Muslims, South Asians and the British mainstream: A national identity crisis?

Abstract

Current popular opinion assumes that Muslims and South Asians in Britain develop anti-mainstream identities because they live in impoverished and segregated ghettos, participate in non-mainstream religions, and politically organise via ethnically and religiously motivated networks. This article uses survey data from the 2003 Home Office Citizenship Survey to challenge each of those points. First, it shows that Muslims and South Asians are almost as likely as whites to identify themselves as British. Second, it argues that discrimination is more important than simple socio-economic difficulties for British identification. In addition, it claims that despite living in ethnically segregated neighbourhoods and retaining ethnic and religious social and political networks, Muslims and South Asians have actively built integrated networks, have trust in mainstream political institutions, and are committed to being a part of the larger British community.

Acknowledgements

This article was written with the support of the Families and Social Capital ESRC Research Group at London South Bank University. The author would like to thank Michel Laguerre and two anonymous reviewers for their comments on earlier versions of the article.

Notes

1. Prior to the late 1980s ‘black’ was a common term used to unite the various non-white ethnic minorities in Britain, i.e. Caribbeans, Africans, Indians, Bangladeshis, and Pakistanis.

2. The analytical categories ‘Muslim’ and ‘South Asian’ are sometimes used interchangeably in Britain, but should be considered as distinct – if overlapping – terms. The 2001 UK Census shows that 74% of Muslims are South Asian (Indian, Bangladeshis or Pakistani), meaning more than 25% are non-South Asian. The 2001 UK Census also shows that 50% of South Asians are Muslim, meaning exactly one-half of South Asians are not Muslim. The groups most likely to be Muslim are Bangladeshis and Pakistanis, each at 92%.

3. The BBC documentary series Panorama aired a special ‘A Question of Leadership’ on 21 August 2005, which questioned whether Muslim leadership in Britain (most notably the Muslim Council of Britain) was sufficiently focused on solving the serious problems of poverty and social exclusion, as well as sufficiently strict with radical elements of British Islam, which supposedly led to the 7 and 21 July 2005 bombings.

4. Together Bangladeshis and Pakistanis comprise 59% of Britain's Muslim population (2001 UK Census).

5. The 2000/01 Households Below Average Income and Family Resources Survey shows that Pakistanis and Bangladeshis were the ethnic group most likely to live in low income households at approximately 60% of individuals, compared to under 30% for both Indians and Caribbeans. (See http://www.statistics.gov.uk/cci/nugget.asp?id=269).

6. The ethnic minority boost sample was obtained by two methods: 1) focused enumeration: two addresses to the left and the right of each core sample address were screened for the presence of ethnic minority adults, and 2) high concentration boost: an additional sample of adults drawn from postcode sectors in which 18% or more of households were headed by a member of an ethnic minority. The combination of methods ensures enough ethnic minorities and a diverse population that does not entirely come from ethnic minority-dense neighbourhoods. For the focused enumeration sample, of the 2,945 in-scope addresses, 62% were interviewed, 14% were not contacted and 24% refused to participate. For the high concentration boost sample, of the 5,315 in-scope addresses, 52% were interviewed, 21% were not contacted and 27% refused to participate. In comparison, for the 14,758 in-scope addresses of the core sample, 64% were interviewed, 11% were not contacted and 24% refused to participate. While the ethnic minority high concentration boost sample has the lowest response rate, the refusal rates are similar across all three samples, suggesting that ethnic minorities were just as likely to respond favourably to the interviewer as whites. This therefore minimises the importance of self-selection effects where respondents from one ethnic group would be more likely to be alienated and less likely to participate. The big difference in response rates is between the 21% of the high concentration ethnic minority boost sample that was not contacted and the 11% of the core sample that was not contacted. This difference is probably due to ethnic minorities being more likely to work in several jobs and during unconventional hours, and the suggested sample weights were used to minimise these and other possible sample biases.

7. The Citizenship Survey includes 16 questions on perceptions of discrimination by various actors: surgeons, local schools, housing departments, local councils, private landlords, armed forces, insurance companies, bank/building societies, job centres, courts, crown prosecution, home office, police, immigration authorities, prison service and probation service. While ethnic groups had slightly different responses according to the different actors, only one of these variables was included because of the high correlation among responses to these questions. Including more than one measure of perceptions of discrimination would introduce multicollinearity, leading to unreliable significance values, confidence intervals and directions of the coefficients. The variable for courts was included because it measures a national institution but is not overly determined to produce a negative response, as the more intrusive crown prosecution, home office, police, immigration authorities, prison service or probation service might have been. However, the results are consistent across all discrimination variables. Another option to these questions on perceived hypothetical discrimination was a series of questions on perceived discrimination in real life cases of being fired from employment or treated poorly at work. While this latter series of questions has the advantage of being closer to actual events, the sample size dwindles dramatically because it only includes people who have been fired from employment, treated poorly at work, etc. In addition the second set of questions remains based on perceptions, and are therefore not as useful as the first set, which retain a larger sample size.

8. An example of a variable in ascending order with equal intervals would be 1 –$10, 2 –$20, 3 –$30, 4 –$40.

9. The 2001 Census shows that 29% of Caribbean men and 20% of Caribbean women are married to a member of another ethnicity compared to less than 10% of Indian, Pakistani and Bangladeshi men and women. Data on ethnicity of partners are not yet available for the 2001 Census, but the 1991 Census revealed that 27% of Caribbean men had white partners, compared to 7% of Indians, 5% of Pakistanis and 3% of Bangladeshis while 15% of Caribbean women had white partners, compared to 4% of Indians, 1% of Pakistanis and 0% of Bangladeshis (Census, April 2001, Office for National Statistics; Berrington 1996).

10. A 1999 Commission for Racial Equality Factsheet presents data from numerous studies which all claim that South Asians are more likely than Caribbeans to suffer from racial harassment and violent attacks (Commission for Racial Equality 1999). See also Nanton (1999) for an in-depth treatment of how Caribbeans are considered more assimilated and closer to the British mainstream than South Asians.

11. Regression results for Whites, Caribbeans, and Africans can be found in the Appendix.

12. The relationship between education and national identification is discussed in various strands of literature. Brooks (2000) writes about the internationalised culture of the highly educated while Pasture and Verberckmoes (1998) write about the importance of national-level identity for working class people.

13. It is not surprising that perceived discrimination is not significant for Indians because they are less likely to be victims of discrimination than Bangladeshis, Pakistanis, or Muslims in general, and the discrimination they do face has not hindered their significant socio-economic achievements in comparison to Bangladeshis, Pakistanis, and Muslims in general (Modood and Berthoud 1997). It is less clear why discrimination is not significant for Bangladeshis. One possible explanation is that Bangladeshis have the smallest sample size of all groups and although discrimination has a P-value of 0.059 in Model B it might become significant at P < 0.05 with a larger sample. Another possible explanation is that Bangladeshis are more recent migrants than Pakistanis or Indians and are more likely to be focused on the initial stages of adjustment instead of discrimination.

14. Trust in both Parliament and the local council would be significant for Pakistanis at p < .10.

15. In the Citizenship Survey 86% of Muslims born abroad have ‘fair’ or ‘very’ strong attachment to Britain compared to 80% of Muslims born in Britain, for Indians it is 88% to 80%, Bangladeshis 89% to 77% and Pakistanis 88% to 75%.