[Air-L] Facebook Data of 1.2 Million Users from 2005 Released

[Fred Stutzman]

At the time - in Facebook - there really was no conception of fully public.  One could only be open to their friends or their networks.  So "public" was a function of the size of the networks in which an individual participated.  To the question of being "open to the network," in 2005 it is fairly safe to assume that 80-90% of profiles were open to the network.  We saw numbers like that fairly consistently across some studies run at the time.

Best,
Fred


On Feb 15, 2011, at 12:45 PM, Alex Leavitt wrote:

> Michael,
> 
> I read through your blog post via a link from Twitter this morning. I agree
> that's it's a big misstep to include the User IDs. But in your write-up, as
> well as on the release page of the dataset, there doesn't seem to be any
> mention of what percentage of the users (at the time of collection) had
> privacy settings enabled versus were fully public. Any idea?
> 
> Alex
> 
> ---
> 
> Alexander Leavitt
> Researcher
> Microsoft Research New England
> Microsoft Entertainment & Devices
> Convergence Culture Consortium (Comparative Media Studies, MIT)
> http://alexleavitt.com
> http://doalchemy.org
> Twitter: @alexleavitt
> 
> 
> 
> 
> On Tue, Feb 15, 2011 at 12:38 PM, Michael Zimmer <zimmerm at uwm.edu> wrote:
> 
>> Dear AoIR-ers:
>> 
>> Some might be interested in this Facebook dataset release, and the issues I
>> raise regarding its "anonymization".
>> 
>> -mz
>> 
>> 
>> Facebook Data of 1.2 Million Users from 2005 Released: Limited Exposure,
>> but Very Problematic
>> 
>> http://michaelzimmer.org/2011/02/15/facebook-data-of-1-2-million-users-from-2005-released/
>> 
>> Recently, a Facebook dataset was released consisting of the complete set of
>> users from the Facebook networks at 100 American institutions, and all of
>> the in-network “friendship” links between those users as they existed at a
>> single moment of time in September 2005. Surprisingly, it initially included
>> each users unique Facebook ID, meaning the presumed “anonymous” dataset
>> could be easily re-identified, potentially putting the personal information
>> of 1.2 million Facebook users at risk.
>> 
>> 
>> 
>> --
>> Michael Zimmer, PhD
>> Assistant Professor, School of Information Studies
>> Co-Director, Center for Information Policy Research
>> University of Wisconsin-Milwaukee
>> e: zimmerm at uwm.edu
>> w: www.michaelzimmer.org
>> 
>> 
>> _______________________________________________
>> The Air-L at 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 at 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/
> 


--
Fred Stutzman
Postdoctoral Fellow
H. John Heinz III College
Carnegie Mellon University
fred at fredstutzman.com