Hi David, Sebastian, Angus, all,

A Data Sharing Agreement is essential if you're dealing with research
that involves human participants in many cases. Even if you've
anonymised a dataset by removing direct identifiers, there's still a
risk that participants will be re-identified using alternative
techniques or through use of overlapping datasets. There's a number of
high profile cases where specific individuals have been re-identified
(http://arstechnica.com/tech-policy/2009/09/your-secrets-live-online-in-databases-of-ruin/).
Most recently, as a result of the NYC taxicab debacle last year
(http://research.neustar.biz/2014/09/15/riding-with-the-stars-passenger-privacy-in-the-nyc-taxicab-dataset/).


Although journal publishers and data repositories encourage Creative
Commons, a lot of data requires additional controls. For our own
repository (due to be launched in a few weeks), I've been directing
people to our legal team if they require controlled access to data to
help them develop a Data Sharing Agreement, as it allows them to
establish additional restrictions on how data is used. The primary issue
with Data Sharing Agreements, however, is that it can take some time to
customise them to individual projects and they're written using legal
terminology which is difficult to understand for the layperson.
Personally, I'd like to see a Creative Commons-like badge system
developed that allows researchers to specify the specific requirements
that they'd like to establish. This would require national/international
collaboration, but would simplify the process of sharing research data
that can't be CC licensed.

Incidentally, I ran a seminar on lunchtime seminar on data licensing a
few weeks ago with my colleague, John Murtagh. Slides can be found at
http://www.slideshare.net/lshtm/introduction-to-data-licences and a
screen recording is available at
https://panopto.lshtm.ac.uk/Panopto/Pages/Viewer.aspx?id=3aa9b483-20fa-462f-82ac-872e5936671d
(begins at 20:14)

Regards,
Gareth



>>> Angus Whyte <[log in to unmask]> 05/06/2015 17:13 >>>
Thanks Sebastian, yes your wording is exactly what I meant i.e. CC-BY
isn't really relevant to subject of identifying individuals.

Angus


On 05/06/2015 17:06, Sebastian Rahtz wrote:
>> On 5 Jun 2015, at 16:52, Angus Whyte <[log in to unmask]> wrote:
>>
>> ….. CC-BY does not permit you to attempt to identify individuals
(etc) in the data…
> did you mean to say that? CC-BY has nothing to say on the subject of
identifying individuals.
>
> But I agree with you:  CC-By's "The license may not give you all of
the permissions necessary for your intended use” covers the “you may
not apply ….legally restrict others from doing anything the license
permits.”
> The licence neither permits you to, nor restricts  you from, breaking
other laws e.g. about privacy.
>
> People get too worked up about trying ti finagle this sort of thing.
The law, generally, isn’t that hard to conform to.
>
> Sebastian Rahtz
> Chief Data Architect
> University of Oxford IT Services
> +44 1865 283431
>
>
>
>
>


--
Dr Angus Whyte
Senior Institutional Support Officer
Digital Curation Centre
University of Edinburgh
Crichton St, Edinburgh EH8 9LE
+44-131-650-9986

The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.