We are testing a new beta website for gov.scot go to new site

Improving Scottish Household Survey income estimates

Introduction

Since Autumn 2008 the Scottish Government income and poverty statistics team have been involved in a long-running project to improve the estimates of household income in the Scottish Household Survey. One of the main aims of this work is to produce more robust estimates of income and poverty at a Local Authority level than those which are currently available. This project is the result of a user need for improved LA-level data that was expressed at the poverty statistics user event in Summer 2008.

The first official estimates of relative poverty at local authority level were published during August 2010 in "Relative poverty across local authorities" which is available at the following link.

Relative poverty across Scottish local authorities

User should note however that the figures in this publication are classed as "data being developed" and are currently undergoing further quality assurance work.

Two reports have been produced which describe different elements of the development of these figures. They are available on this page. The first is a feasability study which was carried out by Ipsos Mori and proposed a methodology for improving the SHS income data. The second is the result of work, following on from the feasability study, which implemented and tested Ipsos Mori's proposed method on Scottish Government systems. This work was carried out by the Office for National Statistics Methodology Consultancy Service.

See below for further detail.

1. Feasability study - Income imputation in the SHS and SHCS

2. Report from Implementation project

SHS income imputation cover
Income imputation in the SHS and SHCS: examining the feasibility of using the FRS to broaden the measure of household income.

During autumn 2008 the Scottish Government commissioned a feasibility study into the possibility of using data from the Family Resources Survey to improve the quality of income estimates produced from the Scottish Household Survey (SHS) and Scottish House Condition Survey (SHCS). The work also included examining the way in which income data in the SHS and SHCS were recorded and processed.

The full report is available to download from the following link: Full Report

A brief summary follows:

Summary and recommendations
  • Differences in the estimates of household income between the FRS and the SHS/ SHCS are driven primarily by the inclusion of income of 'other adult' - adults who are not the Household Reference Person (HRP) or their spouse - in the FRS. Other methodological and definitional differences have a much smaller effect on the estimate of household income.
  • Overall, even though the FRS questionnaire is almost twice the length of the SHS, once this difference has been accounted for, the estimates for income levels in Scotland are very similar.
  • Around half of the variation in the income of other adults in the FRS can be accounted for by variables that are common to both the FRS and the SHS/ SHCS. The main drivers of income that could be used in an imputation strategy are economic status and relationship to the HRP. Less variation in income levels within different economic statuses can be explained by variables common to all three surveys.
  • It is feasible to impute reasonable approximations of income for other adults in the SHS and SHCS to produce estimates of household income that will be robust at the national and local authority level. Imputation of income of other adults will increase the mean estimate of household income across Scotland by around 9%.
  • There are no major practical barriers to imputing income of other adults in the SHS and the SHCS. However, the need to use case level data from the FRS would mean that the estimates of the broader definition of household income, based on the imputed data, would only be available one year after the existing estimates are released.
  • A number of other recommendations are made to improve the estimates of household income in the SHS and the SHCS.
Implementation report - Income imputation in the Scottish Hosuehold Survey

The following report gives detail about a project carried out by the Office for National Statistics Methodology Consultancy Service to test and implement the method proposed by Ipsos Mori for improving the Scottish Household Survey Income data. It is a fairly technical report and gives details of:

  • A review of the Ipsos Mori methodology
  • The results of attempts to replicate this methodology on Scottish Government IT systems and datasets
  • A detailed description of the final method used to impute SHS household incomes
  • A method to estimate the variability of the new imputed income figures
  • An assessment of the robustness of the imputed figures

The report concludes that the quality of the new estimates is "comparable to that of other similar proportion local authority SHS estimates".

To download a pdf version of this report please click on the following link: Implementation report

Further Information

Scottish Household Survey

Scottish House Condition Survey

DWP Family Resources Survey

Poverty Estimates for Scotland