A Proactive Generation Reacts: A Proactive Generation Reacts: Privacy, Disclosure and Privacy, Disclosure and

Strategic Action on FacebookStrategic Action on Facebook

By Zeynep TufekciAssistant Professor

University of North Carolina, Chapel Hill

Twitter: @techsoc

zeynep@unc.edu

Reality versus PerceptionReality versus Perception

A loony, narcissistic generation A loony, narcissistic generation inexplicably oversharing just because inexplicably oversharing just because they’re differentthey’re different

Facebook merely reflecting those Facebook merely reflecting those changeschanges

Numerous assertions by Facebook Numerous assertions by Facebook executives that “one identity, one executives that “one identity, one social graph” is better, moral, normalsocial graph” is better, moral, normal

Privacy and Disclosure are:Privacy and Disclosure are:Historically-situatedHistorically-situated

Grounded by personal Grounded by personal circumstances, identit(ies), contextcircumstances, identit(ies), context

Structured and played out through Structured and played out through architecture, norms and affordancesarchitecture, norms and affordances

Always include strategic action Always include strategic action considerationsconsiderations

Popular Perception:Popular Perception:

BackgroundBackground

Facebook is a de facto norm among Facebook is a de facto norm among college studentscollege students

Facebook architecturally leans towards Facebook architecturally leans towards disclosure and a united social graphdisclosure and a united social graph

Facebook has made multiple privacy Facebook has made multiple privacy changes, often controversialchanges, often controversial

Disclosure levels tend to be highDisclosure levels tend to be high

What We KnowWhat We Know

Contrary to media perceptions, are Contrary to media perceptions, are struggling to adapt and figure outstruggling to adapt and figure out

Documenting struggle, strategy and Documenting struggle, strategy and change: Boyd & Hargittai (2010), change: Boyd & Hargittai (2010), Stutzman (2011), Raynes–Goldie, Stutzman (2011), Raynes–Goldie, Kate (2010), Pew (2010)Kate (2010), Pew (2010)

Strategies & Experiences: Strategies & Experiences: QuestionsQuestions

How have young adult behaviors How have young adult behaviors have changed over time?have changed over time?

What are their experiences with What are their experiences with privacy and disclosure?privacy and disclosure?

What are their strategies?What are their strategies?

Can we untangle experience (i.e. Can we untangle experience (i.e. Bandura’s social learning theory) and Bandura’s social learning theory) and anticipatory behaviors? anticipatory behaviors?

This StudyThis StudyCollege student sampleCollege student sample

Latest survey December 2010 (n=403)Latest survey December 2010 (n=403)

Historical surveys on comparable (but Historical surveys on comparable (but not longitudinal) groups since 2006 not longitudinal) groups since 2006 (n>1000)(n>1000)

mmDiverse, mid-sized school, socio-Diverse, mid-sized school, socio-

economic range, sample generally economic range, sample generally representative of schoolrepresentative of school

N (%)

Male 53.3

Female 45.6

White 43.0

Black 17.5

Hisp./Other 13.0

Asian-Ame. 26.5

MEAN

General Privacy Concern(1-5)

4.06

MEAN

Years on FB 3.58

Sample Characteristics

EVER USE FACEBOOK FOR: (1-4)

Keep in touch with friends in same school

3.1

Keep in touch with friends elsewhere 3.4

To find people with similar interests 1.5

To find potential friends 1.8

How concerned are you that people you do not want to see your profile would

see your profile (1-5)

3.3

Sample Descriptives

EVER DONE THE FOLLOWING BECAUSE OF A PRIVACY OR VISIBILITY CONCERN?

%

Untagged themselves from a photo 73.9Deleted info from profile 81.3Unfriend someone 68.3Deactivate their profile 19.6Those who have: %Never changed privacy settings 9.4Changed their privacy settings last month 36.7Changed their privacy settings in the last year 82.9Of those who changed ever privacy settings (n=363):

Those whose last change was to make their profile LESS visible

90.3

Did the Following Happen to YOU?AT

LEAST ONCE

Lost a job 0.8%

Did not get hired 0.5%

Had issues at work 2.5%

Had a fight with a girlfriend/boyfriend 28.7%

Broke up with a girlfriend/boyfriend 8.0%

Had a fight with a friend 25.7%

Had a fight with a parent 17.2%

Had a legal problem 0.8%

Sample Descriptives

Did the Following Happen to SOMEONE YOU KNOW?

AT LEAST ONCE

Lost a job 19.7%

Did not get hired 19.4%

Had issues at work 30.8%

Had a fight with a girlfriend/boyfriend 70.9%

Broke up with a girlfriend/boyfriend 51.2%

Had a fight with a friend 66.2%

Had a fight with a parent 53.2%

Had a legal problem 21.9%

Did Any of the Following Happen to You? (%) AT LEAST ONCE

Someone you did not want found your profile 73.9%

Other people posted pictures of you that you wish they had not

70.0%

Someone else got upset with you seeing a picture of you with someone else

25.7%

You got upset with someone after seeing a picture of theirs with someone else

39.8%

Someone harassed/stalked you online 33.6%

Someone much older contacted you 54.6%

You found out something very important about a friend 77.8%

You found that someone else had been lying to you by looking at their profile

59.6%

You got caught in a lie by someone else through your profile

32.5%

Profile Visible to Everyone

0

10

20

30

40

50

60

70

Spring 06 Fall 06 Spring 07 Fall 07 Spring 08 Fall 2010

Per

cen

t

Profile visible to everyone (%)

Changes in Profile Visibility in Facebook 2006-2010 (December)

Concerns about Privacy and Visibility of Profile

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Spring 06 Fall 06 Spring 07 Fall 07 Spring 08 Fall 2010

How concerned unwanted people wouldsee your profile? (1-5 very)

How concerned about privacy? (1-5)

Privacy Concerns and Profile Visibility in Facebook 2006-2010 (December)

exp(b) exp(b) exp(b) exp(b)

Female 1.583 1.732* 0.611 0.255**

Black 1.032 0.684 0.428 1.232

Hisp/Other 0.693 0.929 1.828 1.192

Asian 0.676 0.934 0.680 1.861

Happened to me 0.988 0.980 1.044 1.016

I observed 1.099 1.023 1.079 1.085

Surveillance 1.049 1.004 0.754*** 0.937

Far friends 1.181 1.165 0.824 0.925

Near friends 0.892 1.538* 1.186 0.907

Find potential 1.002 1.283 0.740 1.031

Find similar 0.679 0.761 1.801* 1.337

Audience Concern 1.671*** 1.791*** 0.624* 0.375***

Years on FB 1.364** 1.047 0.803 0.773*

_cons 0.097 0.008*** 22.286 8.462

N 383 383 383 383

ll -162.272 223.213 -97.021 -104.967 *

* p<0.05, ** p<0.01, *** p<0.001

Logistic Regression Modeling Odds of Privacy Related Behaviors

Less Visible Last Month Never Changed Public Profile

The Previous Table Says:The Previous Table Says:

Respondent’s concern over unwanted Respondent’s concern over unwanted audiences IS significantly associated audiences IS significantly associated with privacy-preserving actionswith privacy-preserving actions

Respondent’s experience with *or* Respondent’s experience with *or* observations of negative consequences observations of negative consequences is NOT strongly associated with privacy is NOT strongly associated with privacy related actionsrelated actions

Know Don’t Know exp(b)

exp(b) Female -0.098 -0.141

Black 0.100 0.129

Hisp/Other 0.209 -0.021

Asian 0.243** 0.145

Someone else -0.007 -0.003

Me 0.015 -0.032

Surveillance 0.010

0.036***Far friends -0.023 0.104

Near friends 0.017 -0.026

Find potential 0.039

0.156** Find similar 0.029

0.138* Profile concern -0.033 -

0.164***Years on FB 0.005 -0.015

_cons 3.219***

1.494***N 381 381

ll -365.679 -420.871

* p<0.05, ** p<0.01, *** p<0.001

Logistic Regression Modeling Odds of Accepting Friend Request from Person already known / not known

The Previous Table Says: The Previous Table Says: When it Comes to Accepting When it Comes to Accepting

Friend Request from Unknown Friend Request from Unknown Person:Person:

Respondent’s who are interested in Respondent’s who are interested in finding similar people and expanding finding similar people and expanding their networks more likely their networks more likely

Respondents concerned about Respondents concerned about unwanted audiences less likely unwanted audiences less likely

Nickname exp(b) Female 0.527 Black 2.810* Hisp/Other 0.951 Asian 1.194 Someone else 0.929 Me 1.081 Surveillance 1.021 Far friends 0.920 Near friends 1.020 Find potential 1.386 Find similar 0.798 Profile concern 1.196 Years on FB 1.101

_cons 0.031* N 383 ll -103.802

* p<0.05, ** p<0.01, *** p<0.001

Logistic Regression Modeling Odds of Using a Nickname of Facebook

The Previous Table Says: The Previous Table Says: For NicknamesFor Nicknames

African-Americans almost three times the African-Americans almost three times the odds of using a nickname on Facebookodds of using a nickname on Facebook

Descriptive: in 2006-7, about 94% used Descriptive: in 2006-7, about 94% used nicknames, in 2010 December about nicknames, in 2010 December about 90%90%

Also, only 3.5% use multiple profiles on Also, only 3.5% use multiple profiles on FacebookFacebook

Not Shown:Not Shown:

Interaction with “near friends” has Interaction with “near friends” has higher association with untagging and higher association with untagging and deleting information (suggests deleting information (suggests strategic action)strategic action)

Disclosure levels have gone down from Disclosure levels have gone down from 06-07 but are still relatively high06-07 but are still relatively high

Romantic status disclosure also linked Romantic status disclosure also linked to privacy related issues (esp. to privacy related issues (esp. unwanted audiences)unwanted audiences)

ConclusionConclusionFacebook remains a dominant platform Facebook remains a dominant platform

and a strong social norm in Collegeand a strong social norm in College

Young adults in college are adapting and Young adults in college are adapting and changing their behaviors on Facebookchanging their behaviors on Facebook

Privacy related behaviors are associated Privacy related behaviors are associated with pro-active (concerns) rather than with pro-active (concerns) rather than re-active (personal incidents or re-active (personal incidents or observations) considerationsobservations) considerations

Privacy/disclosure actions have a strong Privacy/disclosure actions have a strong strategic component.strategic component.

Instead: Strategic disclosureInstead: Strategic disclosure

Instead: Cost/benefit analysisInstead: Cost/benefit analysis

Instead: Impression Instead: Impression ManagementManagement

Instead: Adaptation and Instead: Adaptation and ChangeChange

Thank you! Questions?Thank you! Questions?

By Zeynep TufekciAssistant Professor

University of North Carolina, Chapel Hill

Twitter: @techsoc

zeynep@unc.edu