Disclosure control and small area statistics

1
Disclosure control and small area statistics
Grids or blobs? With comparison of two small
areas
Vilni Verner Holst Bloch Cand.scient. landscape
ecology and natural resources Statistics
Norway Otervegen 23 N - 2225 Kongsvinger Tel
47 / 62 88 50 62 Fax 47 / 62 88 54 59
E-mail vvh_at_ssb.no
2
Increasing demand for disclosure control
Even smaller area statistics - easier to
identify/recognice persons or communities
More detailed statistics more sensitive data
More available statistics easier to combine
and derive other information (classic ex.
elder single person with lots of money)
3
Disclosure depend on geography
(radius 75 m)
Addresses
Oslo municipality 66 000 addresses (coordinate
pairs)
Persons
4
Disclosure depend on topic
5
Disclosure varies with time and society
Norwegian population cencus 1910
- about amongst others returned
norwegian-americans, dissidents, blind people,
deaf people and lunitics
6
Advantages of grid
Easy data handling do not need GIS
7
Advantages of blobs
High geographical accuracy
Easy to read less noise in the map
8
A comparison
Stovner
Central Oslo
9
Information on address level
Stovner
Central Oslo
10
Residents per 100 x 100 m gridcell
Stovner
Central Oslo
11
Employees per 100 x 100 m gridcell
Stovner
Central Oslo
12
Education in neighbourhood
Stovner
Central Oslo
13
Disclosure rules in Statistics Norway
Blobs
Statistics based on this method given criterias
of radius minimum 75 m, minimum 10
livingaddreses and 10 units may be published.
Grids
Standard criterias for disclosure control yet to
be decided. In practice minimum classes has been
1-4 residents or 1-19 employees.