Social Learning - Worms (Miguel & Kremer) Flashcards
Effect of worms
Infect 1/4 people WW , reduce school attendance, concentration and ability
Miguel & Kremer Research design
Deworming medicine randomly given to schools
Problems with this research design
Treating certain kids still benefits other non-treated kids since social externalities - less spread!
Key: So treatment effects can be underestimated since both seem more healthy.
How did they rollout this program
75 schools. 3 groups of 25.
Phased over 3 years: first group 1998, 2nd 1999, 3rd 2000.
Also looked at kids in non-treated schools but have links to the early treatment schools (e.g friends, family etc!) , and compare this to some with no links!
Findings after 1 year on absence
B) Cost effective?
Absence falls from 25% (control) to 16% (treated)
B) Extremely cost effective! A $0.49 pill worth $3.50 per extra school yeart!
Effect of links
More links to early treatment schools reduces uptake by 3% per link!
(Why???)
Why does links to early treated schools despite not having treatment themselves reduce uptake?
Free riding - friends are treated, so you benefit!
So how are 3 ways peers can benefit non-treated
Disease environment improved (allow them to free ride)
Peer imitation - friends take so I take
Peer provides information on use and benefits of technology
Why did linked children reduce own adoption of deworming pills? (2)
Free rider - reduction in their risk since links were treated.
Deworming itself is unpleasant!
So social returns>private returns in many health technologies.
When this happens, what does this require
Large subsidies to sustain takeup. E.g may even need to pay people to take deworming to get full desired effect and escape free riding!
