computer-supported cooperative work and social computing - … · computer-supported cooperative...

Computer-supported Cooperative Work and Social ComputingCSE510 Guest Lecture

Benjamin Mako Hillmakohill@uw.edu

University of WashingtonDepartment of Communication

Assistant Professor

Harvard UniversityBerkman Center for Internet and Society

Faculty Affiliate

February 23, 2016

I. Mapping Computer-supportedCooperative Work

2 / 36

Individual

Small Group

Project

Organization

PC Applications

Networked PCs

Minis, networks,GDSS

Mainframe systems

MIT/IS CSCW HCI

Grudin (1994a, b) showing “development and research contexts” in the academicstudy of computer use in computer science. On the left side are the sub-fields orresearch streams in computer science. On the top are the types of user beingserved. On the bottom are the types of products being produced.

3 / 36

[Johansen (1988); Baecker (1995); image from Wikimedia Commons]4 / 36

Typologies of TasksQuadrant IGenerate

Quadrant IIChoose

Quadrant IIINegotiate

Quadrant IVExecute

Generating Ideas Generating Plans

ExecutingPerformance

Tasks

ResolvingConflictsof Power

Resolving Conflictsof Interest

Resolving Conflictsof Viewpoint

Deciding Issuesw/ No Right Answer

Solving Problemsw/ Correct Answers

Type 2:Creativity tasks Ty

pe

1: Pl

annin

g t

asks

Type 3: Interactive tasks

Type 4: Decisio

n-making ta

sks

Take

5: C

ogni

tive

con

flict

tas

ks Type 6: Mixed-m

otive tasks

Type 7: Contests/battles

Type 8: Performances

Conceptual Behavioral

Conflic

tC

oopera

tion

[McGrath 1984]5 / 36

II. Classic Approaches to CSCW

6 / 36

Syllabus from MIT 16.499 (Circa 2005)

… Group Interaction Theory: Theories in Verbal Communication & Non-VerbalCommunication

… Group Interaction Theory: Distributed Cognition… Group Interaction Theory: Activity Theory… Methodologies for Studying Groups & CSCW Technologies: Intro and Quantitative

Approaches… Methodologies for Studying Groups & CSCW Technologies: Qualitative Approaches… Techniques for Modeling Group Interactions… Awareness in Collaboration: Intro & Workspace Awareness… Awareness in Collaboration: Team Situation Awareness… Design Considerations for CSCW Technologies Computer Support for Co-located

Collaboration Computer Support for Distributed Collaboration

7 / 36

Distributed Cognition

[Hutchins (1990): Technology of Team Navigation]8 / 36

Activity Theory

[e.g., Nardi 1995: Context and Consciousness: Activity Theory and Human-Computer Interaction]9 / 36

Awareness (Synchronous)

[Dourish and Bellotti (1992)]10 / 36

Awareness (Asynchronous)

[Hill et al. CHI’92: “EditWear and Readwear”]11 / 36

Coordination

A schematic illustration of the roles and information flows in software testing in the S4000 project.The flows in the diagram indicate the intended flow according to the bug handling protocol.

[Schmidt and Simone (1996); Malone and Crowston (CSCW’92)]12 / 36

Organization and Social Structure

[Orikowski 1992: “Learning from Notes”]13 / 36

II. Social Computingand Peer Production

14 / 36

Peer Production

New modes of collective productionmade possible by lowered transactioncosts through new communicationtechnologies. (Benkler 2003, 2006)

15 / 36

Peer Production?

Individual

Small Group

Project

Organization

PC Applications

Networked PCs

Minis, networks,GDSS

Mainframe systems

MIT/IS CSCW HCI

It’s not particularly obvious where peer production would fit. It’s certainly notobvious that it fit within traditional CSCW spaces.

16 / 36

Peer Production in CSCW

.2.2

5.3

.35

.4.4

5P

redi

cted

Cha

nge

in Q

ualit

y

0 2 4 6 8Number of editors (log2)

High editor concentration Low editor concentration

By number of editors and editor concentrationPredicted change in quality

Figure 4. Joint influence of number and concentration of editors on changes in quality.

.25

.3.3

5.4

.45

Pre

dict

ed C

hang

e in

Qua

lity

0 2 4 6 8Number of editors (log2)

High editor communication Low editor communication

By number of editors and amount of communicationPredicted change in quality

Figure 7. Joint influence of number of editors and communication on changes in quality.

-.20

.2.4

.6P

redi

cted

Cha

nge

in Q

ualit

y

1 2 3 4 5 6Initial quality

High editor concentration Low editor concentration

By inital qualityPredicted change in quality

Figure 6. Joint influence of initial quality and concentration of editors on changes in quality.

.25

.3.3

5.4

.45

.5P

redi

cted

Cha

nge

in Q

ualit

y

0 20 40 60 80Months since start of article

High editor concentration Low editor concentration

By months since start of articlePredicted change in quality

Figure 5. Joint influence of article age and concentration of editors on changes in quality.

[Kittur and Kraut CSCW’2008]17 / 36

[Viégas et al. CHI’2004: HistoryFlow]18 / 36

IV. My Peer ProductionResearch

19 / 36

Almost Wikipedia

CitationHill, Benjamin Mako. (2013) “Almost Wikipedia: Eight Early EncyclopediaProjects and the Mechanisms of Collective Action.” In Essays on Volunteer Mo-

bilization in Peer Production. Doctoral Dissertation. Massachusetts Institute ofTechnology.

20 / 36

Why Wikipedia? Instead of...

21 / 36

Synthesis

Innovativeness of Goal/ProductFamiliar Novel

Inno

vativ

enes

s of

Pro

cess

/Too

lsN

ovel

Fam

iliar

Traditional products using traditional methods and tools.

"Like Encylopedia Britannica — just online and free."

New products using traditional methods and tools.

"A new type of encyclopedia, but produced like the old ones."

Traditional products using new methods and tools.

"Like Encyclopedia Britannica, but produced in a radically new way."

New products using novel methods and tools.

"A new type of encyclopedia produced in a radically new way."

22 / 36

The Remixing Dilemma

CitationHill, Benjamin Mako, Andrés Monroy-Hernández. “The Remixing Dilemma: TheTrade-off between generativity and originality.” Published in American Behavioral

Scientist, 2013.23 / 36

Remixing

The reworking and recombination of existing creative artifacts.Most commonly in reference to music, video, and interactive media.

… Widespread, and an important new communication modality (e.g.,

Manovich 2005; Lessig 2009)

… Especially among use youth (Jenkins 2006; Palfrey and Gasser 2008)

24 / 36

Research Questions

… What qualities of Scratch projects and theircreators are associated with more generativeprojects?

… What qualities are associated with more originalremixing? (e.g., Keen 2007; Lanier 2010)

25 / 36

(Resnick et al. 2009)

Results (RQ1)Testing Theories of Generativity

Ceteris paribus (including exposure)...

1A) After a threshold is reached, simpler projects aremore generative, because they are more likely to beincomplete and to invite elaboration.

… “Release early, release often” (Raymond 1999)

… Principle of procrastination (Zittrain 2008)

1B) Remixing relies on common reference points makingthe work of more prominent creators more generative.

… (Sinnreich 2010; Cheliotis and Yew, 2009)

1C) Remixing involves elaboration and iteration makingworks that are remixes themselves more generative thande novo projects.

… (Murray and O’Mahoney 2007; Cheliotis and Yew, 2009)

Complexity

Gen

erat

ivity

Hypothesis 1A

Results (RQ1)Testing Theories of Generativity

Ceteris paribus (including exposure)...

1A) After a threshold is reached, simpler projects aremore generative, because they are more likely to beincomplete and to invite elaboration.

… “Release early, release often” (Raymond 1999)

… Principle of procrastination (Zittrain 2008)

1B) Remixing relies on common reference points makingthe work of more prominent creators more generative.

… (Sinnreich 2010; Cheliotis and Yew, 2009)

1C) Remixing involves elaboration and iteration makingworks that are remixes themselves more generative thande novo projects.

… (Murray and O’Mahoney 2007; Cheliotis and Yew, 2009)

Complexity

Gen

erat

ivity

Hypothesis 1A

?

Results (RQ1)Testing Theories of Generativity

Ceteris paribus (including exposure)...

1A) After a threshold is reached, simpler projects aremore generative, because they are more likely to beincomplete and to invite elaboration.

… “Release early, release often” (Raymond 1999)

… Principle of procrastination (Zittrain 2008)

1B) Remixing relies on common reference points makingthe work of more prominent creators more generative.

… (Sinnreich 2010; Cheliotis and Yew, 2009)

1C) Remixing involves elaboration and iteration makingworks that are remixes themselves more generative thande novo projects.

… (Murray and O’Mahoney 2007; Cheliotis and Yew, 2009)

Complexity

Gen

erat

ivity

Hypothesis 1A

?

Creator Prominence

Gen

erat

ivity

Hypothesis 1B

Results (RQ1)Testing Theories of Generativity

Ceteris paribus (including exposure)...

1A) After a threshold is reached, simpler projects aremore generative, because they are more likely to beincomplete and to invite elaboration.

… “Release early, release often” (Raymond 1999)

… Principle of procrastination (Zittrain 2008)

1B) Remixing relies on common reference points makingthe work of more prominent creators more generative.

… (Sinnreich 2010; Cheliotis and Yew, 2009)

1C) Remixing involves elaboration and iteration makingworks that are remixes themselves more generative thande novo projects.

… (Murray and O’Mahoney 2007; Cheliotis and Yew, 2009)

Complexity

Gen

erat

ivity

Hypothesis 1A

?

Creator Prominence

Gen

erat

ivity

Hypothesis 1B

Results (RQ1)Testing Theories of Generativity

Ceteris paribus (including exposure)...

1A) After a threshold is reached, simpler projects aremore generative, because they are more likely to beincomplete and to invite elaboration.

… “Release early, release often” (Raymond 1999)

… Principle of procrastination (Zittrain 2008)

1B) Remixing relies on common reference points makingthe work of more prominent creators more generative.

… (Sinnreich 2010; Cheliotis and Yew, 2009)

1C) Remixing involves elaboration and iteration makingworks that are remixes themselves more generative thande novo projects.

… (Murray and O’Mahoney 2007; Cheliotis and Yew, 2009)

Complexity

Gen

erat

ivity

Hypothesis 1A

?

Creator Prominence

Gen

erat

ivity

Hypothesis 1B

Cumulativeness

Gen

erat

ivity

Hypothesis 1C

Results (RQ1)Testing Theories of Generativity

Ceteris paribus (including exposure)...

1A) After a threshold is reached, simpler projects aremore generative, because they are more likely to beincomplete and to invite elaboration.

… “Release early, release often” (Raymond 1999)

… Principle of procrastination (Zittrain 2008)

1B) Remixing relies on common reference points makingthe work of more prominent creators more generative.

… (Sinnreich 2010; Cheliotis and Yew, 2009)

1C) Remixing involves elaboration and iteration makingworks that are remixes themselves more generative thande novo projects.

… (Murray and O’Mahoney 2007; Cheliotis and Yew, 2009)

Complexity

Gen

erat

ivity

Hypothesis 1A

?

Creator Prominence

Gen

erat

ivity

Hypothesis 1B

Cumulativeness

Gen

erat

ivity

Hypothesis 1C

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008… 2B: Sinnreich 2010… 2C: Cheliotis and Yew, 2009

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008

… 2B: Sinnreich 2010… 2C: Cheliotis and Yew, 2009

Complexity

Orig

inal

ity

Hypothesis 2A

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008

… 2B: Sinnreich 2010… 2C: Cheliotis and Yew, 2009

Complexity

Orig

inal

ity

Hypothesis 2A

?

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008… 2B: Sinnreich 2010

… 2C: Cheliotis and Yew, 2009

Complexity

Orig

inal

ity

Hypothesis 2A

?

Creator Prominance

Orig

inal

ity

Hypothesis 2B

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008… 2B: Sinnreich 2010

… 2C: Cheliotis and Yew, 2009

Complexity

Orig

inal

ity

Hypothesis 2A

?

Creator Prominance

Orig

inal

ity

Hypothesis 2B

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008… 2B: Sinnreich 2010… 2C: Cheliotis and Yew, 2009

Complexity

Orig

inal

ity

Hypothesis 2A

?

Creator Prominance

Orig

inal

ity

Hypothesis 2B

Cumulativeness

Orig

inal

ity

Hypothesis 2C

Results (RQ2)Testing Theories of Originality

But we also care about the originality of resultingremixes. (Keen 2007; Lanier 2010)

2A-C) The Remixing Dilemma:

Attracting more remixers will result in less

skilled, and/or less motivated, remixers who

will, ceteris paribus, remix projects less

originally.

… 2A: Zittrain 2008… 2B: Sinnreich 2010… 2C: Cheliotis and Yew, 2009

Complexity

Orig

inal

ity

Hypothesis 2A

?

Creator Prominance

Orig

inal

ity

Hypothesis 2B

Cumulativeness

Orig

inal

ity

Hypothesis 2C

Takeaways

Projects are more likely to be remixed when they are:

… Moderately complicated… Created by prominent creators… Cumulative

But, there is a tradeoff in that each of these factors isalso associated with less original forms of remixingbehavior.

Promoting complexity seems like the best option.

31 / 36

Takeaways

Projects are more likely to be remixed when they are:

… Moderately complicated… Created by prominent creators… Cumulative

But, there is a tradeoff in that each of these factors isalso associated with less original forms of remixingbehavior.

Promoting complexity seems like the best option.

31 / 36

Laboratories of Oligrachy

CitationShaw, Aaron, Benjamin Mako Hill. “Laboratories of Oligarchy? How The IronLaw Extends to Peer Production.” Journal of Communication 64, no. 2 (April2014): 215–38.

32 / 36

Peer Production as Participatory Democracy?

Peer production projects have been citedfunction as a novel form of participatoryorganization...

… ... with a broad democratizing potentialinspiring waves of social movement activistsand theorists. (e.g., Benkler, 2006; Castells, 1996;

Fuster Morell, 2012; Hess and Ostrom, 2011; Wilson

and Tufekci, 2012)

… ... and a model of leaderless organizatione.g., (Shirkey 2008; Konieczny, 2009)

33 / 36

Robert Michels’ “Iron Law”

“He who says organization saysoligarchy.”

As organizations increase in size andcomplexity, they have a tendency to developoligarchy leadership that pursuesconservative goals consistent withorganizational maintenance.

… Consolidation of power among elites… Transformation of goals as elite interests

diverge from members

(Michels 1915; Lipset et al.

1956 Leach, 2005; Voss and

Sherman, 2000)

34 / 36

Robert Michels’ “Iron Law”

“He who says organization saysoligarchy.”

As organizations increase in size andcomplexity, they have a tendency to developoligarchy leadership that pursuesconservative goals consistent withorganizational maintenance.

… Consolidation of power among elites

… Transformation of goals as elite interestsdiverge from members

(Michels 1915; Lipset et al.

1956 Leach, 2005; Voss and

Sherman, 2000)

34 / 36

Robert Michels’ “Iron Law”

“He who says organization saysoligarchy.”

As organizations increase in size andcomplexity, they have a tendency to developoligarchy leadership that pursuesconservative goals consistent withorganizational maintenance.

… Consolidation of power among elites… Transformation of goals as elite interests

diverge from members

(Michels 1915; Lipset et al.

1956 Leach, 2005; Voss and

Sherman, 2000)

34 / 36

Robert Michels’ “Iron Law”

“He who says organization saysoligarchy.”

As organizations increase in size andcomplexity, they have a tendency to developoligarchy leadership that pursuesconservative goals consistent withorganizational maintenance.

… Consolidation of power among elites… Transformation of goals as elite interests

diverge from members

(Michels 1915; Lipset et al.

1956 Leach, 2005; Voss and

Sherman, 2000)

34 / 36

35 / 36

0.004

0.008

0.012

1.2

1.3

1.1

1.2

1.3

1.4

1.5

M1: P(N

ew Adm

in)M

2: Project Edits by Admin

M3: Adm

in Reverts

0 200 400 600 800

Total Registered UsersPrototypical Plots