Disclosure Control: Neighbourhood Statistics NeSS

1
Disclosure ControlNeighbourhood Statistics
(NeSS)

• David McLennan
• University of Oxford

2
What is NeSS?

• National Strategy for Neighbourhood Renewal
  (2001)
• Policy Action Team 18 Better Information (2000)
• comprehensive and good quality information at a
  small-area level to enable a better understanding
  of local problems and more effective targeting of
  solutions
• ONS website (www.neighbourhood.statistics.gov.uk)

3
What data does NeSS contain?

• 12 domains of indicators
• 2001 Census
• Access to Services
• Community Well-Being / Social Environment
• Crime and Safety
• Economic Deprivation
• Education, Skills and Training

• Health and Care
• Housing
• People and Society
• Physical Environment
• Work Deprivation
• Indices of Deprivation and Classification

4
E.g. of a data feed into NeSSThe English
Indices of Deprivation 2004

• The English Indices of Deprivation 2004 (ID 2004)
  consist of seven domains of deprivation measured
  at small area level (LSOA)
• Income Deprivation
• Employment Deprivation
• Health Deprivation and Disability
• Education, Skills and Training Deprivation
• Barriers to Housing and Services
• Living Environment Deprivation
• Crime

Index of Multiple Deprivation
5
Our role as data provider to NeSS E.g. The
English Indices of Deprivation 2004
6
Problems of disclosure controlE.g. ID 2004
Unemployment Indicator

• Employment Deprivation Domain 25 of overall
  Index of Multiple Deprivation Score
• Employment Deprivation Domain Unemployment Rate
  Work Limiting Illness Rate
• Unemployment Rate Numerator Claimant Count
  New Deal participants (DWP)
• Unemployment Rate Denominator estimated working
  age population (m18-64 f18-59)

7
Problems of disclosure controlE.g. ID 2004
Unemployment Indicator

• DWP applied disclosure control to Unemployment
  Rate Numerator
• Counts rounded up or down to multiple of five
• E.g. a count of 24 could change to 20 or 25
• Rounded counts of 10 or less suppressed
• So, a count of 11 could be rounded to 15 and
  retained, while a count of 14 count be rounded to
  10 and suppressed.

8
Histogram showing frequency distribution of ID
2004 Unemployment Numerator at LSOA level
9
Scatter plot showing LSOA ranks True ID 2004
rate -vs- Disclosure controlled NeSS rate
10
Problems of Disclosure Control

• Disclosure Control prevents replication of Index
  methodology by other researchers or users
• Loss of valuable data on neighbourhoods, although
  lesser impact on upper end of distribution

11
A solution?

• Continue with planned expansion of NeSS, plus
• Secure intra-net style NeSS?
• Accessible only to government departments and
  authorised researchers
• No (or minimal) disclosure control applied to
  internal data
• Emphasis on applying disclosure control to end
  product of analysis