Increasing adoption of conservation agriculture technologies: A framed field experiment in Ghana
How to cite
Ambler, Kate; Brauw, Alan de; Murphy, Michael. 2019. "Increasing adoption of conservation agriculture technologies: A framed field experiment in Ghana." AEA RCT Registry. AEARCTR-0003973. https://doi.org/10.1257/rct.3973-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
This study uses a framed field experiment to study ways to increase the adoption of conservation agriculture (CA) technologies in Ghana. The fact that CA practices often result in production benefits only after several consecutive seasons of adoption creates a two-fold challenge. First, it is difficult to convince farmers to invest over multiple seasons without a payoff, and second, evaluation of any strategies is complicated by the four season or more time frame that is needed. Our study addresses the first issue by studying both multi-season incentives in an environment of uncertainty over when payoffs will occur and information on the adoption decisions of peers. To address the second issue, we conduct a framed field experiment, allowing us to simulate many different seasons of implementation in a single session. The results can be used to inform randomized impact evaluations of similar programs.
Intervention
The framed field experiment will study ways to encourage the adoption of conservation agriculture technologies. The two interventions to be tested are incentives for adoption and peer information. We will also study the interaction of incentive timing with uncertainty around when production benefits from the practice adoption will kick in. Because conservation agriculture practices require adoption in multiple seasons before gains are realized, the framed field experiment allows us to simulate the long time horizon and understand the potential of these interventions without conducting a years-long study. See the project documents for more information.
Experimental design
This framed field experiment contains three separate randomizations. More detail is available in the project description.
Randomization 1: Adoption incentive
The incentive treatment will consist of one treatment arm and a control group, with 2/3 of the sample in the treatment group and 1/3 in the control group. The treatment will consist of an incentive for farmers who choose to adopt CA practices in any of the first four rounds of the experiment. The bonus will be paid in the relevant round following the confirmation of the adoption decision for that round by the farmer, in the first four rounds of the experiment only.
Randomization 2: Uncertainty regarding production payoff
An additional element of the experiment models the uncertainty regarding when farmers can expect to see a production benefit from the application of CA techniques. All farmers will be told that they can expect to see benefits in 4 to 6 seasons, and incentives will last for four seasons. Farmers will be randomly allocated to a group that receives benefits in 4 seasons, 5 seasons, or 6 seasons (across both the treatment and control groups).
Randomization 3: Peer information
The peer information treatment will be cross-randomized with the bonus treatment evenly across one treatment and control group. Treated participants will receive an additional prompt about the adoption behavior of a peer-farmer in the previous round, prior to making their decision for the current round. This information will be randomized by round and farmer, so each farmer will receive different information in every round. The information will include the adoption choice of the peer farmer in that round and the prior adoption history (in how many consecutive rounds if any the peer farmer had adopted CA practices). The information will also include the returns experienced by the peer farmer based on their adoption choice, history, and that season’s rainfall. The rainfall experienced by the peer farmer will always be the same as the rainfall experienced by the participant.
Design details
Cited by
No citation data available for this registration yet.