trust employed. We then go on to look at patterns of inter-ethnic friendship amongst children aged 10- 11 to consider how diversity and segregation affect interaction, taking friendship between children to be a building block of local social capital.

B. Existing studies of local diversity and social capital

In their book ‘Better Together’ Putnam and Feldstein (2003), point to the ethnic homogeneity of Portland Oregon as one of the factors underlying the vibrant local democracy. In that study they are careful to point out that political ideas underlie the decentralised, participatory democratic structures and that class differences in outlook between different citizens persist. Putnam has gone on to see whether the link between ethnic composition, trust and participation he identified in Portland can be replicated across cities in the USA, but has not yet published the material. Goodhart (2004) nevertheless draws upon Putnam to make his argument, as well as the work of Alesina (2000) on Welfare States and some British evidence from Mori(2004).

Alesina’s work is said to provide evidence that welfare state provision as a whole is vulnerable to increased population diversity (Goodhart 2004), in that public good provision is better the more homogenous the local US city area. Alesina’s explanation of the link is that participation in city-wide associations is higher, the more homogenous the city, whether you measure diversity, by income, race, or ethnicity (2000).

Not only does that ignore systems of trust building beyond formal associations and the fact that prior attitudes affect involvement in such associations, detailed reading of Alesina’s research shows that heterogeneity only limits involvement in certain activities - churches rather than youth clubs or schools – and does not explain why these forms of association, but not others, are reflected in differences in public good provision. The ability of certain groups to mobilize power looks to be a more satisfactory explanation than ethnic diversity per se.

Causation flows in existing studies of diversity and social capital from local population diversity through social interaction, to social capital and generalised trust, generally without interrogating the diversity-social interaction and the social interaction-social capital links adequately. Neighbourhoods and ethnic categories are treated as fixed and reified, rather than explored as dynamic processes.

Our alternative sees ethnicity as variable, ethnic differences as subjectively experienced, depending on the forms of social solidarity historically established, not least through the schooling system, just as generalized trust can be shown to feed into, rather than result from, associational activity (Uslaner 2002; Stolle 2003).

1. Meaning of ‘Diversity’

Ethnicity is of course fluid. Not only are there a large number of dimensions of difference/ diversity, they can cross cut in important ways. Connolly (1998) describes how a group of naughty 6 year old boys- one black, two mixed and one white, were all described as black. An Asian boy who was good at football was seen as ‘becoming white’, while a white girl who played with Asian girls was shunned by other white girls ‘because she smelled’ . (p182).

Existing studies of children’s friendships show that preference for same ethnicity friendship is highly variable (Smith and Tomlinson 1989) and that the traditional ethnic hierarchy of preference, with whites at the apex, is far from universal. Ethnic borrowing of dialect, music and dance can lead to a downplaying of ethnicity, with white boys in particular being wannabe blacks more often than not (Peskin 1991 Back 2002).

Our sample of 600 year 6 children were surveyed about their own identity and that of their friends. As others have found, they express only a limited sense of their own ethnic identity (Ali 2003; Richardson 1971), a factor which complicates our more quantitative analysis. One Afro-Chinese boy in our sample for example gave us a self description as ‘Asian, reliable and black’.

Half our sample of school children were given a free choice of adjectives to select to describe themselves. These included ones relating to ethnic background, religion and personality. A high proportion of white children and quite a few ethnic minority children opted not to specify any national or ethnic identity. The other half were forced to select or explicitly reject an ethnic and religious identity and asked to classify their friends by ethnic identity. The fluidity and uncertainty about ethnicity was very evident. Not only did children fail to describe friends by ethnicity, where they did, the same child might be ascribed a very different ethnicity by different friends. Relatively few children identified themselves by religion, even when explicitly asked about their religion. Though this is revealing in itself, it rather undermined our intent to compare religion and national background in friendship preferences.

When asked to write down differences between themselves and their friends only a quarter of children whose friends were of a different ethnicity noted any relevant difference, for example that their friend ‘came from a different country’; or ‘had a different culture’. Some wrote their friend ‘looked different’, but this could refer to differences of height, weight, hair colour or skin tone, rather than ethnic background. On the other hand one or two children described children from same ethnic background- white European - as also culturally different, pointing out, for example, that ‘he’s Spanish’. So the salience of ethnic difference, still less faith difference, was not well established for this age group.

• 2. Measurement of ‘Diversity’

This is not the place to get into a discussion of the technical measure of diversity, just to point out that some measures, based purely on a white:black, or British: New Commonwealth distinction are highly problematic. We found that the link between the proportion of white children in an area or within a school and levels of diversity when measured across all possible groups is limited, especially where contrasts are made within London.

It is clear is that a very large number of adults and children in UK cities fall outside simple ethnic categorizations. In London 25% of households of two or more people contain people identifying themselves in different ethnic categories ( Census 2001GLA Commissioned Table M243). Not only are an increasing proportion ‘mixed’, large numbers fall into ‘other’ categories. How these groups are classified will effect any correlates of diversity; when Black Africans and Afro-Caribbeans are distinguished from one another, or Irish from British whites, the meaning of diversity changes, even if the technical measure can accommodate varying degrees of differentiation.

We found it impossible to attribute sameness or difference unambiguously to children classified as ‘mixed’ or ‘other’, reinforcing the problematic of ethnic diversity as often measured. In an ideal situation, it would be helpful to observe the friendship patterns of children from each type of mixed heritage backgrounds, but numbers do not permit this and it does not seem legitimate to treat all mixed-background children as if they necessarily shared an identity, so we sometimes excluded them from the analysis.

• 3. Concept of proximity/ processes of social capital formation.

Much of the literature on diversity and place lacks a feel for geography. Places are taken as givens, based on administrative boundaries or commuting patterns. Schools are points of social interaction in ways in which local neighbourhoods aren’t. So instead of inferring social contact between proximate neighbours, we know that there is some interaction between children in a school class from which friendships can develop.

The advantage of looking at schools is that we are dealing with a bounded social/ geographical group. This means that sorting processes which can pre-judge people’s attitudes to their neighbours are less relevant to primary school children.

Of course children in different circumstances will not always mean the same thing by describing a child as a friend; school friendships may be more friendships of convenience than others; they certainly aren’t always reciprocated. To allow for the possibility that school-friendships are qualitatively different from others, we divided the sample into two and allowed the second group of children to choose four friends both from within their class and outside it.

• 4. Varieties of Social Capital/ ‘diversity’ and the bonding: bridging distinction

What is most striking about the literature on diversity and social capital is the limited conception of social capital employed. Generally trust based on homogenous bonding is recognized to have potentially negative effects and possible racist overtones. It is bridging social capital that cuts across difference that is socially valuable. However the studies on which Goodhart constructs his case contrast social capital in heterogeneous areas with that in more homogenous places. There is a fundamental contradiction between the approach which identifies heterogeneous/ diverse areas as those where trust and participation and social capital are at their lowest and the finding- re-iterated in the Cantle report (Home Office 2001)- that distrust is generated by segregation and lack of interaction. The diversity and social capital discussion in the UK is in danger of a double whammy for minority groups: fuelling populist fears of ‘diversity’ while blaming minority groups alone for living parallel or separate lives.

Before being able to develop broad social solidaristic policy that allow for diversity and social justice, the stereotyping that takes diversity of demographic origins to equate to diversity of cultural practices and both to inhibit social interaction needs to be disentangled. By asking how experience at primary school might affect friendship patterns at secondary school in the context of systems of segregation of pupils between and within schools, our longer term research aim is to reflect on policy changes that might be needed to foster social solidarity across diversity.

Here we are only reporting on what happens at primary school level. We are not yet in a position to say whether having cross-ethnic friendships at primary school affects a child’s values and behaviour at secondary school. But we hope the analysis of primary school friendship as well as providing a baseline for the longer term study can tell us something about the way difference and diversity are constructed in primary schools today.