Scottish Government

1. Introduction

1.1. SIMD 2004

The Scottish Index of Multiple Deprivation ( SIMD) 2004 ¹ is an area-based measure of deprivation defined for each of the 6,505 Scottish Neighbourhood Statistics ( SNS) Data Zones ². It is a combination of 6 measures of different aspects, or domains of deprivation:

• Current Income;
• Employment;
• Health;
• Education, Skills and Training;
• Geographic Access and Telecommunications;
• Housing.

A major application of the SIMD 2004 is the identification of small areas of severe multiple deprivation, so that Community Regeneration funds can be allocated to Local Authorities ( LAs) in relation to their needs ³.

Calculation of the SIMD 2004 was performed by the Office of the Chief Statistician ( OCS) for the Scottish Executive ( SE), using methods previously developed by the Social Disadvantage Research Centre ( SDRC) at the University of Oxford. The methodology employs a wide range of statistical techniques, many of them representing significant innovation in the small-area measurement of deprivation. However, there has been little empirical work to evaluate these methods or to compare the resultant indices with those produced using alternative methods.

1.2. Scottish Executive Strategy

In December 2002, the SE's Central Statistics Unit commissioned the Scottish Centre for Research on Social Justice to make recommendations for developing a long-term strategy for measuring deprivation in Scotland ⁴. These recommendations largely reinforced the methodology used for the production of the SIMD 2003 ⁵ and, subsequently the SIMD 2004. However, one of the specific recommendations was to commission further work to improve the shrinkage techniques used, and to reconsider the use of factor analysis for the combination of indicators in the calculation of some domain scores.

1.3. Project Brief

In November 2004, following the publication of the SIMD 2004, the SE issued invitations to tender for the provision of an "Evaluation of the Statistical Techniques in the Scottish Index of Multiple Deprivation". This was in line with the long-term strategy for the measurement of deprivation in Scotland, but was more wide ranging than originally recommended. Five areas in particular were highlighted for review; shrinkage, factor analysis, exponential transformation, weights for combining domains, and uncertainty.

1.4. Purpose Of This Report

This Report describes the work carried out by the authors, in collaboration with members of the Expert Advisory Group ( EAG), in accordance with the Evaluation Plan that evolved from the Project Brief, the Tender submitted by the Robertson Centre for Biostatistics, University of Glasgow, and from discussions between the authors, the EAG and representatives of the OCS.

Section 2 describes the initial phase of the project, in which the programs used by the OCS to produce the SIMD 2004 were validated by replicating the algorithm in another computer language and confirming that the resultant domain scores and SIMD match those produce by the OCS.

Section 3 outlines the plan that was adopted for the evaluation of the statistical methods used in the SIMD 2004. Each area of the original brief is covered separately; however, the assessment of uncertainty is discussed first, reflecting the consideration that this aspect of the project could form the main basis for the evaluation of alternative methods in other areas. The methods of evaluation of each of the five areas are then described in more detail in Sections 4 through 8. The results of these evaluations are presented in Section 9. Section 10 discusses the results of the evaluation and Section 11 lays out the recommendations of the authors. Section 12 lists the references used in this report, and a series of Appendices are included with additional Tables and Figures.

This Report is necessarily detailed and, in places, technically complex. An Executive Summary has also been written, outlining the main findings and recommendations arising from this work.