What is the relationship between global non-response rates and imputation rates for specific sets of variables?
To quote the 2011 NHS Ethnic Origin Reference Guide:
"The imputation rates for the NHS Ethnic origin variable are similar to those of the 2006 Census (see Table 1). The NHS imputation rate for Ethnic origin at the national level is 5.8% which compares with the 2006 Census imputation rate for ethnic origin of 5.9%."
Does this suggest that the NHS data and 2006 Census data are of comparable quality at the national level for ethnic origin? If the global non-response rate (again, at the national level) is higher that the imputation rate for a given set of variables (and it was for all the groups of variables in the Reference Guides), what exactly does that mean?
Yes, the NHS data and 2006 Census data are of comparable quality at the national level for ethnic origin - as the client quoted directly from the Ethnic Origin Reference Guide below.
The GNR global non-response rate is an important measure of the quality of NHS estimates. It combines household and item non-response. This measure is used for the 2011 Census, just as it was in 2006 for dissemination of the Census, including the long form. In the specific case of the NHS, the global non-response rate is weighted to take account of the initial sample and the subsample used in non-response follow-up. It is calculated and presented for each geographic area.
The National Household Survey User Guide provides more detail - Chapter 6 would be of particular interest. http://www12.statcan.gc.ca/nhs-enm/2011/ref/nhs-enm_guide/guide_5-eng.cfm#A_6_3
Imputation replaces missing, invalid or inconsistent responses with plausible values after non-response follow-up has occurred. The National Household Survey User Guide provides more information on data processing: http://www12.statcan.gc.ca/nhs-enm/2011/ref/nhs-enm_guide/index-eng.cfm