viralvcd
TRANSCRIPT
Nithya Sambasivan, UC , IrvineEd Cutrell, Microsoft Research IndiaKentaro Toyama, UC, Berkeley
ViralVCDTracing Information-Diffusion Paths with Low Cost Media in Developing Communities
• ICT4D projects assess need, relevance, and scope for development
– Often community-centered
• Study technology penetration /use in a community
• Understand key players and dynamics of diffusion
• Surveys and interviews limited by lack of in-situmonitoring
• Experience sampling and probes fail to scale in communities
The problem
• A low-cost, rapid data elicitation technique for low-income contexts
• Employs physical media and mobile phone questionnaires
• Gathers data that is:
– Social: networks underpinning diffusion
– Technological: ownership, access, and usage
– Developmental: assessment of baseline
• Combines probes and snowball sampling
– Contextualized in technology use
– Caters to local practices
Enter ViralVCD
• Preliminary ethnographic study in 2 slums in Bangalore, India (female domestic workers)
• Total: 7 slums in deployment
• In collaboration with NGOs
• Topics for conversation: healthcare and education
• 30 / 64 households owned a DVD player
The site
The process
Participatory videos + VCDs
Screenings Mobile call-in contest
Diffusion
• Useful and interesting content
– 13 minutes at most
– Showcase local practices to other communities
– But a lens into its own diffusion
• Observation of a subset of social networks
– Interested in “development”
Participatory videos
Conducted a cooking contest•Openly competitive
Nutritive cooking
•Local knowledge
Nutritive cooking
•Taste and nutrition as criteria
Childhood education•Role play between two non-literate women
Dissemination•Screenings, 2-3 VCDs per person
98411 21111
ப ோன்
ண்ணுங்க!
ಫೋನ್ ಮಾಡಿ!
Call-in contest•Missed calls, researcher calls back
Call-in contest•Entertainment idiom
Call-in contest•Entertainment idiom
• We determined
– Unique identifier / photo
– Socio-economic profile and baseline
– Source of VCD (person received from)
– Unique content-related question
• Correct answers
– Utilitarian incentive (bed sheets or utensils)
– Varied experimental conditions to test effects
• They then pass the VCD to someone else
– Contest ended in a week
Call-in contest
Results• Gave 131 VCDs to 64 attendees
• Call response: 31.25% (50 unique callers)
Social insightsMicro-level
• Two forms of diffusion
– Peer-to-peer (A→B→C)
– Actor-driven (A→ (B and C)).
Macro-level
• Diffusion reflected social solidarity
• Helped understand relationships in an ecology
Technological insights• Understand in-situ usage
• 50% of callers had proximate access (neighborhood access)
• All mobile, except for 4 pay phones
• Strong correlation between owing mobiles and DVD players
Developmental insights• Development extensions
• Contest at the end necessitated viewing of videos
• No difference in diffusion across income levels
Opportunities• Provided a direct interview opportunity
• Identify key players in communities
• Assess the socio-technical makeup of the setting.
• Across income levels, with appropriate incentive
• Medium technological penetration may be sufficient
• Identify and recruit peers of the same stratum, possibly across communities.
Conclusion• Avoided additional infrastructure
• Helped us understand
– Social networks
– Technological access and ownership
– Developmental baseline
• Could use other pervasive technologies
– Cassettes or notebooks
• More broadly applicable to CHI
OK ta-ta bye-bye!Please refer to the paper for a lengthier discussion of method and results.