Imagined communities Awareness, information sharin

ImaginedCommunities:Awareness,InformationSharing,andPrivacyontheFacebookPre-proceedingsversion.PrivacyEnhancingTechnologiesWorkshop(PET),2006AlessandroAcquisti1andRalphGross21H.JohnHeinzIIISchoolofPublicPolicyandManagement2DataPrivacyLaboratory,SchoolofComputerScienceCarnegieMellonUniversity,Pittsburgh,PA15213Abstract.OnlinesocialnetworkssuchasFriendster,MySpace,ortheFacebookhaveexperiencedexponentialgrowthinmembershipinrecentyears.Thesenetworksofferattractivemeansforinter-actionandcommunication,butalsoraiseprivacyandsecurityconcerns.InthisstudywesurveyarepresentativesampleofthemembersoftheFacebook(asocialnetworkforcollegesandhighschools)ataUSacademicinstitution,andcomparethesurveydatatoinformationretrievedfromthenet-workitself.Welookforunderlyingdemographicorbehavioraldifferencesbetweenthecommunitiesofthenetwork’smembersandnon-members;weanalyzetheimpactofprivacyconcernsonmembers’behavior;wecomparemembers’statedattitudeswithactualbehavior;andwedocumentthechangesinbehaviorsubsequenttoprivacy-relatedinformationexposure.Wefindthatanindividual’sprivacyconcernsareonlyaweakpredictorofhismembershiptothenetwork.Alsoprivacyconcernedindivid-ualsjointhenetworkandrevealgreatamountsofpersonalinformation.Somemanagetheirprivacyconcernsbytrustingtheirabilitytocontroltheinformationtheyprovideandtheexternalaccesstoit.However,wealsofindevidenceofmembers’misconceptionsabouttheonlinecommunity’sactualsizeandcomposition,andaboutthevisibilityofmembers’profiles.1Introduction“Studentslivinginthescholarshiphalls[ofKansasUniversity]werewrittenupinearlyFebruaryforpicturesonfacebook.comthatindicatedapartyviolatingthescholarshiphallsalcoholpolicy”[1].“‘StanSmith’(nothisrealname)isasophomoreatNorwichUniversity.HeismajoringincriminaljusticeeventhoughheadmitstoshopliftingonhisMySpacepage”[2].“Corporationsareinvestingintext-recognitionsoftwarefromvendorssuchasSAPandIBMtomonitorblogsbyemployeesandjobcandidates”[3].Althoughonlinesocialnetworksareofferingnovelopportunitiesforinteractionamongtheirusers,theyseemtoattractnon-users’attentionparticularlybecauseoftheprivacyconcernstheyraise.Suchconcernsmaybewellplaced;however,onlinesocialnetworksarenolongernichephenomena:millionsofpeoplearoundtheworld,youngandold,knowinglyandwillinglyuseFriendster,MySpace,Match.com,LinkedIn,andhundredothersitestocommunicate,findfriends,dates,andjobs-andindoingso,theywittinglyrevealhighlypersonalinformationtofriendsaswellasstrangers.Nobodyisliterallyforcedtojoinanonlinesocialnetwork,andmostnetworksweknowaboutencourage,butdonotforceuserstoreveal-forinstance-theirdatesofbirth,theircellphonenumbers,orwheretheycurrentlylive.Andyet,onecannothelpbutmarvelatthenature,amount,anddetailofthepersonalinformationsomeusersprovide,andponderhowinformedthisinformationsharingis.Changingculturaltrends,familiarityandconfidenceindigitaltechnologies,lackofexposureormemoryofegregiousmisusesofpersonaldatabyothersmayallplayaroleinthisunprecedentedphenomenonofinformationrevelation.Yet,onlinesocialnetworks’securityandaccesscontrolsareweakbydesign-toleveragetheirvalueasnetworkgoodsandenhancetheirgrowthbymakingregistration,access,andsharingofinformationuncomplicated.Atthesametime,thecostsofminingandstoringdatacontinuetodecline.Combined,thetwofeaturesimplythatinformationprovidedevenonostensiblyprivatesocialnetworksis,effectively,publicdata,thatcouldexistforaslongasanybodyhasanincentivetomaintainit.Manyentities-frommarketerstoemployerstonationalandforeignsecurityagencies-mayhavethoseincentives.Inthispaperwecombinesurveyanalysisanddataminingtostudyonesuchnetwork,cateredtocollegeandhighschoolcommunities:theFacebook(FB).WesurveyarepresentativesampleofFBmembersataUScampus.Westudytheirprivacyconcerns,theirusageofFB,theirattitudestowardsitaswellastheirawarenessofthenatureofitscommunityandthevisibilityoftheirownprofiles.Inparticular,welookforunderlyingdemographicorbehavioraldifferencesbetweenthecommunitiesofthenetwork’smembersandnon-members;weanalyzetheimpactofprivacyconcernsonmembers’behavior;wecomparemembers’statedattitudeswithactualbehavior;andwedocumentthechangeinbehaviorsubsequentinformationexposure:whousestheFacebook?Why?Aretheresignificantdifferencesbetweenusersandnon-users?Whydopeoplerevealmoreorlesspersonalinformation?Howwelldotheyknowtheworkingsofthenetwork?Ourstudyisbasedonasurveyinstrument,butiscomplementedbyanalysisofdataminedfromthenetworkbeforeandafterthesurveywasadministered.WeshowthattherearesignificantdemographicdifferencesbetweenFBmemberandnon-members;thatalthoughFBmembersexpress,ingeneral,significantconcernabouttheirprivacy,theyarenotparticularlyconcernedfortheirprivacyonFB;thataminorityyetsignificantshareoftheFBpopulationattheCampuswesurveyedisunawareoftheactualexposureandvisibilityoftheinformationtheypublishonFB;andwedocumentthatprimingaboutFB’sinformationpracticescanaltersomeofitsmembers’behavior.Therestofthepaperisorganizedasfollows.InSection2wediscusstheevolutionofonlinesocialnetworksandFBinparticular.InSection3wehighlightthemethodsofouranalysis.InSection4wepresentourresults.InSection5wecomparesur