Notes on 2001 relationship imputation markers



RELPIMP = 1 when at least one relationship of a household member with the LS sample member has been imputed. Applies to 8.2% of LS members at 2001 Census.
RELFIMP = 1 when at least one relationship between members of the LS sample memberís family has been imputed, regardless of whether this involves the LS member or not. Applies to 10.1% of LS members at 2001 Census.
RELHIMP = 1 whenever there is at least one imputed relationship within the LS sample memberís household, regardless of whether this is in the LS memberís family or not. Applies to 11.0% of LS members at 2001 Census.

Note: the references in the Data Dictionary to FRELAT0 and HRELAT0 are misleading.

Identifying an imputed relationship between LS member and other household member

It is possible to identify which relationship between an LS member and another household member is imputed. However, this is not possible for imputed relationships between other household members (i.e. that do not involve the LS member directly).

Characteristics of LS members with relationship imputation (flagged by RELPIMP/RELFIMP/RELHIMP)

Age(group): as age increased so in general the % of imputations decreased, taking into account the decreased % of the older population, compared to the younger population.
Sex: The imputations are split at virtually 50/50 between the sexes, for all three relationship imputation levels.
Marital status: Over half the imputations are for single people, even though married people are slightly greater in number in the total population (single: 41.9%, married: 42.9%).
Education qualifications: Those with no qualifications and those who are not applicable (0-15 years and 75+ years) are more likely to be imputed than people with qualifications. These 2 subgroups make up nearly 60% of all imputations and make up nearly 50% in the total population.
Care giving: Those providing no care contribute approximately the same % to the imputed sub-population as they do to the total population. Those providing 1-19 hrs of care per week contribute less to the imputed sub-population, compared to their contribution to the total population. In contrast those providing 20+ hrs of care per week contribute more to the imputed sub-population than they do to the overall population.
Self-reported health: The distribution of relationship imputation is not noticeably different to the overall distribution of this variable.
Limiting long term illness: Interestingly a greater proportion of people with imputed relationships than of those without any do not have an llti.
Economic activity: The distribution for those with imputed relationships does not differ noticeably from the overall distribution of this variable.

Back to list for household composition

Last modified 29 April 2009