Download presentation
Presentation is loading. Please wait.
Published byReginald Nash Modified over 9 years ago
1
Child social exclusion: development of a small area indicator for Australia Justine McNamara
2
2 Acknowledgments Fellow authors: Robert Tanton, Anne Daly, Ann Harding and Mandy Yap This paper is part of a study funded by the Australian Research Council to develop spatial indicators of social exclusion for Australia’s children (DP 560192). Based on data provided by the Australian Bureau of Statistics from the 2001 Census of Population and Housing
3
3 Social exclusion n Multidimensional measure of disadvantage n Limitations of income-based measures of disadvantage
4
4 Social exclusion “Social exclusion happens when people or places suffer from a series of problems such as unemployment, discrimination, poor skills, low incomes, poor housing, high crime, ill health and family breakdown” (British Social Exclusion Unit 1997)
5
5 Spatial differences n Increasing interest in Australia in examining geographical differences in advantage and disadvantage n No previous child-focused summary measures of disadvantage at a small area level for the whole of Australia
6
6 Methodology n Composite index, but also analysis of individual variables n Exploring best way to treat income within the index
7
7 Methodology Data: Australian 2001 Census of Population and Housing Spatial Unit: Statistical Local Area (SLA) Statistical method: Principal Components Analysis
8
8 Statistical Local Areas n There are 1353 SLAs in Australia n Not evenly distributed across states/territories n Population sizes vary n Have aggregated SLAs in Brisbane and Canberra
9
9 Calculating the index n Used principal components analysis (PCA) to summarise variables into a single measure of child social exclusion risk. n PCA transforms a set of correlated data into a set of new variables or components. n The first new variable or component captures most of the variation in the original set of variables, and is used as the index.
10
10 Variables – original model
11
11 Treatment of income within the index n Rationale for making index variables dependent on income was to reflect importance of financial well- being as a variable, and to strengthen the limited range of variables available in the Census n Relatively high cut-off for “low income” meant that cross-tabulated variables not too restrictive
12
12 Treatment of income within the index Also arguments against cross-tabulating income with other index variables: n problems with income as a measure of disadvantage n missing income data n relatively modest correlation between income and other disadvantage variables in the literature n can still include it on same level as other variables
13
13 Variables – revised models
14
14 PCA results – original model and new models
15
15 Analysis of spatial distribution n Use child-population weighted quintiles to analyse distribution of social exclusion risk across Australia n Quintiles are quintiles of children, not SLAs
16
16 Models Model 1 = income inclusive variables Model 2 = non-income inclusive variables Model 3 = non-income inclusive variables, with government school variable removed
17
17 Spatial distribution of social exclusion: Model 1 versus Model 3
18
18 Model 1 versus Model 3: capital cities Note: 62% of all children live in capital cities
19
19
20
20
21
21 Conclusion In our index, treatment of income makes a difference to spatial patterns Will continue to evaluate best approach
22
22 Further work n Spatial distribution of child social exclusion over time n Use of spatial regression techniques
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.