Emerging Research Agendas at the
Intersection of Communication
& Computational Social Science
DR. KATY PEARCE, UNIVERSITY OF WASHINGTON
@KATYPEARCE
Drawing constellations
amongst the stars:
Opportunities and
challenges of studying
vulnerable populations
via computational social
science
1
SO WHAT?
• We all have to deal
with ethical issues in
computational social
science
• Looking at extreme
cases of vulnerable
populations can
illuminate issues
• I’m suggesting this is
helpful for those with
contextual knowledge
2
RESEARCH IN AUTHORITARIAN
STATES IS NOT FOR THE MEEK
AUTHORITARIAN STATES
ACCESS
RISK
CREDIBILI
TY
BLURRED
LINES
3
ENTER THE INTERNET
“Social networking sites are
a social scientist’s wet
dream” – Halavais, 2011
4
WHAT CAN THE INTERNET DO FOR
YOU?
AFFORDANCES
ACCESS
RISK
CREDIBILI
TY
BLURRED
LINES
5
INTERNET &
AUTHORITARIAN STATES
ACCESS
RISK
CREDIBILI
TY
BLURRED
LINES
6
Computational methods can allow to
see beyond the visible spectrum of
traditional analysis (Cioffi-Revilla,
2010) – this is rich, detailed,
interrelated, timely, and low-cost
data (Kitchin, 2013) that can reveal
patterns of individual and group
behavior (Lazer, et al., 2009)
ENTER
COMPUTATIONAL METHODS
7
ACCESS
• Access is easy
• But perhaps
unethical
• Informed consent
(boyd & Crawford,
2012; Gleibs, 2014)
RISK
• Data collection
reduces risk greatly
• Analysis
introduces entirely
new risk (Oboler,
Welsh, & Cruz,
2012)
BLURRED
LINES
• The issue remains:
am I friend or foe
or merely a
researcher?
CREDIBILITY
• Computational
analysis may seem
“creepy”
COMPUTATIONAL &
AUTHORITARIAN STATES
• Hard to access participants
• Everyone is at risk
• Credibility and blurred lines
are always an issue
• Computational methods
give me something and
complements other
methods
• Less risky to access
• Analysis is probably putting
people at greater risk
AT STAKE
CHALLENGES OPPORTUNITIES
WHAT HAVE I DONE AND WHAT SHOULD I DO?
78
• Busting “bad guys” feels good
• Oppositionists ask me to do
these analyses to determine if
campaigns were successful, to
identify powerful nodes, and
to identify new allies
• Embassy staff find analyses
useful
• If oppositionists and embassy
staff find analyses useful, so
does the regime
• Regime probably knows some
of this, but not everything
– Example: after analyses
where I determined a
young photographer to be
the most powerful social
media player, increased
threats to him and regular
blocking of his accounts
• Retaliatory society
WHAT TO DO?
HELPFUL HARMFUL
1
7
TAKEAWAYS
The Internet and especially computational methods solve many of
the issues related to research in authoritarian states
But it creates many new problems
Given the challenges related to privacy and consent in
computational social science, how should we be thinking about
vulnerable populations?
1
8
THANK YOU
1
9
EXTRA SLIDES
1
9
1
RESEARCH IN AUTHORITARIAN
STATES IS NOT FOR THE MEEK
AUTHORITARIAN STATES
ACCESS
•Researchers face access
barriers – getting visas,
funding to do fieldwork,
language and culture
differences, finding willing
participants, building
rapport
RISK
• Participants can take on
great risk to speak to a
researcher
• IRB cares more about
participants than researcher
• Researcher can easily risk
denial of visa or slanderous
content
BLURRED
LINES
• Friendships or relations in
an authoritarian state
increase risk for all involved
• When does friendship
begin and research end?
CREDIBILITY
• Make or break
• Very difficult to establish
• Researchers are assumed to
be spies
• Trust and credibility can be
helped by demonstrating
specific knowledge
• Using connections and
dropping names can help
1
INTERNET AFFORDANCES
ACCESS
• Easier to find individuals,
easier to find time to
conduct interviews
• Allows for participants to
have greater reflection than
in a face-to-face interview
RISK
• Digital traces can increase
likelihood of information
getting out
• Research in a digital space
may not be considered as
authentic
BLURRED
LINES
• Online contexts and social
media blur lines between
friend and researcher
• If a research subject is a
“friend” and posts things
online, where is the
informed consent?
CREDIBILITY
• Difficult to demonstrate
authenticity in a mediated
environment
• Rapport can be build online
in a different way
• Self-presentation online can
be carefully crafted
• Researchers can make
credentials available online
WHAT CAN THE INTERNET DO FOR
YOU?

Ica cs spanelv2

  • 1.
    Emerging Research Agendasat the Intersection of Communication & Computational Social Science DR. KATY PEARCE, UNIVERSITY OF WASHINGTON @KATYPEARCE Drawing constellations amongst the stars: Opportunities and challenges of studying vulnerable populations via computational social science
  • 2.
    1 SO WHAT? • Weall have to deal with ethical issues in computational social science • Looking at extreme cases of vulnerable populations can illuminate issues • I’m suggesting this is helpful for those with contextual knowledge
  • 3.
    2 RESEARCH IN AUTHORITARIAN STATESIS NOT FOR THE MEEK AUTHORITARIAN STATES ACCESS RISK CREDIBILI TY BLURRED LINES
  • 4.
    3 ENTER THE INTERNET “Socialnetworking sites are a social scientist’s wet dream” – Halavais, 2011
  • 5.
    4 WHAT CAN THEINTERNET DO FOR YOU? AFFORDANCES ACCESS RISK CREDIBILI TY BLURRED LINES
  • 6.
  • 7.
    6 Computational methods canallow to see beyond the visible spectrum of traditional analysis (Cioffi-Revilla, 2010) – this is rich, detailed, interrelated, timely, and low-cost data (Kitchin, 2013) that can reveal patterns of individual and group behavior (Lazer, et al., 2009) ENTER COMPUTATIONAL METHODS
  • 8.
    7 ACCESS • Access iseasy • But perhaps unethical • Informed consent (boyd & Crawford, 2012; Gleibs, 2014) RISK • Data collection reduces risk greatly • Analysis introduces entirely new risk (Oboler, Welsh, & Cruz, 2012) BLURRED LINES • The issue remains: am I friend or foe or merely a researcher? CREDIBILITY • Computational analysis may seem “creepy” COMPUTATIONAL & AUTHORITARIAN STATES
  • 9.
    • Hard toaccess participants • Everyone is at risk • Credibility and blurred lines are always an issue • Computational methods give me something and complements other methods • Less risky to access • Analysis is probably putting people at greater risk AT STAKE CHALLENGES OPPORTUNITIES WHAT HAVE I DONE AND WHAT SHOULD I DO? 78
  • 18.
    • Busting “badguys” feels good • Oppositionists ask me to do these analyses to determine if campaigns were successful, to identify powerful nodes, and to identify new allies • Embassy staff find analyses useful • If oppositionists and embassy staff find analyses useful, so does the regime • Regime probably knows some of this, but not everything – Example: after analyses where I determined a young photographer to be the most powerful social media player, increased threats to him and regular blocking of his accounts • Retaliatory society WHAT TO DO? HELPFUL HARMFUL 1 7
  • 19.
    TAKEAWAYS The Internet andespecially computational methods solve many of the issues related to research in authoritarian states But it creates many new problems Given the challenges related to privacy and consent in computational social science, how should we be thinking about vulnerable populations? 1 8
  • 20.
  • 21.
  • 22.
    1 RESEARCH IN AUTHORITARIAN STATESIS NOT FOR THE MEEK AUTHORITARIAN STATES ACCESS •Researchers face access barriers – getting visas, funding to do fieldwork, language and culture differences, finding willing participants, building rapport RISK • Participants can take on great risk to speak to a researcher • IRB cares more about participants than researcher • Researcher can easily risk denial of visa or slanderous content BLURRED LINES • Friendships or relations in an authoritarian state increase risk for all involved • When does friendship begin and research end? CREDIBILITY • Make or break • Very difficult to establish • Researchers are assumed to be spies • Trust and credibility can be helped by demonstrating specific knowledge • Using connections and dropping names can help
  • 23.
    1 INTERNET AFFORDANCES ACCESS • Easierto find individuals, easier to find time to conduct interviews • Allows for participants to have greater reflection than in a face-to-face interview RISK • Digital traces can increase likelihood of information getting out • Research in a digital space may not be considered as authentic BLURRED LINES • Online contexts and social media blur lines between friend and researcher • If a research subject is a “friend” and posts things online, where is the informed consent? CREDIBILITY • Difficult to demonstrate authenticity in a mediated environment • Rapport can be build online in a different way • Self-presentation online can be carefully crafted • Researchers can make credentials available online WHAT CAN THE INTERNET DO FOR YOU?

Editor's Notes

  • #4 Access Researchers face access barriers – getting visas, funding to do fieldwork, language and culture differences, finding willing participants, building rapport Risk Participants can take on great risk to speak to a researcher IRB cares more about participants than researcher Researcher can easily risk denial of visa or slanderous content Credibility Make or break Very difficult to establish Researchers are assumed to be spies Trust and credibility can be helped by demonstrating specific knowledge Using connections and dropping names can help Blurred lines Friendships or relations in an authoritarian state increase risk for all involved When does friendship begin and research end?
  • #6 Access Easier to find individuals, easier to find time to conduct interviews Allows for participants to have greater reflection than in a face-to-face interview Risk Digital traces can increase likelihood of information getting out Research in a digital space may not be considered as authentic Credibility Difficult to demonstrate authenticity in a mediated environment Rapport can be build online in a different way Self-presentation online can be carefully crafted Researchers can make credentials available online Blurred Lines Online contexts and social media blur lines between friend and researcher If a research subject is a “friend” and posts things online, where is the informed consent?
  • #7 Access Access issues are aided by the Internet A question of representativeness arises Risk Digital-based research may reduce risks for researchers Outcomes of angering governments do not differ Digital may make risks to participants seem less, but may increase risk because of digital records Digital research is safer, but not completely safe Credibility Have trusted connections introduce researcher online Let Facebook be your character witness Thus, authoritarian + digital can make for an easier path to trust building, although it also presents new challenges Blurred Lines Should a researcher have a disclaimer or consent agreement when accepting friend requests? How does the digital space differ from face-to-face?
  • #8 Computational methods can allow to see beyond the visible spectrum of traditional analysis (Cioffi-Revilla, 2010) – this is rich, detailed, interrelated, timely, and low-cost data (Kitchin, 2013) that can reveal patterns of individual and group behavior (Lazer, et al., 2009). These are patterns that would not be seen without computational tools as well as interpretation that is build on a theoretical base and contextual scholarly and applied knowledge. Context matters (boyd & Crawford, 2012) The ethics of the collection and analysis of social media data are up for debate (boyd & Crawford, 2012). Gleibs (2014) argues that social media data is not public and requires informed consent – social media users unwittingly become “human subjects” when their data is used in computational analysis and those users had an expectation of privacy, even if technically all this information is “public”. Researchers are not in the imagined audience (boyd & Crawford, 2012). Can these patterns, analysis, and interpretation put individuals at risk? Oboler, Welsh, and Cruz (2012) suggest yes. When those being analyzed are living in an authoritarian regime, and thus under greater risk already, this can be even worse. For me, as a researcher and as someone who cares about democracy in my fieldsite, I am conflicted. I know that some oppositionists have found my analyses useful. For example, being able to say that a hashtag campaign was “successful” because they saw the reach in my analysis can help a movement. My identifying government zombie social media accounts may raise morale amongst oppositionists. Embassy employees often tell me how much they appreciate these analyses as it gives them a new perspective on a situation – especially a protest or a campaign. But if oppositionists and embassy staff are reading my blog analyses, the regime itself probably is too.
  • #19 Can these patterns, analysis, and interpretation put individuals at risk? Oboler, Welsh, and Cruz (2012) suggest yes. When those being analyzed are living in an authoritarian regime, and thus under greater risk already, this can be even worse. For me, as a researcher and as someone who cares about democracy in my fieldsite, I am conflicted. I know that some oppositionists have found my analyses useful. For example, being able to say that a hashtag campaign was “successful” because they saw the reach in my analysis can help a movement. My identifying government zombie social media accounts may raise morale amongst oppositionists. Embassy employees often tell me how much they appreciate these analyses as it gives them a new perspective on a situation – especially a protest or a campaign. But if oppositionists and embassy staff are reading my blog analyses, the regime itself probably is too.
  • #23 Access Researchers face access barriers – getting visas, funding to do fieldwork, language and culture differences, finding willing participants, building rapport Risk Participants can take on great risk to speak to a researcher IRB cares more about participants than researcher Researcher can easily risk denial of visa or slanderous content Credibility Make or break Very difficult to establish Researchers are assumed to be spies Trust and credibility can be helped by demonstrating specific knowledge Using connections and dropping names can help Blurred lines Friendships or relations in an authoritarian state increase risk for all involved When does friendship begin and research end?
  • #24 Maybe trim down or cut all