Dear all,
Some time ago, my colleague James Wilson asked this list for advice on
the risks of anonymized data being de-anonymized as a result of
increased quantities of data being available, and there therefore being
increased opportunities for cross-referencing.
A plethora of helpful responses were received (thanks!), which I've
attempted to summarize in the attached document. If I've missed anything
important (or if any of the quoted individuals feel I've misunderstood
them), do please let me know.
The document is CC licensed, but please note that this is intended
primarily as an informal summary of advice, not as a formal publication.
(If time permits, the DaMaRO Project may use this as a basis for some
more polished guidance on this topic for researchers, which will appear
in due course.)
Best wishes,
Meriel.
--
Dr Meriel Patrick
Analyst and Training Officer, DaMaRO Project
http://damaro.oucs.ox.ac.uk/
IT Services, University of Oxford
13 Banbury Road, Oxford, OX2 6NN
Tel. 01865 (2)73260
http://www.it.ox.ac.uk/
Usual working days: Mondays, Wednesday afternoons, and Fridays
|