Skip to main content

Table 7 Effects of locational heterogeneity of training on information spillover on non-training participants

From: Does training location matter? Evidence from a randomized field experiment in Rural Indonesia

 (1)(2)(3)(4)(5)(6)
VariablesKnowing conservation techniqueKnowing grafting methodsAdopting conservation technique if knowing = 1Adopting grafting methods if knowing = 1Knowing conservation and grafting techniquesAdopting conservation and grafting techniques if knowing both = 1
Numbers of network who went to inter-island training * Post 20130.05050.0982*− 0.0787− 0.01760.139**− 0.123*
(0.0573)(0.0557)(0.0609)(0.0508)(0.0656)(0.0686)
Numbers of network who went to intra-island training * Post 20130.04340.0718− 0.05050.06400.09640.0114
(0.0728)(0.0712)(0.0801)(0.0658)(0.0840)(0.0892)
Numbers of network who went to training * Post 2013− 0.0279− 0.05850.0512− 0.0157− 0.06450.0647
(0.0406)(0.0396)(0.0441)(0.0366)(0.0467)(0.0500)
Numbers of network who went to inter-island training− 0.0168− 0.07550.03190.00810− 0.08980.0586
(0.0528)(0.0512)(0.0557)(0.0461)(0.0607)(0.0626)
Numbers of network who went to intra-island training− 0.0580− 0.05650.00949− 0.00949− 0.104− 0.0460
(0.0592)(0.0580)(0.0642)(0.0526)(0.0686)(0.0699)
Numbers of network who went to training0.03190.0712**− 0.008710.01280.0778*− 0.0159
(0.0371)(0.0362)(0.0395)(0.0328)(0.0428)(0.0446)
Year of 20130.04820.0240− 0.00745− 0.03520.0428− 0.0409
(0.0377)(0.0369)(0.0411)(0.0346)(0.0433)(0.0471)
Year of 20140.006500.03480.01020.03690.02140.00827
(0.0362)(0.0355)(0.0401)(0.0333)(0.0416)(0.0459)
Constant0.822***0.903***0.852***0.762***0.807***0.708***
(0.0783)(0.0751)(0.0874)(0.0719)(0.0928)(0.101)
Observations537543476491543464
Ethnicity fixed-effectsYesYesYesYesYesYes
Farmers group fixed-effectsYesYesYesYesYesYes
F test of null R20.0008630,000000.000000.000000.000000.00000
P value of Hausman test0.81960.95590.08960.39150.82370.0102
R squared0.1110.1470.1820.1510.1470.198
  1. Estimation is based on LATE random-effects instrumental variable models. Standard errors are in parentheses. ***, **, and * signify statistical significance at the 1%, 5% and 10% level, respectively. All training dummy is instrumented by all invitation result. No of information sources who are training participants are instrumented with no of information sources who are selected to participate according to invitation result