Heterogeneous quality of agricultural commercial inputs and learning through experimentation

How to cite

Macours, Karen; Laajaj, Rachid. 2014. "Heterogeneous quality of agricultural commercial inputs and learning through experimentation." AEA RCT Registry. AEARCTR-0000268. https://doi.org/10.1257/rct.268-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Karen Macours* , Rachid Laajaj

* Primary investigator

Registry ID
AEARCTR-0000268
DOI
https://doi.org/10.1257/rct.268-2.0
First registered
Last updated
Status
In development
Trial dates
2014-02-18 – 2017-12-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/268

Abstract

This study aims to analyze the learning about product quality by farmers who are encouraged to experiment with inputs. We will study whether such experimentation leads to higher adoption of products that were shown to have higher returns on a farmer’s own farm. After identifying farmers who are interested in participating in an agronomical trial in 96 villages, we will randomly select half of the villages to participate. They will be invited to experiment with several inputs and practices, under guidance of an agronomical researcher. By comparing these treatment farmers with the control farmers (the other half) we will test whether the reduction in asymmetry of information about the returns to different inputs translates into a higher take up of the good while crowding out the lower return ones. We focus on whether learning-by-doing differs with soil characteristics and farmers’ own characteristics (in particular their technical, cognitive and non-cognitive skills). The random assignment generates an experimental control group that also allows analyzing potential biases in agronomical field trials.

Intervention

Farmers will invited to participate in an agronomical trial in which they'llexperiment with several inputs and practices, under guidance of an agronomical researcher.

Experimental design

o First 96 villages were randomly selected from a map of all the villages in the division, and the geographical concentration of the villages per sub-location was also randomized.
o Among the 96 villages, 1/4 will be randomly assigned to participate in agronomical trials for 3 seasons starting in season 1; ¼ will be randomly assigned to participate in agronomical trials for 3 seasons starting in season 2; and ½ stay in the control group. Assignment of villages to treatment and control occurs through public lotteries.
o In each of the 96 villages, candidates to participate in the research trials were obtained in the following way: 5 farmers were selected by the community during a village gathering; and another 5 farmers were randomly drawn among all households in the village. The later randomization was done in office by computer.
o Within each treatment village, selected farmers are randomly assigned to participate in agronomical trials with maize (40%), with soya (30%), or soya-maize intercrop (30%). This randomization will be done in office by computer.
o Each of the participating farmers will be invited to participate in an agronomical trial on a small part of his land. This plot will be sub-divided into 6 sub-plots, which will randomly be assigned to 6 different input combinations. This randomization will be done in office by computer.

Design details

Randomization method
Village level randomization by public lottery All other randomizations done in office by computer
Randomization unit
Village-level for overall treatment. Individual for variations of treatment
Clusters
96 villages
Observations
900
Arms
1/4 for treatment first year; 1/4 for treatment second year; 1/2 control
Primary outcomes
o Adoption of inputs and practices with demonstrated high return o Heterogeneity in adoption and returns based on baseline skills and soil characteristics.

Cited by

No citation data available for this registration yet.

Keywords