|
Field
Trial Title
|
Before
Social learning to scale-up the agroecological transition
|
After
Early Adopters, Peer Leaders and the Diffusion of Agroecological Knowledge
|
|
Field
Abstract
|
Before
Peer effects are a subject of increasing attention in many areas of economic research. Peer influence can create social multiplier effects, whereby an initial investment targeting one small group can lead to larger changes, as individuals close to the target group are directly influenced by its actions. When it comes to the green transition, a central question is that of learning about green agricultural technologies. The adoption of a new technique sometimes requires specific technical assistance, which cannot easily be provided to all eligible farmers because of the high costs it would entail. In this context, observational learning can play a crucial role in the diffusion of new practices. However, there are many reasons why social learning might not ultimately happen; and it is often difficult to identify and measure it accurately. The literature provides several examples of the diffusion of agricultural innovations through social networks and peer effects in developing countries. There is, however, no prior evidence of the existence of the diffusion of agroecological practices through social learning in developed countries. We aim at filling this gap by studying the knowledge spillovers of an intervention that trains a number of volunteer farmers in new agricultural practices, with the aim of indirectly transforming those of the members of their network – we call a network of farmers a cluster. The experiment requires the implementation of a pilot program, in which the clusters are randomly assigned to training. The proportion of farmers within each cluster who are offered training varies between treatment arms. We call them the injection points of new knowledge. This design allows first to infer the causal effect of training on treated groups, then to check the presence of the diffusion of new knowledge inside the clusters, and finally to highlight a threshold in the proportion of injection points necessary for the diffusion of new knowledge.
|
After
Peer effects are a subject of increasing attention in many areas of economics research. Influence from peers can indeed generate social multiplier effects, whereby an initial investment targeting one small group can lead to larger changes, since individuals close to the target group tend to imitate them and learn from their experience. When it comes to the green transition in agriculture, the diffusion of new knowledge through peers could play a crucial role in promoting the adoption of agroecological practices. However, the conditions for successful peer learning in this context are still poorly understood. In particular, it is unknown whether the profile of the first individuals in peer groups to receive information affects its diffusion to other peers.
To answer this question, we run a randomized-controlled-trial (RCT) from a large sample of voluntary French farmers, where individuals are randomly assigned to peer groups and receive informative content for 18 months through a digital platform. Only one farmer per peer group receives the information directly, serving as the injection point of the group. Treatment varies depending on the profile of the designated injection point, which is either chosen randomly among early adopters of agroecological practices (Treatment~A) or among ordinary peers (Treatment~B). Farmers assigned to the control arm are also placed in peer groups but they receive no specific information on innovative practices (Treatment~C).
We can first verify that farmers' knowledge of agroecology can be improved by injecting informative content using a digital platform. In addition, we can measure the sharing of information between peers and test the hypothesis that receiving second-hand information can improve knowledge level. A major advantage of this protocol is that it can test the hypothesis that early adopters do better than ordinary peers, whether it is improving their own knowledge or transmitting newly acquired knowledge to their peers. Focusing on ordinary peers who receive information second-hand, who are the most representative farmers and therefore the target of public programs outside the experiment, we can further verify whether they acquire more knowledge when the information comes from an early adopter rather than from another ordinary peer.
|
|
Field
Trial Start Date
|
Before
September 03, 2023
|
After
March 15, 2024
|
|
Field
Trial End Date
|
Before
June 30, 2024
|
After
June 30, 2026
|
|
Field
Last Published
|
Before
March 30, 2023 02:55 PM
|
After
July 18, 2024 08:39 AM
|
|
Field
Intervention (Public)
|
Before
The objective of this experiment is to measure the dissemination, within farmer clusters, of new technical knowledge enabling the transition to pesticide-free agriculture. The experience is based on the implementation of a training program, which a randomly chosen proportion of farmers within each cluster are invited to follow. The treatment is the proportion of trained farmers within a cluster.
|
After
The digital platform, managed by the service provider Landfiles, has been customized for the experiment. In particular, participants do not have the possibility to share the content outside the platform, or even outside their peer group. Only the injection points receive the information specially designed for the study; they can only share it with members of their group. Members of the same peer group can exchange with each other, but not with members of other peer groups.
While all participants receive some generic content linked to environmental regulations (non-technical and not focused on agroecological practices) throughout the experiment, participants assigned to treatment arms A and B will also receive a specific content on ecological practices, which we call the intervention. The content of the intervention will take the form of portraits of innovative farms, follow-up of agronomic trials, and newsletters. It will include several agroecological topics, including companion plants, plant cover, non-tillage associated with the reduction of herbicides, cereal-vegetable association, foliar fertilization. In practice, the injection points in arms~A and~B will receive one portrait of an innovative farm every two weeks, one agronomic trial follow-up every three weeks, and one newsletter every month.
|
|
Field
Intervention Start Date
|
Before
October 02, 2023
|
After
September 01, 2024
|
|
Field
Intervention End Date
|
Before
November 30, 2023
|
After
June 30, 2026
|
|
Field
Primary Outcomes (End Points)
|
Before
The main outcomes are scores of technical knowledge related to the agroecological transition.
|
After
The primary outcomes measure participants’ access to information through user data automatically collected through the digital platform. These data includes information about user connections, viewing posts, reacting to posts, and reaction time to posts.
|
|
Field
Primary Outcomes (Explanation)
|
Before
The training provided in groups A and B aims to improve the knowledge necessary for a transition to pesticide-free agriculture. The acquisition of this knowledge by the farmers (whether they are injection points or ordinary members of the cluster) can be assessed using a knowledge test. If the injection points pass the knowledge test, the proportion of successful tests within the treatment groups will be that of the injection points (1/3 and 1/2 respectively). If the ordinary members also pass the test (through knowledge sharing), the proportion of successful tests within the treatment groups will be higher than 1/3 and 1/2. If there is a threshold effect, i.e. it takes at least a certain proportion of injection points for diffusion to take place, then the proportion of successful tests in group B will be greater than 1.5 times that in group A.
|
After
Primary outcomes that record the activity of each user on the digital platform aim at evaluating participants’ interest in the information shared through the platform. This also includes whether injection points share the informative content with their peer group or not – and if yes, at what frequency and at what speed. We expect treated groups to behave differently, with a dissemination of information that is both faster and more systematic in groups whose injection point is an early adopter of agroecology, because the latter is expected to play its role as injection point more effectively than an ordinary peer.
|
|
Field
Experimental Design (Public)
|
Before
We will run a randomized control trial (RCT) to compare the efficiency of treatments A and B in improving the technical knowledge needed to switch to pesticide-free agriculture. Each volunteer participant will be asked to sign an FPIC form specifying that they understand that their participation in the pilot program does not guarantee that they will benefit from free training at the end, as the pilot includes two phases of randomization.
Upstream of the randomization, we will form clusters of farmers on the basis of criteria defined by the project's partner cooperatives. The main criterion will be the geographical location of the members of the cluster, the latter having to be geographically close. In any case, each of the farmers volunteering to participate in the experiment must be a member of one and only one cluster.
In the first phase of randomization (lottery #1), the clusters are randomly assigned to the arms of the trial, which include one control group and two treatment groups. People in the treatment groups then participate in a second lottery (lottery #2), which will designate individuals inside each cluster who will be trained. People in the control group are not offered such an opportunity during the experimental period.
|
After
We run a stratified randomized control trial (RCT) that include three arms: two treatment arms(A and B) and one control arm. The first strata includes farmers previously identified as early adopters. The second strata include ordinary peers. In the first phase of randomization, individuals are randomly assigned to the arms of the trial. Thereafter, we form peer groups of around ten farmers within each treatment arm. After peer groups are defined, we then randomly select injection points within peer groups of treatment arms A and B. In treatment arm A (resp., B), the injection point is randomly selected among early adopters (resp., ordinary peers) in each peer group. In the control group, no one is designated as an injection point (neither early adopters nor ordinary peers) since no information is disseminated in this treatment arm.
|
|
Field
Randomization Method
|
Before
Random assignment to the arms of the trial (lottery #1), as well as random selection of injection points (lottery #2) will be performed using computer software.
|
After
Random assignment to the arms of the trial, as well as random selection of injection points among ordinary peers in treatment arm B will be performed using computer software.
|
|
Field
Randomization Unit
|
Before
We will randomize clusters.
|
After
We will randomize individuals.
|
|
Field
Was the treatment clustered?
|
Before
Yes
|
After
No
|
|
Field
Planned Number of Clusters
|
Before
480 clusters of 6 farmers each
|
After
Randomization is conducted at the individual level, not based on clusters.
|
|
Field
Planned Number of Observations
|
Before
Total number of farmers is 2880.
|
After
1,500 farmers
|
|
Field
Sample size (or number of clusters) by treatment arms
|
Before
Sample size by treatment arm is 960 farmers (160 clusters)
|
After
500 individuals in treatment arm A, 500 individuals in treatment arm B, 500 individuals in the control group.
|
|
Field
Power calculation: Minimum Detectable Effect Size for Main Outcomes
|
Before
With alpha = 0.05 and power = 0.8 the minimum detectable effect size, i.e. the minimum difference in knowledge test pass rates is 0.0832.
|
After
For plausible values of standard deviations and intra-cluster correlations, we should be able to detect a difference of less than 2 points between the scores of two treatment arms.
|
|
Field
Additional Keyword(s)
|
Before
social learning, agroecological transition
|
After
Social Learning, Peer Learning, Peer Effects, Agroecological Transition, Agricultural Technology
|
|
Field
Keyword(s)
|
Before
Agriculture, Finance
|
After
Agriculture, Behavior, Environment And Energy
|
|
Field
Intervention (Hidden)
|
Before
The experiment requires the implementation of a pilot program, in which the clusters are randomly assigned to training. The proportion of farmers within each cluster who are offered training varies between treatment arms. We call these farmers the injection points of new knowledge. In the first treatment arm (group A), we set the proportion of injection points at 1/3. In the second treatment arm (group B), the proportion is set at 1/2. Clusters from the control group (group C) are not offered to participate in the training. In the case of a linear effect of the diffusion of knowledge, the impact in group B will be 1.5 times that of the impact in group A. Otherwise, we can conclude that there is a non-linearity of diffusion (threshold effect).
|
After
The digital platform, managed by the service provider Landfiles, has been customized for the experiment. In particular, participants do not have the possibility to share the content outside the platform, or even outside their peer group. Only the injection points receive the information specially designed for the study; they can only share it with members of their group. Members of the same peer group can exchange with each other, but not with members of other peer groups.
While all participants receive some generic content linked to environmental regulations (non-technical and not focused on agroecological practices) throughout the experiment, participants assigned to treatment arms A and B will also receive a specific content on ecological practices, which we call the intervention. The content of the intervention will take the form of portraits of innovative farms, follow-up of agronomic trials, and newsletters. It will include several agroecological topics, including companion plants, plant cover, non-tillage associated with the reduction of herbicides, cereal-vegetable association, foliar fertilization. In practice, the injection points in arms~A and~B will receive one portrait of an innovative farm every two weeks, one agronomic trial follow-up every three weeks, and one newsletter every month.
|
|
Field
Secondary Outcomes (End Points)
|
Before
Other outcomes include indicators of the environmental performance of farmers, the use of chemical pesticides and the adoption of alternative practices notably.
|
After
Secondary outcomes include scores of knowledge tests relating to the content distributed to the treated groups through the platform. The control group clusters take the knowledge tests too.
|
|
Field
Secondary Outcomes (Explanation)
|
Before
Insofar as the training provided has adapted to the problem faced by the participants, it is interesting to look beyond the simple acquisition of new knowledge and to evaluate the change in practices likely to be triggered by the acquisition of this new knowledge.
|
After
The scores on the knowledge tests which are taken several times during the protocol and whose objective is to evaluate to what extent the information disseminated in the content within the groups has been assimilated by the recipients of the content. We expect that the acquisition of knowledge will be better in clusters whose injection point is a peer leader, since they are expected to receive more informative content, on a more regular basis. We also expect that the acquisition of knowledge will be better among injection points, since they receive first-hand information; it is also expected to be better among early adopters since they are expected to be more familiar with and more open to new techniques.
Insofar as the information reaches the recipients, we also evaluate the change in practices likely to be triggered by the acquisition of new knowledge. Here again, we expect that clusters whose injection point is a peer leader will outperform other clusters, that recipients of first-hand information will outperform recipients of second-hand information, and that peer leaders will outperform ordinary peers.
|