Hi, If you're looking for a way to share social media data with additional privacy and confidentiality protections, please reach out to us at the Social Media Archive at ICPSR (somar-help@umich.edu). For sensitive data like you mention, we use a "restricted use" model that includes a strong data use agreement. a virtual data enclave, and manual disclosure risk review. This ensures that data is available for replication and extension but with protections for folks represented in the data. You can learn more about how ICPSR approaches confidentiality here: https://www.icpsr.umich.edu/web/pages/datamanagement/confidentiality/index.h... You can sign up for updates about SOMAR here: https://www.socialmediaarchive.org/ Given the terms under which you collected the data (the understanding you and the admins shared), this data may require additional constraints on its use (e.g., replication only maybe, much like if you had explicit informed consent for a specific research question but not others). We'd be happy to talk to you about options. We are currently testing our infrastructure and curating data for release. We expect to officially launch the archive in early 2023. Take care, Libby -- Libby Hemphill pronouns: she/her/hers Director, Resource Center for Minority Data <http://www.icpsr.umich.edu/RCMD>, ICPSR <http://www.icpsr.umich.edu/icpsrweb/> Director, Social Media Archive <http://socialmediaarchive.org/>, ICPSR <http://www.icpsr.umich.edu/icpsrweb/> Associate Director, Center for Social Media Responsibility <http://csmr.umich.edu/> Research Associate Professor, Institute for Social Research <http://home.isr.umich.edu/> Associate Professor, School of Information <https://www.si.umich.edu/> University of Michigan On Thu, Nov 17, 2022 at 10:54 PM Luke Munn via Air-L < air-l@listserv.aoir.org> wrote:
Marcela is correct to highlight the dangers of re-identification. You're probably thinking of Latanya Sweeney's work in this space, which has been around for a couple decades and has been highly cited.
https://www.forbes.com/sites/adamtanner/2013/04/25/harvard-professor-re-iden...
https://dataprivacylab.org/projects/identifiability/paper1.pdf
https://dl.acm.org/doi/10.1142/S0218488502001648
ngā mihi / best, Luke
On Fri, 18 Nov 2022 at 12:43, Marcela Canavarro via Air-L < air-l@listserv.aoir.org> wrote:
Hello!
Modern data protection laws rule that anonimyzed data is when the person is not identified OR identifiable. You can say that, given available means today, you cannot guarantee that the individuals cannot be identifiable through data inference.
Check out GDPR - the European Data Protection law - and LGPD - the Brazilian one. They both define anonymized data in that way. It is also a good idea to check how the californian data protection act treat this topic.
Also, there is a famous study where a huge anonymized data set was openly available and it took just a few hours for many people to be identified through data inference. It might be a nice context for your argument.
I hope that helps.
Marcela.
Em quinta-feira, 17 de novembro de 2022, Dr. Emma Briant via Air-L < air-l@listserv.aoir.org> escreveu:
Hi there, It sounds a very reasonable concern. I would just explain this in a written response to that point in the reviewer’s comments. You don’t necessarily need to do everything each reviewer says, you just need to address their concern, show you’ve considered it and explain why you chose to do what you did, that you have good reason for taking this approach. Best of luck, Emma
On Thu, 17 Nov 2022 at 20:17, Xanat Meza via Air-L < air-l@listserv.aoir.org> wrote:
Hello everyone! I have a situation with some social media data we collected for a paper that is currently under revision. The situation is complex because we collected this data from a Facebook group dedicated to a rare medical condition back in 2017, when social media data rules were not as strict as they are now. When we requested ethics review from our institution, they even told us we did not have to do anything in particular and got the study approval without much difficulty. Therefore, we requested permission to the group administrators to collect posts, posted a permission request for the group members, asking them to send us a message if they wished to opt out from the study, and collected one thousand posts by hand. We noted that there were many researchers in this group and that the members participated in surveys and medical studies frequently and with enthusiasm, so we thought at that time that an opting-out format would be enough. The data basically consists on user name, the texts in the posts, date and time of the posts, number of replies, and reactions. We separated the user names and replaced them with alphanumeric codes. Now, a reviewer of our paper is insisting that we MUST share this data openly because it is anonymized. However, we think that it should be available upon request, as social media data management has become stricter in recent years, particularly on the case of data from vulnerable communities. If we place this data related to a rare medical condition in an open repository, even people who are not researchers may have access to it and use it to bully this Facebook group, even if in theory, they could not target specific users. Does anyone have any ideas or advise on how we can respond politely to this reviewer that it is safer for everyone to keep the data available upon request? Xanat V. Meza
Ph.D. Kansei, Behavioral and Brain SciencesUniversity of Tsukuba M.A. Media and Communication Yeungnam University B.D. Graphic Communication Design Universidad Autonoma Metropolitana _______________________________________________ The Air-L@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
-- Dr Emma L Briant
Owner: Maven of Persuasion LLC Fellow at Central European University's Center for Media, Data and Society Associate at University of Cambridge, Center for Financial Reporting & Accountability Patreon: https://www.patreon.com/emmalbriant Website: www.emma-briant.co.uk
Author of: *Propaganda and Counter-Terrorism: Strategies for Global Change *from Manchester University Press Co-Author of: *Bad News for Refugees* with Prof. Greg Philo and Dr. Pauline Donald from Pluto Press. Sign up for my Newsletter <http://emma-briant.co.uk/sign-up-for-important-updates/>! Follow me on Twitter @emmalbriant
"I thought yesterday was the first day of the rest of my life but it turns out today is." - Steve Martin _______________________________________________ The Air-L@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/ listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
_______________________________________________ The Air-L@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
_______________________________________________ The Air-L@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/