Scottish Indices of Deprivation 2003
Chapter 1: Measuring multiple deprivation at the small area level: the conceptual and technical framework
Background
The need for information about the geographical distribution of relative deprivation in Scotland has long been recognised. The most recent official study was completed in 1998 and resulted in the Scottish Area Deprivation Index for Scottish Postcode Sectors (Central Research Unit, 1998).
However, advances in the collection and use of non-census data sources, such as administrative data, have allowed analysis of deprivation at small area level in the inter-censal period. Such advances have also released the analysis from using proxy indicators from the Census in favour of using direct measures of deprivation, such as low income. In addition, recent work (see Noble, Smith, Wright et al, 2000; Noble, Smith, Penhale et al, 2000; and Noble, Smith, Wright et al, 2001) has developed the conceptualisation of multiple deprivation and its component parts, and this has been built in to the construction of the new Scottish Indices of Deprivation.
Poverty and Multiple Deprivation
In his account of Poverty in the United Kingdom, Townsend (1979, p. 31) sets out the case for defining poverty in terms of relative deprivation. Thus, 'Individuals, families and groups can be said to be in poverty if they lack the resources to obtain the types of diet, participate in the activities and have the living conditions and amenities which are customary, or at least widely encouraged or approved in the societies to which they belong'. Though 'poverty' and 'deprivation' have often been used interchangeably, many have argued that a clear distinction should be made between them. It could be argued that the condition of poverty means not having enough financial resources to meet needs. Deprivation on the other hand refers to unmet need, which is caused by a lack of resources of all kinds, not just financial. Atkinson (1998, pp. 1-20) notes that in recent debates on 'Social Europe', the terms poverty and social exclusion have been used on occasions interchangeably, but defines poverty as a 'lack of money or material possessions'. Townsend himself concurs. In his article 'Deprivation', Townsend (1987, pp. 131 and 140) argues that 'people can be said to be deprived if they lack the types of diet, clothing, housing, household facilities and fuel and environmental, educational, working and social conditions, activities and facilities which are customary …'. People are in poverty if they lack the resources to escape deprivation.
In his 1987 article Townsend elaborates on the distinctions between social and material deprivation. Social deprivation, which he acknowledges as more difficult to measure, is defined as 'providing a useful means of generalising the condition of those who do not or cannot enter into ordinary forms of family or other relationships'. In this he is anticipating some aspects of what one might now call 'social exclusion'. The more easily measured material deprivation relates to diet, health, clothing, housing, household facilities, environment and work (Townsend, 1987, p.136). Townsend also lays down the foundation for articulating multiple deprivation as an accumulation of single deprivations - a concept which is developed further in the design of the new Scottish Indices of Deprivation.
Though Townsend's work mainly (though not entirely) referred to individuals experiencing deprivations - single or multiple - the arguments can, in modified form, extend to area based measures. However, data availability inevitably means that some of the sophistication of his original concept is lost. At an area level it is very difficult to measure the percentage of the population experiencing one, two or more deprivations. It is possible to look at single deprivations at an area level and state that a certain proportion of the population experiences that deprivation, a proportion experiences some other form of deprivation etc., and at an area level describe the combination of single deprivations as area level multiple deprivation. This says little of the individual experience of multiple deprivation.
The Social Disadvantage Research Centre (SDRC) has developed the earlier work in Scotland and moved forward the measurement of multiple deprivation both conceptually and practically. The current approach is to conceptualise multiple deprivation as a composite of different dimensions or domains of deprivation. However, each dimension is measured independently using the best indicators available to generate a score or Domain Index for each aspect of deprivation. These domain scores are then combined with explicit weighting to generate an Index of Multiple Deprivation which is an aggregate of the component domains. As will be discussed, the availability of new data has allowed these domains to be described with more precision, and in a more robust and consistent way than has been possible before.
This approach allows the separate measurement of different dimensions of deprivation, such as income deprivation, education deprivation and health deprivation. There is a question as to whether there should be an additional domain for low income, or one that measures the lack of socially perceived necessities (e.g. adequate diet, consumer durables, ability to afford social activities etc.). To follow Townsend, within a Multiple Deprivation Index only the deprivations flowing from a low income would be included so low income itself would not be a component, but socially perceived necessities would. However, there are no readily available small area data on the lack of socially perceived necessities and therefore low income is an important indicator for these aspects of material deprivation. Moreover, there are arguments that measures of consumption are themselves problematic as the lack of certain items may be by choice rather than due to the inability to pay for them. Therefore, it is appropriate to measure low income itself rather than the possession of certain items.
The research team recognise income deprivation in its own right but would not argue that it should be the only measure of area deprivation. Many dimensions of deprivation are measured in the Scottish Indices of Deprivation more directly than before. The other dimensions of deprivation contribute crucial further information about an area. However, low income remains a central component of the definition of multiple deprivation for the Scottish Indices of Deprivation. As Townsend (1987, p.131) writes 'while people experiencing some forms of deprivation may not all have low income, people experiencing multiple or single but very severe forms of deprivation are in almost every instance likely to have very little income and little or no other resources.'
Multiple deprivation is not a separate form of deprivation. It is simply a combination of more specific forms of deprivation, which themselves can be more or less directly measurable. It is an empirical question whether combinations of these different forms of deprivation are more than the sum of their parts, that is, they are not simply additive but interact and may have more impact, if found in certain combinations.
This perspective accommodates the reality of varying combinations of deprivation and disadvantage in different types of areas, which has been a persistent finding on the geographical distribution of different forms of deprivation and disadvantage since the pioneering work of Richard Webber in Liverpool in the 1970s. But it does raise questions about the addition of items to form an Index of 'multiple deprivation'. And if multiple deprivation cannot be directly quantified then there are problems in technically validating any overall Multiple Deprivation Index as 'validation' requires something against which the Multiple Deprivation Index can itself be measured. If this is correct, then the question of how components in the overall Multiple Deprivation Index might be weighted rightly becomes a central question.
Individual, Household and Area Level Deprivation
Measuring different aspects of deprivation and combining these into an overall Multiple Deprivation Index raises a number of questions about the links between different forms of deprivation at the individual, household and area level. First, how far do individuals and families experiencing deprivation in fact cluster together geographically, and how far are other individuals and families who are not experiencing deprivation affected by the overall level of deprivation in their area? Though much of the data collected may, in the final analysis, be based on individual or household levels of deprivation, the results in any composite Index are likely to be presented in the form of an area level aggregate score. However, this may combine deprivations experienced by many different groups within that area.
The approach to measuring multiple deprivation
This brief debate on poverty and deprivation has underlined the importance of the financial component in any overall measurement of deprivation. This theme runs through the research team's approach to multiple deprivation and the basis for weighting the components in any overall multiple deprivation measure.
The debate also confirms the idea of separate 'domains' of deprivation, which any individual may experience singly or in combination. The intention has been to find ways of assessing the major forms of deprivation to create a robust deprivation Index for each domain - that is for income, unemployment, education etc. separately.
This approach implies rather more items in total than in earlier overall measures of deprivation. It also requires 'domain specific' items (and not what often happens when for example, 'educational deprivation' is measured by a mix of education, social and economic factors in combination). It also requires procedures for combining items within any Domain Index according to clear rules.
Particular Issues
In a country with large rural areas there is a particular need for indicators to capture the deprivation experienced by people living in rural areas. Several features of the Scottish Indices of Deprivation address these concerns. Firstly, all indicators included in the Scottish Indices of Deprivation had to be applicable to the whole of Scotland, so that fair scores and comparisons between areas could be made. This means that all of the indicators in the Scottish Indices of Deprivation are applicable to both rural and urban areas. For example, high rates of unemployment can in principle occur in both rural and urban wards, and would then be captured by the rate of people claiming the relevant benefits. Second, indicators which have different 'meanings' depending on their location have not been included. Thus, car ownership, which has previously been used as a proxy for low income, has not been included, as in some areas people might choose to make other financial sacrifices because they need a car to get to work if there is limited public transport. Low income itself is measured directly in the Income Domain by the inclusion of counts of people claiming benefits. It is also appreciated that 'rural' areas do not necessarily share all the same characteristics, and that these differences will be revealed in the ward scores and ranks. This is of course also true of the differences between and within 'urban' areas.
The Geographical Scale of the Indices of Deprivation
An ideal measure of area deprivation would be able to quantify deprivation at the level of 'neighbourhoods' or small areas with a homogeneity of characteristics and a standard population size. Such a measure would be able to offer to policy makers and funding programmes a robust method of describing the geographic distribution of deprivation.
This raises several issues for a potential index of multiple deprivation. First, the possible availability of data and population estimates for small areas; second, the problem of the variation in population size between areas and third, the heterogeneity of areas.
The availability of data for small areas in the inter-censal period had been considerably advanced by the increasing accessibility of benefits and other data at a sub-local authority level. Population estimates at ward level for mid-2000 were made available to the project by the General Register Office for Scotland (GROS). Denominators have been constructed for the indicators using these estimates. An adjustment was made to construct estimates for children under the age of 16, and to obtain a working-age population (16-59).
In order to compare concentrations of deprivation between areas it is necessary to calculate rates in order to standardise for differences in the size of the population. Although wards vary in population size, the problem of varying sizes is inherently much greater for local authorities. This raises the problem of the heterogeneity or homogeneity of an area, in terms of the intensity of deprivation. Rates for large areas, such as a whole city, may mask a great variation within it. This means that because of the heterogeneity within local authorities, as well as the variations in population size, comparisons of the intensity of deprivation at the local authority level are problematic. The ward level thus presents the most robust small area option for the Index of Multiple Deprivation.
Because of the availability of small area data, and the techniques advanced by the research team to harness other data sets, it was possible to construct the Scottish Indices of Deprivation at ward level. The Indices are based on the 1999 electoral ward geography. Ward boundaries in Scotland changed significantly in 1999 and the Indices have been constructed to reflect these revisions. The thirty-two local authorities were largely unchanged at this time.