I test a theory of spillover formation based on behavioral biases related to the interpretation of RCTs awareness and assignment methods.
Experimental design
I will provide free access to a soil quality test similar to the Soil Health Card for the treatment group within a village. The variation is between villages: I will randomize villages into 1) ambiguous treatment assignment vs visual treatment assignment; 2) high percentage of eligible population into treatment group vs low percentage of eligible population into treatment group; 3) pure control arm: villages where no farmer will be targeted; 4) pure treatment arm: villages where all the farmers are targeted.
If resources allow: 5) intervention introduced by local organization vs outsider organization;
Design details
Randomization method
Between villages, randomization will be done in the office by a computer (Stata), making sure villages that are less than 5km away from each other are clustered inside the same treatment arm; within villages, randomization will be done by a lottery (visual vs non-visual). The visual lottery consists of removing balls from a bag.
Randomization unit
First level of randomization: village; second level of randomization: individual
Clusters
120 villages
Observations
4,000 farmers
Arms
35 villages into visual treatment assignment, 35 villages into non-visual treatment assignment
25 villages pure control
25 villages pure treatment
Primary outcomes
Demand for information (willingness to pay, contact for extension services' programs, contact for government programs), changes in links (networks), ability perceptions, trust, crop yields, technology adoption, agricultural knowledge, locus of control, life satisfaction, religiosity, confidence, productivity, income, attrition
Cited by
No citation data available for this registration yet.