1 Disclosure control and small area statistics Vilni Verner Holst Bloch Cand.scient. landscape ecology and natural resources Statistics Norway Otervegen 23 N Kongsvinger Tel : ++47 / Fax : ++47 / Grids or blobs? With comparison of two small areas
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 Persons Addresses Disclosure depend on geography Oslo municipality – addresses (coordinate pairs) (radius 75 m)
4 Disclosure depend on topic
5 Disclosure varies with time and society Norwegian population cencus about amongst others; returned norwegian-americans, dissidents, blind people, deaf people and lunitics
6 Stable land units makes measuring change in geographical concentration easier (?) Measuring co-variation when lacking exact geographical match Easy data handling – do not need GIS Advantages of grid
7 Advantages of blobs High geographical accuracy Easy to read – less ”noise” in the map
8 A comparison Central Oslo Stovner
9 Central Oslo Stovner Information on address level
10 Residents per 100 x 100 m gridcell Central Oslo Stovner
11 Employees per 100 x 100 m gridcell Central Oslo Stovner
12 Education in neighbourhood Central Oslo Stovner
13 Disclosure rules in Statistics Norway Blobs Grids Statistics based on this method – given criterias of radius minimum 75 m, minimum 10 livingaddreses and 10 units – may be published. Standard criterias for disclosure control yet to be decided. In practice minimum classes has been 1-4 residents or 1-19 employees.