Estimating an occupational based wage in the census: a mixed model approach to generate empirical bayes estimates
Dibben, C. & Clemens, T. (2012) SLS Research Working Paper 10. Longitudinal Studies Centre Scotland: Edinburgh/St Andrews, 9 October 2012. [SLS]
Commonly, a lack of income information in the census is mitigated analytically through the use of proxy measures. Area based deprivation scores are the most common of these in which various domains of deprivation, obtained from the census, are aggregated to different spatial scales to give an indication of the characteristics of an area. Furthermore, other measures of socio-economic position such as occupational social class and education are often used to proxy some of the effect of material and financial circumstances. Though both measure independent components of socio-economic confounding (SEP) they may not capture entirely the health effects of income (Galobardes, Shaw et al. 2006). An alternative approach is to produce an estimated synthetic measure using regression modelling. However, whilst such a technique has been utilised for the production of aggregated area estimates of income (Williamson and Voas 2000), the potential for estimates generated at the individual scale, within the UK census, have yet to be investigated. The aim of this working paper is to investigate the usefulness of multilevel models based on Standard Occupation Classification (SOC) groups to estimate an occupational based wage measure in the census.
Download output document: SLS Research Working Paper 10 (PDF 392KB)