Scottish Index of Multiple Deprivation 2006: General Report

Methodology

The methodology used to construct the SIMD 2006 remains fundamentally the same as that used for the SIMD 2004, with some minor improvements. The methodology is based on the techniques developed by Oxford University and is described in full in the SIMD 2004 technical report. For the SIMD 2006, there have been several improvements to the methodology and data sources.

The main changes to the index are the addition of a new crime domain and new public transport sub-domain, the removal of shrinkage from the process of constructing the health and education indicators, and changing from direct to indirect age-sex standardisation for selected health and education indicators.

These changes along with improvements to individual indicators are described in full in the SIMD 2006 technical report. The general methodology is described briefly below.

Constructing the index

The domains included in the SIMD 2006 are:

Current Income
Employment
Health
Education, Skills and Training
Geographic Access to Services
Housing
Crime

Each domain is made up of individual indicators which are listed in Annex 2. The domain score is calculated differently for each domain, depending on the indicators available.

The current income, employment, housing and crime domains are created by summing counts and dividing by the appropriate population denominator taken from the Census or Small Area Population Estimates ( SAPEs). For the 2006 SIMD the income and employment domains are constructed by counting the number of people claiming relevant benefits, and dividing by the appropriate population taken from the 2004 SAPE. Thus the domain scores are a simple percentage.

The housing domain is the sum of people in households that are overcrowded or have no central heating, divided by the total household population from the 2001 Census. The crime domain is a count of selected recorded crimes, called SIMD crimes, divided by the 2004 SAPE total population, but is shown as a rate of SIMD crime per 10,000 population rather than a percentage of the population.

The health, education and geographic access domains are constructed using factor analysis, which is a statistical technique that calculates weights for each indicator before they are added together to create the domain score. The indicators cannot simply be summed as they are not all counts and use different denominators. This means that the scores for these three domains are relative rather than absolute values and, as such, can not be used to measure absolute differences or absolute change.

The overall index is a weighted sum of the seven domain scores. Prior to weighting, the domains are standardised by ranking the scores. The ranks then undergo exponential transformation to avoid high ranks in one domain 'cancelling out' low ranks in another. The weights are applied to each of the domains in the SIMD 2004 and SIMD 2006 to create the overall indices. The weights are provided in Annex 2.

The resulting SIMD scores for each data zone are then ranked from one (most deprived) to 6,505 (least deprived).

A flow diagram summarising the SIMD 2006 methodology is available below.