Dear Lee and all,
This thread has been very interesting and I have followed the topic of "where are the limits of open" for a long time.
In environmental research the most common element that might be omitted from metadata, or obfuscated, is location data. This might be the case for nesting sites of endangered species, to protect sites from vandalism, or similar cases. I assume, the latter is a common case in archaeology, too.
Regards,
Jens
--
Dr Jens Klump
Science Leader Earth Science Informatics
Geoscience Analytics Team Leader
Mineral Resources
CSIRO
E [log in to unmask] T +61 8 6436 8828
CSIRO ARRC, 26 Dick Perry Avenue, Kensington, WA 6151, Australia
www.csiro.au
http://ORCID.org/0000-0001-5911-6022
-----Original Message-----
From: Research Data Management discussion list [mailto:[log in to unmask]] On Behalf Of RESEARCH-DATAMAN automatic digest system
Sent: Monday, 20 March 2017 8:01 AM
To: [log in to unmask]
Subject: RESEARCH-DATAMAN Digest - 18 Mar 2017 to 19 Mar 2017 (#2017-61)
There are 2 messages totaling 2205 lines in this issue.
Topics of the day:
1. OA & metadata exposure in full RE: Could a dataset's metadata record be
considered sensitive? (2)
----------------------------------------------------------------------
Date: Sun, 19 Mar 2017 10:46:42 +0000
From: Marieke Polhout <[log in to unmask]>
Subject: Re: OA & metadata exposure in full RE: Could a dataset's metadata record be considered sensitive?
Dear Falk and all,
It is a very interesting question. Not all research can be archived as Open Access due to legislation and rights. The metadata is always open access though.
When it comes to the exposure of names of researchers in open access metadata of archival systems, this could cause a problem for the security of researchers in sensitive subjects like animal testing - unfortunately. The question was how to fit this into an archival system (if any).
At DANS we have some experience with sensitive information in datasets when it comes to interviewees, not researchers of animal testing as we don't store that data. (But who knows what we'll get in the future.) The anonymisation of the interviewee information only takes place in the open access metadata, not in the rest of the files. Oral History project leaders want their interviewees to tell their story without having to fear for consequences.
Is there anyone in this group who has stored animal testing data and could share how to deal with this?
Do you mention names of the researchers in archival systems?
Is this open access to the public or just a small group of people?
And are there any problems at all?
Kind regards,
Marieke Polhout MA
[log in to unmask]
________________________________
Van: Research Data Management discussion list <[log in to unmask]> namens Falk Huettmann <[log in to unmask]>
Verzonden: zaterdag 18 maart 2017 23:15
Aan: [log in to unmask]
Onderwerp: OA & metadata exposure in full RE: Could a dataset's metadata record be considered sensitive?
Dear all,
Thanks for those details,
but wait a minute:
What do people mean not to expose project and science details, when funded by the public and/or industry for the wider good ?
That’s not right, and it violates many underlying concepts and assumptions we all pursue, and we all work under.
So is somebody saying animal testing and industry people have something to hide, or they get special rules?
Are there double-standards then ?
And are we saying metadata are to be read between the lines for the details, or even worse,
they are not meaningful for what the data and projects are about ?
That’s obviously not what metadata do and want to achieve and for what we do here and teach to students
(fine-tuning metadata on purpose is certainly not smart but harmful).
Sorry to report this to you, but it is 100% wrong and unacceptable. And yes, I do live in the real world.
Unless I am informed differently, this looks like destructive science, and not repeatable and not transparent
(=the pillars of science and its entities).
If we see such things, we fully expose it and put it out in public and at the category of ‘poor’ science and not Open Access.
Funds not well spend and to be inquired further to clarify.
Just as a side comment:
I always was told animal experiments are useful and serve the wider purpose; same with most industrial research; nothing to hide.
So I really get confused now.
Same applies to oil and gas, fracking and mining, and pharmaceutical research! We see this problem daily though, btw.
If there is a funding and withdrawing public data argument behind it in such metadata, data and its research, then we all
see exactly the crux: big money cannot buy secrecy but still trying to look good and transparent in the public eye!
We all play to the rules; so same applies here (remember Volkswagen Diesel etc air pollution scandal …or Watergate etc ?)
So in metadata and open access there is not much wiggle room, diplomacy and for getting by on the cheap and in ‘the dark’;
it’s all pretty black/white and exposed. And thanks god that is.
Please clarify as needed! Happy to discuss more as you require.
Very best
Falk Huettmann
Falk Huettmann PhD, Associate Professor
-EWHALE lab- Biology and Wildlife Dept., Institute of Arctic Biology
419 IRVING I, University of Alaska Fairbanks AK 99775-7000 USA
Email [log in to unmask]<mailto:[log in to unmask]> Phone 907 474 7882 Fax 907 474 6716
From: Research Data Management discussion list [mailto:[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Marieke Polhout
Sent: Saturday, March 18, 2017 2:01 AM
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Re: Could a dataset's metadata record be considered sensitive?
Hi Lee, Daureen and others,
I agree that removing the name of the species would make the data useless. However, I think the name of the researcher who is conducting the animal testing experiment is meant here as sensitive information, not the name of the species.
At DANS (The Netherlands) we manage the online archiving system EASY: https://easy.dans.knaw.nl
Recently published in EASY - Home - EASY<https://easy.dans.knaw.nl/ui/home>
easy.dans.knaw.nl<http://easy.dans.knaw.nl>
Search scope. EASY can be searched via a general free-text search. It searches in the metadata of all published datasets, but it does not extend into the contents of ...
Any metadata in EASY is open access, thus open to the public. We make depositors of sensitive data aware of this. For example, in sensitive interview projects we advise not to mention the names of the interviewees in the metadata. A description for an interview with sensitive information could be: "This woman (1929) is from a family which supported the German occupation of The Netherlands. In this interview she tells us about her life before, during and after World War II...." etc. Only after a granted permission request from the initial researcher one is allowed to see this interview.
If mentioning the name of the researcher in metadata is a problem, one could choose to just mention the organization as the Creator (DC-terms). There is no personal attribution then, which could prevent problems for the researcher. However, the researcher is not getting the (positive) credits for his/her work either, in the archive at least. It depends on the situation, there is definitely some risk management involved here.
Best,
Marieke Polhout MA
Datamanager DANS
https://dans.knaw.nl/en
DANS — English<https://dans.knaw.nl/en>
dans.knaw.nl<http://dans.knaw.nl>
News Your data paper in Research Data Journal for the Humanities and Social Sciences. RDJ, published by Brill publishers and DANS, is a peer reviewed e-only open ...
________________________________
Van: Research Data Management discussion list <[log in to unmask]<mailto:[log in to unmask]>> namens Daureen Nesdill <[log in to unmask]<mailto:[log in to unmask]>>
Verzonden: vrijdag 17 maart 2017 18:21
Aan: [log in to unmask]<mailto:[log in to unmask]>
Onderwerp: Re: Could a dataset's metadata record be considered sensitive?
Hi all,
“controversial (animal testing)” - By leaving out the name of the species will make it difficult for researchers wanting data derived from XXX species.
Daureen Nesdill, MS, MLIS
Research Data Management Librarian
The Faculty Center @ the Marriott Library
[log in to unmask]<mailto:[log in to unmask]>
801-585-5975
ORCID http://orcid.org/0000-0003-0126-5038
From: Research Data Management discussion list [mailto:[log in to unmask]] On Behalf Of Gareth Knight
Sent: Friday, March 17, 2017 9:28 AM
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Re: Could a dataset's metadata record be considered sensitive?
Hello Lee, all,
Interesting question. I guess it depends upon the description contained in the metadata record. It would be sensitive if it contained information that could be used to identify participants or reveal something that wasn’t in the public domain (e.g. politician XYZ had participated in a HIV study), but wouldn’t be if it simply stated that the dataset contains identifiable variables such as names, addresses, etc. but didn’t provide examples. It could also be sensitive if it revealed that people were performing research on a topic that was classified (e.g. defence), controversial (animal testing), or potentially related to embargoed material (I’m thinking of Nature articles in particular).
Regards,
Gareth
From: Research Data Management discussion list [mailto:[log in to unmask]] On Behalf Of Lee Wilson
Sent: 17 March 2017 14:55
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Could a dataset's metadata record be considered sensitive?
Hello,
I have been unable to find a definitive answer to this question and am curious to hear the community’s thoughts. If a researcher submits sensitive data (containing PII or PHI) to a repository that is capable of providing adequate security protection and restricting access to the data, would the metadata record for this dataset also be considered sensitive? To push the question further, is it possible and under what circumstances would a metadata record for research data be considered sensitive?
I am posing the question from a Canadian context, but am also interested in hearing how this issue would be considered in other countries.
Many thanks,
Lee Wilson
Research Consultant, Data Management
------------------------------
Date: Sun, 19 Mar 2017 15:51:32 +0000
From: Daureen Nesdill <[log in to unmask]>
Subject: Re: OA & metadata exposure in full RE: Could a dataset's metadata record be considered sensitive?
Hi All,
The names of the researchers and the species of the animals being used in the research are stated in the published article. The information is already out there, so why not in the open access metadata?
Daureen
________________________________
From: Research Data Management discussion list [[log in to unmask]] on behalf of Marieke Polhout [[log in to unmask]]
Sent: Sunday, March 19, 2017 4:46 AM
To: [log in to unmask]
Subject: Re: OA & metadata exposure in full RE: Could a dataset's metadata record be considered sensitive?
Dear Falk and all,
It is a very interesting question. Not all research can be archived as Open Access due to legislation and rights. The metadata is always open access though.
When it comes to the exposure of names of researchers in open access metadata of archival systems, this could cause a problem for the security of researchers in sensitive subjects like animal testing - unfortunately. The question was how to fit this into an archival system (if any).
At DANS we have some experience with sensitive information in datasets when it comes to interviewees, not researchers of animal testing as we don't store that data. (But who knows what we'll get in the future.) The anonymisation of the interviewee information only takes place in the open access metadata, not in the rest of the files. Oral History project leaders want their interviewees to tell their story without having to fear for consequences.
Is there anyone in this group who has stored animal testing data and could share how to deal with this?
Do you mention names of the researchers in archival systems?
Is this open access to the public or just a small group of people?
And are there any problems at all?
Kind regards,
Marieke Polhout MA
[log in to unmask]
________________________________
Van: Research Data Management discussion list <[log in to unmask]> namens Falk Huettmann <[log in to unmask]>
Verzonden: zaterdag 18 maart 2017 23:15
Aan: [log in to unmask]
Onderwerp: OA & metadata exposure in full RE: Could a dataset's metadata record be considered sensitive?
Dear all,
Thanks for those details,
but wait a minute:
What do people mean not to expose project and science details, when funded by the public and/or industry for the wider good ?
That’s not right, and it violates many underlying concepts and assumptions we all pursue, and we all work under.
So is somebody saying animal testing and industry people have something to hide, or they get special rules?
Are there double-standards then ?
And are we saying metadata are to be read between the lines for the details, or even worse,
they are not meaningful for what the data and projects are about ?
That’s obviously not what metadata do and want to achieve and for what we do here and teach to students
(fine-tuning metadata on purpose is certainly not smart but harmful).
Sorry to report this to you, but it is 100% wrong and unacceptable. And yes, I do live in the real world.
Unless I am informed differently, this looks like destructive science, and not repeatable and not transparent
(=the pillars of science and its entities).
If we see such things, we fully expose it and put it out in public and at the category of ‘poor’ science and not Open Access.
Funds not well spend and to be inquired further to clarify.
Just as a side comment:
I always was told animal experiments are useful and serve the wider purpose; same with most industrial research; nothing to hide.
So I really get confused now.
Same applies to oil and gas, fracking and mining, and pharmaceutical research! We see this problem daily though, btw.
If there is a funding and withdrawing public data argument behind it in such metadata, data and its research, then we all
see exactly the crux: big money cannot buy secrecy but still trying to look good and transparent in the public eye!
We all play to the rules; so same applies here (remember Volkswagen Diesel etc air pollution scandal …or Watergate etc ?)
So in metadata and open access there is not much wiggle room, diplomacy and for getting by on the cheap and in ‘the dark’;
it’s all pretty black/white and exposed. And thanks god that is.
Please clarify as needed! Happy to discuss more as you require.
Very best
Falk Huettmann
Falk Huettmann PhD, Associate Professor
-EWHALE lab- Biology and Wildlife Dept., Institute of Arctic Biology
419 IRVING I, University of Alaska Fairbanks AK 99775-7000 USA
Email [log in to unmask]<mailto:[log in to unmask]> Phone 907 474 7882 Fax 907 474 6716
From: Research Data Management discussion list [mailto:[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Marieke Polhout
Sent: Saturday, March 18, 2017 2:01 AM
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Re: Could a dataset's metadata record be considered sensitive?
Hi Lee, Daureen and others,
I agree that removing the name of the species would make the data useless. However, I think the name of the researcher who is conducting the animal testing experiment is meant here as sensitive information, not the name of the species.
At DANS (The Netherlands) we manage the online archiving system EASY: https://easy.dans.knaw.nl
Recently published in EASY - Home - EASY<https://easy.dans.knaw.nl/ui/home>
easy.dans.knaw.nl<http://easy.dans.knaw.nl>
Search scope. EASY can be searched via a general free-text search. It searches in the metadata of all published datasets, but it does not extend into the contents of ...
Any metadata in EASY is open access, thus open to the public. We make depositors of sensitive data aware of this. For example, in sensitive interview projects we advise not to mention the names of the interviewees in the metadata. A description for an interview with sensitive information could be: "This woman (1929) is from a family which supported the German occupation of The Netherlands. In this interview she tells us about her life before, during and after World War II...." etc. Only after a granted permission request from the initial researcher one is allowed to see this interview.
If mentioning the name of the researcher in metadata is a problem, one could choose to just mention the organization as the Creator (DC-terms). There is no personal attribution then, which could prevent problems for the researcher. However, the researcher is not getting the (positive) credits for his/her work either, in the archive at least. It depends on the situation, there is definitely some risk management involved here.
Best,
Marieke Polhout MA
Datamanager DANS
https://dans.knaw.nl/en
DANS — English<https://dans.knaw.nl/en>
dans.knaw.nl<http://dans.knaw.nl>
News Your data paper in Research Data Journal for the Humanities and Social Sciences. RDJ, published by Brill publishers and DANS, is a peer reviewed e-only open ...
________________________________
Van: Research Data Management discussion list <[log in to unmask]<mailto:[log in to unmask]>> namens Daureen Nesdill <[log in to unmask]<mailto:[log in to unmask]>>
Verzonden: vrijdag 17 maart 2017 18:21
Aan: [log in to unmask]<mailto:[log in to unmask]>
Onderwerp: Re: Could a dataset's metadata record be considered sensitive?
Hi all,
“controversial (animal testing)” - By leaving out the name of the species will make it difficult for researchers wanting data derived from XXX species.
Daureen Nesdill, MS, MLIS
Research Data Management Librarian
The Faculty Center @ the Marriott Library
[log in to unmask]<mailto:[log in to unmask]>
801-585-5975
ORCID http://orcid.org/0000-0003-0126-5038
From: Research Data Management discussion list [mailto:[log in to unmask]] On Behalf Of Gareth Knight
Sent: Friday, March 17, 2017 9:28 AM
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Re: Could a dataset's metadata record be considered sensitive?
Hello Lee, all,
Interesting question. I guess it depends upon the description contained in the metadata record. It would be sensitive if it contained information that could be used to identify participants or reveal something that wasn’t in the public domain (e.g. politician XYZ had participated in a HIV study), but wouldn’t be if it simply stated that the dataset contains identifiable variables such as names, addresses, etc. but didn’t provide examples. It could also be sensitive if it revealed that people were performing research on a topic that was classified (e.g. defence), controversial (animal testing), or potentially related to embargoed material (I’m thinking of Nature articles in particular).
Regards,
Gareth
From: Research Data Management discussion list [mailto:[log in to unmask]] On Behalf Of Lee Wilson
Sent: 17 March 2017 14:55
To: [log in to unmask]<mailto:[log in to unmask]>
Subject: Could a dataset's metadata record be considered sensitive?
Hello,
I have been unable to find a definitive answer to this question and am curious to hear the community’s thoughts. If a researcher submits sensitive data (containing PII or PHI) to a repository that is capable of providing adequate security protection and restricting access to the data, would the metadata record for this dataset also be considered sensitive? To push the question further, is it possible and under what circumstances would a metadata record for research data be considered sensitive?
I am posing the question from a Canadian context, but am also interested in hearing how this issue would be considered in other countries.
Many thanks,
Lee Wilson
Research Consultant, Data Management
------------------------------
End of RESEARCH-DATAMAN Digest - 18 Mar 2017 to 19 Mar 2017 (#2017-61)
**********************************************************************
|