This paper attempts to explore whether schools’ ethnic composition influences the formation of students’ friendships and whether these friendships are associated with students’ educational attainment. Friendship formation across and within ethnic groups is highly relevant to social capital theory which is currently widely used by the academic community. During the last decade it has also occupied a significant part in policy debates. Educational policies under the New Labour government in the UK have been significantly influenced by social capital theory, and particularly by Putnam’s approach (Gewirtz et al., 2005). Putnam (2007) - one of the main social capital theorists - has argued quite controversially on the basis of his area-level data in the US that ethnic diversity reduces social capital both across and within ethnic groups, even when controlling for income levels. How far would this apply in other national contexts, and what are its implications for education? Does this mean that students in multi-cultural schools form less intraethnic and interethnic friendships and trust each other less? If social capital is important for learning, and if schools are considered as communities, does this imply that diversity in school intakes is likely to be detrimental for educational outcomes? This paper investigates the extent to which Putnam’s latest suggestions and their theoretical extension are valid in another setting such as multi-cultural London.

This research is based on data gathered from Year 12 students, in four secondary schools in London. All four of them were co-educational state-schools and they were chosen according to the ethnic profile of their intakes. One had ethnically diverse and balanced intake, whereas the other three had predominantly White British, British Indian and British African-Caribbean intake, respectively. The reason behind this choice was the need for comparison of students’ friendships and attitudes in both cases of an ethnically mixed and an ethnically predominant environment. For all four cases mixed methods of collecting and analysing data were employed. A survey questionnaire was administered to approximately 300 students and in-depth, semi-structured interviews were carried out with about 60 students.

Preliminary analysis of the data suggests that the theoretical extension of Putnam’s model does not hold true in our sample. Whereas ethnic diversity in schools encourages interethnic friendships it does not have a negative impact on either intraethnic friendships or overall trust. Furthermore, students’ educational attainment does not seem to be negatively associated with friendships between students of certain ethnic backgrounds. Yet, it appears that the ethnic mix of the school is only one of the factors that influence friendship formation. The way students form friendships is a complicated procedure which does not only lie on their ethnic background but also on their social class background and their gender. Students form different types of friendships based on their multiple identities (social class, ethnicity, gender) and each type of friendship serves a different purpose.