Ethnic group: identity and change 1991-2001

Funders: ESRC (as part of a Fellowship)
Name: Ludi Simpson
Dates: October 2003-September 2004

Outputs

• Quantifying stability and change in ethnic group Full report of evidence from the Longitudinal Study 1991-2001
• Journal of Official Statistics paper (2007)

• Stability and change in ethnic groups in England and Wales Population Trends 121, 2005
• Were researchers' expectations confirmed? 1 page
• Change in ethnic group Bola Akinwale MSc dissertation, City University, 2005

The results quantify the unreliability of ethnic group recording over time due to question change, amibiguity of the question for some respondents (including those not born in the country or region associated with the ethnic group label), errors, and conscious change of affiliation.

The analysis provides recommendations for analysts of individual microdata and aggregate data. The classification which best fits both the 1991 and 2001 Census categories are the seven categories (with an eighth residual category) is given opposite. It has been adopted in ONS guidance on comparing 1991 and 2001 Census ethnic group data.

The final category 'Other' is not intended for comparison between the two years: it is made up of a different mix of categories in each year.

This was one of the ONS Longitudinal Study's beta test projects, designed to check and validate the data for individuals linked across censuses up to and including 2001. Project proposal.

Reports to LS (reports 3 and 4 in particular contain material not published elsewhere):

1 Concepts

2 Group amalgamation

3 Imputation Sex, Country of Birth, Ethnic group

4 Small cells - impact of suppression on analysis