I know that this is a "Friday afternoon" type discussion but, suppose the research was being carried out by a Local Authority/Residential Social Landlord that had 1,000 properties that became vacant over the year and "x" in my example was 999. The researcher could use this to show that:
(1) The non white person took longer to be rehoused (say, for example, no white person took more than 8 months to be rehoused, there would obviously be a cause for concern)
(2) There were less non white people rehoused than the proportion of non white people in the country's/city's population.
Whilst not showing the causes for this situation, it would be a valid basis for the organisation to initiate more investigation.
I don't see how you would "follow the individual" in the initial research.
Tony
-----Original Message-----
From: This list is for those interested in Data Protection issues
[mailto:[log in to unmask]]On Behalf Of Ian Welton
Sent: 17 February 2005 16:33
To: [log in to unmask]
Subject: Re: 'pure' research
Tim Trent on Thursday, February 17, 2005 at 3:11 PM said:-
> It comes down to sample definition.
>
> The smaller the sample size the less sensible it is to use a
> statistical research mechanism because you will be chasing
> results from sub populations of one.
>
> For the purposes of the type of study you outline it seems
> more sensible to "follow the individual" (ignoring any DP
> implications) rather than to attempt a nonsensical anonymisation.
Sample definition and ;
acceptable methodology matching the required ethical stance of those
concerned, which, one hopes, maintains appropriate levels of respect (by
various means) towards the data subject(s).
Ian W
> -----Original Message-----
> From: This list is for those interested in Data Protection
> issues [mailto:[log in to unmask]] On Behalf Of Tim Trent
> Sent: Thursday, February 17, 2005 3:11 PM
> To: [log in to unmask]
> Subject: Re: 'pure' research
>
>
> It comes down to sample definition.
>
> The smaller the sample size the less sensible it is to use a
> statistical research mechanism because you will be chasing
> results from sub populations of one.
>
> For the purposes of the type of study you outline it seems
> more sensible to "follow the individual" (ignoring any DP
> implications) rather than to attempt a nonsensical anonymisation.
>
> -----Original Message-----
> From: Smith, Tony [mailto:[log in to unmask]]
> Sent: 17 February 2005 14:04
> To: Tim Trent; [log in to unmask]
> Subject: RE: 'pure' research
>
> Surely there could be "pure" research, for example, that
> looked at the time that it took local people in a small area
> to get rehoused. The results may then say that in one
> specific ward there were x people who were white who were
> waiting to be rehoused and they were so within 6 months and
> that there was 1 person who was not white who was not
> rehoused for a year. Would this be pure research? It should
> not take too much research to find out who the person was!
> Additionally, pure research could look at why the person
> needed rehousing in the first place!
>
> Tony
>
> -----Original Message-----
> From: This list is for those interested in Data Protection
> issues [mailto:[log in to unmask]]On Behalf Of Tim Trent
> Sent: 17 February 2005 12:23
> To: [log in to unmask]
> Subject: Re: 'pure' research
>
>
> One has to look at what "research" is.
>
> Pure research is an anonymised statistical study of a group
> of unidentified research subjects. It creates an analysis of
> results by set. If it identified an individual that would
> limit its value as pure research since perceptions of the
> individual would be used to interpret the results.
>
> Passing the results of the individual to the individual are
> the passing of a sample of one to a sample of one. The
> individual has a right to see the answers, naturally.
>
> Passing attributable results to a third party are outside the
> terms of reference of pure research. It is possible to
> declare at the start of a study "By taking part you agree
> that your attributable results may be passed to (list of
> third parties or statement about third party access) for the
> following purposes (list of purposes)", and that removes the
> DPA issue.
>
> It also taints the research since the sample is necessarily
> skewed to include only those who do not mind their data being
> passed to third parties. Conducting pure research on a skewed
> sample invalidates the results of the research. There are
> circumstances where one could argue against this, including
> the comparison of purposely skewed samples against other
> purposely skewed samples, but that is way beyond the initial question.
>
> -----Original Message-----
> From: This list is for those interested in Data Protection
> issues [mailto:[log in to unmask]] On Behalf Of Ian Welton
> Sent: 17 February 2005 12:15
> To: [log in to unmask]
> Subject: Re: [data-protection] 'pure' research
>
> Tim Trent on Thursday, February 17, 2005 at 10:25 AM said:-
>
> > At the point of delivering attributable data to the school
> (or to any
> > third
> > party) the research is no longer pure.
>
> Are all the exemptions contained within the DPA permissible,
> or is research purity tainted in the same way?
>
> Ian W
>
> > -----Original Message-----
> > From: This list is for those interested in Data Protection issues
> > [mailto:[log in to unmask]] On Behalf Of Tim Trent
> > Sent: Thursday, February 17, 2005 10:25 AM
> > To: [log in to unmask]
> > Subject: Re: 'pure' research
> >
> >
> > At the point of delivering attributable data to the school
> (or to any
> > third
> > party) the research is no longer pure.
> >
> > -----Original Message-----
> > From: This list is for those interested in Data Protection issues
> > [mailto:[log in to unmask]] On Behalf Of Okey, Andrew
> > Sent: 17 February 2005 10:21
> > To: [log in to unmask]
> > Subject: [data-protection] 'pure' research
> >
> > Hi folks,
> >
> > Today's HE teaser is -
> >
> > Research staff undertake work in schools. A group of
> pupils, some with
> > learning difficulties, see a play and are then asked what they
> > remember about it (the point of the research is to
> establish whether
> > children with learning difficulties are potentially less reliable
> > witnesses in court). As part of the set-up for the work,
> all children
> > likely to be involved are given an adapted IQ test to inform the
> > analysis of their later recollections.
> >
> > So far, so 'pure' and, of course, pure research enjoys certain
> > exemptions from DP law, notably from access under section
> 7, and from
> > principles 2 and 5.
> >
> > Now, the researchers suspect that they may get enquiries
> from parents
> > whose children are involved in the work and who are interested in
> > knowing their IQ results. That looks OK, because section
> 33.5 of the
> > act says that research is still pure even where the data is
> disclosed
> > to "the data subject of a person acting on their behalf".
> However, the
> > researchers are concerned that the schools involved may also
> > ask/demand to know such results. Now, research is only pure when it
> > does not "support measures or decisions" made on/against
> individuals,
> > and one could imagine that what the schools would be trying
> to obtain
> > are measures of children's IQ on which they might conceivably make
> > decisions - e.g. about streaming.
> >
> > At which point, if the researchers gave this data to
> schools would the
> > research no longer be pure, thereby removing all the
> exemptions they'd
> > usually enjoy?
> >
> > Andrew Okey
> > Lancaster University
> >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > All archives of messages are stored permanently and are
> > available to the world wide web community at large at
> > http://www.jiscmail.ac.uk/lists/data-protection.html
> > If you wish to leave this list please send the command
> > leave data-protection to [log in to unmask]
> > All user commands can be found at : -
> > http://www.jiscmail.ac.uk/help/commandref.htm
> > Any queries about sending or receiving message please send
> to the list
> > owner
> > [log in to unmask]
> > (all commands go to [log in to unmask] not the list please)
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > All archives of messages are stored permanently and are
> > available to the world wide web community at large at
> > http://www.jiscmail.ac.uk/lists/data-protection.html
> > If you wish to leave this list please send the command
> > leave data-protection to [log in to unmask]
> > All user commands can be found at : -
> > http://www.jiscmail.ac.uk/help/commandref.htm
> > Any queries about sending or receiving message please send
> to the list
> > owner
> > [log in to unmask]
> > (all commands go to [log in to unmask] not the list please)
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >
>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> All archives of messages are stored permanently and are
> available to the world wide web community at large at
> http://www.jiscmail.ac.uk/lists/data-protection.html
> If you wish to leave this list please send the command
> leave data-protection to [log in to unmask]
> All user commands can be found at : -
> http://www.jiscmail.ac.uk/help/commandref.htm
> Any queries about sending or receiving message please send to
> the list owner
> [log in to unmask]
> (all commands go to [log in to unmask] not the list please)
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> All archives of messages are stored permanently and are
> available to the world wide web community at large at
> http://www.jiscmail.ac.uk/lists/data-protection.html
> If you wish to leave this list please send the command
> leave data-protection to [log in to unmask]
> All user commands can be found at : -
> http://www.jiscmail.ac.uk/help/commandref.htm
> Any queries about sending or receiving message please send to
> the list owner
> [log in to unmask]
> (all commands go to [log in to unmask] not the list please)
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> All archives of messages are stored permanently and are
> available to the world wide web community at large at
> http://www.jiscmail.ac.uk/lists/data-protection.html
> If you wish to leave this list please send the command
> leave data-protection to [log in to unmask]
> All user commands can be found at : -
> http://www.jiscmail.ac.uk/help/commandref.htm
> Any queries about sending or receiving message please send to
> the list owner
> [log in to unmask]
> (all commands go to [log in to unmask] not the list please)
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
All archives of messages are stored permanently and are
available to the world wide web community at large at
http://www.jiscmail.ac.uk/lists/data-protection.html
If you wish to leave this list please send the command
leave data-protection to [log in to unmask]
All user commands can be found at : -
http://www.jiscmail.ac.uk/help/commandref.htm
Any queries about sending or receiving message please send to the list owner
[log in to unmask]
(all commands go to [log in to unmask] not the list please)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
All archives of messages are stored permanently and are
available to the world wide web community at large at
http://www.jiscmail.ac.uk/lists/data-protection.html
If you wish to leave this list please send the command
leave data-protection to [log in to unmask]
All user commands can be found at : -
http://www.jiscmail.ac.uk/help/commandref.htm
Any queries about sending or receiving message please send to the list owner
[log in to unmask]
(all commands go to [log in to unmask] not the list please)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|