Derived variables: ethnic group based on country of birth reported in 1981

STATA CODE

Name of derived variable: ETHGRP8
Variable label: Ethnic grouping at 1981 Census
Categories:

1 "White”
2 "Black or mixed"
3 "Indian, Pakistani, Bangladeshi, Sri Lankan"
4 "Chinese, Japanese, Far East"
5 "Other"

Derived from: COB8
Study group: People present at 1981 Census
Exclusions: None
Notes:

It is clearly unsafe to assume that a person belongs to the majority ethnic grouping of his/her country of birth. For many LS projects it will be possible to classify a sample member's ethnic group by his/her Census report at 1991 or 2001. However, this code maybe useful for sample members who were present at neither, or (using NCOB8) for other members of an LS member's household in 1981. If results from this code are compared with the LS member's reported ethnicity at the 1991 Census (for members who were found at 1991), the main difference is that this variable underestimates the proportion of 'Whites'.

The aim is to create a variable which is compatible with ETHGRP9, and the exploratory work which informed the design of that variable has been assumed to be valid in 1981 as well. For example, in 1991 most respondents born in Uganda or Kenya identified themselves as ethnically Indian so people reporting in 1981 that they were born in Uganda or Kenya have been coded as ethnically Indian. The possibility of inaccuracy at an individual level is obvious.


Back to list for ethnicity

Last modified 3 October 2011