Transcript Geog3025_Multivariate

Multivariate neighbourhood
indicators
GEOG3025
Multivariate neighbourhood
indicators
• Lecture overview:
Objectives of lecture
Introductory questions
Multivariate deprivation
Deprivation indicators
Calculation of indicators
Lecture summary
GEOG3025
Objectives
• Understand the concept of multiple
material deprivation
• Understand the principle of multivariate
deprivation indicators
• Be able to calculate and interpret
general-purpose indicators
GEOG3025
Introductory questions…
What makes an area
“deprived”?
How might you
measure social
deprivation?
GEOG3025
Multiple deprivation
• Recognition that deprivation is a
multifaceted phenomenon (Black Report,
1980)
• Particular focus on health care
applications
• Measurement requires combination of
multiple variables – multivariate
indicators
GEOG3025
Research areas
• Material deprivation as cause of health
inequality
• Deprivation and mortality
• Deprivation and educational
achievement
• Use in resourcing public services (e.g.
standard spending assessment)
GEOG3025
Deprivaton or exclusion?
• Multiple deprivation
– Derived directly from census data
– Derived from census and administrative
data
• Social exclusion
– Incorporating concepts of exclusion from
normal activities of citizens e.g.
consumption, savings, production, political
social (Burchardt et al. 2002)
GEOG3025
Multivariate deprivation…
Photo: Dave Martin
Photo: Dave Martin
Remoteness, elderly
population?
GEOG3025
Photo: Dave Martin
Poor transportation,
unemployment,
lone parents?
Overcrowding,
unskilled jobs?
Census-based deprivation indicators
• Jarman
– Deprviation payments to GPs 1980s, 90s
• Carstairs
– Health-related, esp. Scotland
• Townsend
– Health-based, widely used
• Department of the Environment
GEOG3025
Census deprivation index
construction: issues
•
•
•
•
•
•
Data source choice
Variable choice
To combine or not?
To standardise or not?
To transform or not?
To weight or not?
GEOG3025
Two distributions…
Number
of areas
Unemployed
Not owner occupied
Percentage
GEOG3025
Transform and/or standardise?
Standardise:adjust for
spread
GEOG3025
Transform: adjust
for shape
Example: Townsend score
• Input variables
– No car
– More than one person per room
– Not owner occupied
– Unemployed
GEOG3025
Townsend calculation
• Express variables as percentages
• Log transformation of unemployment
and overcrowding values (+ 1)
• Z score standardisation of all variables
• Sum the standardised scores with equal
weighting
GEOG3025
Townsend score for single area
Variable
No car
Overcr.
Not own
Unemp.
Score
GEOG3025
Percent
Log
Z
Townsend score for single area
Variable
Percent
No car
17
Overcr.
6
Not own
45
Unemp.
9
GEOG3025
Log
Score
Townsend score for single area
Variable
Percent
Log
No car
17
-
Overcr.
6
0.84
Not own
45
-
Unemp.
9
1
GEOG3025
Z
Townsend score for single area
Variable
Percent
Log
Z
No car
17
-
0.9
Overcr.
6
0.84
1.5
Not own
45
-
2.3
Unemp.
9
1
1.1
GEOG3025
Townsend score for single area
Variable
Percent
Log
Z
No car
17
-
0.9
Overcr.
6
0.84
1.5
Not own
45
-
2.3
Unemp.
9
1
1.1
Score
GEOG3025
5.8
Source: devon.gov.uk
Deprivation mapping
GEOG3025
Non-census/hybrid indicators
•
•
•
•
•
Index of local conditions
Welsh Jarman
Index of local deprivation
Index of multiple deprivation
Combining census and non-census
variables:
– House condition, income support, derelict
land, low GCSE grades, house insurance
weights, etc.
GEOG3025
IMD 2000
• Index of multiple
deprivation
• 6 ‘domains’:
– Income
– Employment
– Health and disablity
– Education, skills and training
– Housing
– Geographical access to
services
GEOG3025
MAUP/ecological fallacy
• Levels of measured deprivation
dependent on area boundaries
• Large areas can conceal small areas
(‘pockets’ of deprivation)
• Area-level deprivation indicators will
misrepresent many individuals and
households within the areas
GEOG3025
Structural problems…
• Different regions have different
deprivation structures
• Claims of London-bias may just
represent population figures
• Urban-rural deprivation differences
• Welsh Jarman – additional variables
GEOG3025
Assignment
•
•
•
•
•
•
•
•
Select variable set and study region
CASWEB retrieval (including denominators)
Run Geodep spreadsheet
Import dataset
Define denominators
Design indicator
Run/re-run
Interpret scores
GEOG3025
Lecture summary
• Problems with specification and design
• Data limitations – especially early
versions
• Particularly relevant to general-purpose
indicators
• Best to use purpose-specific indicators
• Key review reference: Senior (2002) in
Census Data System
GEOG3025