EVALUATION OF THE EARLY IMPACTS OF THE BETTER COTTON INITIATIVE ON SMALLHOLDER COTTON PRODUCERS

Last registered on March 23, 2018
View Trial History

Pre-Trial

Trial Information

General Information

Title
EVALUATION OF THE EARLY IMPACTS OF THE BETTER COTTON INITIATIVE ON SMALLHOLDER COTTON PRODUCERS
RCT ID
AEARCTR-0000986
Initial registration date
January 19, 2016

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
January 19, 2016, 11:13 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
March 23, 2018, 11:30 AM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Country
India
Region
State - Andhra Pradesh

Primary Investigator

Name
Ravinder Kumar
Affiliation
Natural Resources Institute, University of Greenwich, UK
Contact Primary Investigator

Other Primary Investigator(s)

Additional Trial Information

Status
On going
Start date
2015-03-16
End date
2018-11-30
Keywords
Agriculture, Environment & Energy
Additional Keywords
Voluntary sustainability standards, better cotton initiative, social impact pathway, economic impact pathway, environmental impact pathway, value chain impact pathway
JEL code(s)
Secondary IDs
Abstract
The ISEAL Alliance Secretariat works with its sustainability standard members on various projects aimed at strengthening their approach to M&E systems, learning more about the impacts of standard systems, and determining how to increase the effectiveness of standards. The ‘Demonstrating and Improving Poverty Impacts’, has this aim and is funded by the Ford Foundation. ISEAL has commissioned a consortium led by the Natural Resources Institute, University of Greenwich, and including the Gujarat Institute of Development Research, the Centre for Economic and Social Studies, and Pragmatix Research and Advisory Services, to conduct an impact evaluation study of the early impact of pre-certification technical assistance and certification on previously uncertified smallholders.The Better Cotton Standard System is a holistic approach to sustainable cotton production that covers all three pillars of sustainability: environmental, social and economic. The study research questions are as follows:

• To what extent has the process of becoming or being certified under BCI sustainability standards had an impact (positive or negative, expected or unexpected) upon smallholders (farmers and households) in Kurnool district? What are the economic (yield, productivity, incomes, food security) and social (child labour, farm workers, no discrimination in wages for women) impacts?
• To what extent do we see an improvement in environment variables connected with cotton production (uptake of fertiliser use, reduction in pesticide use, efficient water use, soil health, habitat /biodiversity)?
• To what extent can Producers Unit and /or Farmer Producer Company ‘empower’ cotton farmers and households – both economically and socially?
• Can we see an increase in Better Cotton availability and uptake in the district /beyond? How can this be strengthened? What are the relative benefits and costs of meeting BCI standards and achieving certification for intended beneficiaries and supply chain actors?

In order to be able to measure and attribute impact, but also to understand what has created impact and identify lessons, the study employs a theory based evaluation approach. The theory of change lays out the anticipated chain of inputs, outputs, outcomes and impacts, and the causal linkages between them. A Randomized Control Trial (RCT) is found feasible primarily due to the willingness of the implementing partner to rollout their programme following a randomization strategy identified by the evaluation team. A cluster-RCT approach is proposed with the attribution of impacts of the BCI intervention package analysed by comparison of pre- and post-situation of intervention farmers and pre and post comparisons between intervention and non-intervention groups. The level of BCI project exposure to the farmers will also be assessed so that the analysis takes account of variations in implementation. Matched pair randomisation is used based on statistical data (village /cluster wise) from various existing sources. Observational approaches were employed following the lines of comparison of the experimental design to further interrogate and gather evidence on the theory of change (and alternative theories of change)– to assess what change has happened and why, through participatory field research techniques including household survey, focus group discussions, household panel and key informant interviews.
External Link(s)

Registration Citation

Citation
Kumar, Ravinder. 2018. "EVALUATION OF THE EARLY IMPACTS OF THE BETTER COTTON INITIATIVE ON SMALLHOLDER COTTON PRODUCERS ." AEA RCT Registry. March 23. https://doi.org/10.1257/rct.986-2.0
Former Citation
Kumar, Ravinder. 2018. "EVALUATION OF THE EARLY IMPACTS OF THE BETTER COTTON INITIATIVE ON SMALLHOLDER COTTON PRODUCERS ." AEA RCT Registry. March 23. https://www.socialscienceregistry.org/trials/986/history/27117
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
In order to be able to measure and attribute impact, but also to understand what has created impact and identify lessons, the study employs a theory based evaluation approach. The theory of change lays out the anticipated chain of inputs, outputs, outcomes and impacts, and the causal linkages between them. By identifying this expected impact chain and the associated assumptions inherent within it, we can gather evidence to establish whether the theory of change holds true and where it does not, i.e. where there are weak linkages, as well as unexpected and unintended consequences of the intervention. The theory of change allows for a consideration of the relative contribution of the intervention vis-à-vis other interventions. Through the combination of theory of change, which allows us to understand how impact has or has not been achieved, with randomized control trial –the latter enabling a rigorous attribution of impact -we can assess the impact of the intervention – in this case the BCI project implemented by PRDIS.

A Randomized Control Trial (RCT) combined with statistical analysis is found feasible primarily due to the willingness of the implementing partner to rollout their programme following a randomization strategy identified by the evaluation team. As BCI and implementation partner targets and resources are limited for Kurnool, in terms of outreach to farmers, randomisation is a fair approach to reach out to farmers and households (not unfair to other farmers not selected). The research team along with the implementation partner considered various options for randomization before finalizing the best suitable approach for the context of the BCI project.

A cluster-RCT approach is proposed with the attribution of impacts of the BCI intervention package analysed by comparison of pre- and post-situation of smallholders within the intervention mandal, and pre and post comparisons between intervention and non-intervention groups within the intervention mandal. The level of exposure of farmers will also be assessed so that the analysis takes account of variations in implementation.

Observational/ethnographic approaches will also be employed (following the lines of comparison of the experimental design) to interrogate the theory of change – to assess what change has happened and why through participatory field research techniques including focus group discussions, household panel survey and key informant interviews. The evaluation team will seek to understand the context of implementation, unexpected outcomes, and the relative influence of different drivers of change.
Intervention Start Date
2015-05-01
Intervention End Date
2018-11-30

Primary Outcomes

Primary Outcomes (end points)
Pesticide Use: Kilograms/hectare/for each active ingredient

Fertiliser Use: Kilograms/hectare/for each active ingredient

Water use for Irrigation: Cubic metres/hectare

Yield: Total cotton produced in kilograms of lint/total cotton production in hectares

Cotton profitability: Gross margin/hectare

Elimination of Child Labour - A: Leveraging partnership with local specialist organisations Existence of partnership(s) established by or on behalf of the Producer Unit with credible social organisations to address child labour, in particular to identify and reduce barriers to formal schooling.

Elimination of Child Labour – B: Improving understanding and awareness; Percentage of farmers who can accurately differentiate between acceptable forms of children’s work and hazardous child labour

Inclusion of women: Number of farmers and workers receiving BCI training who are women by training topic
Primary Outcomes (explanation)
Knowledge and practice adoption by farmers: The BCI verification /certification is awarded to the producer unit (based on a three-tier assessment of sample members). The experimental research charts out the progression of each individual member over a period of time in terms of their knowledge and application of BCI recommended practices. It also tracks the outcome variables (cost of production, yield, profitability, pesticide use etc.). The correlation between practices and outcomes are analysed. The research team have developed an index called Better Cotton Composite Index (BCCI) which tracks every member of the learning group (also those who are not part learning group in the intervention set and those belonging to ‘control’ set) in terms of their knowledge and application of BCI recommended practices. This is a simple and potentially replicable analytical tool.

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
A cluster RCT design is proposed as this is a more appropriate design given the study context of specific heterogeneity within a broadly homogenous mandal. The village was considered to be a cluster and the unit of random assignment, as methodologically it was not possible to randomly assign farmers /households to the intervention within a village, given the saturation approach. The saturation approach entails targeting all farmers for the programme, wherein a few join voluntarily in the first year and then progressively more become part of the project promoted learning groups /producer unit in the second and third year. The unit of intervention by the BCI project is the learning group (member households within the learning group) and the producer unit. The unit of assessment and analysis for this study is the household, learning group and clusters.
The cluster RCT is expected to reduce or eliminate the influence of confounding variables as cluster based randomisation will ensure that the randomly assigned clusters represent all typical situations available in the mandal. The cluster RCT design can address the control of the spread effect to a certain extent as the clusters are spread out across the mandal since they are selected based on stratification using bio-physical and socio-economic parameters and not on the basis of geographical proximity.
The implementation partner wanted to prioritise black or mixed soil areas in the first phase of project implementation (first 3 years) and hence this bio-physical measure (dense black, medium black, mixed soil) was used as a filter to create the universe for random selection of the clusters /villages. The sampling universe of 21 clusters (so obtained after applying the filter) was divided into 10 best matched pairs (using existing bio-physical and socio-economic parameters) and then from each pair, a treatment and control cluster/ village was randomly assigned.
Experimental Design Details
Randomization Method
Randomisation was done using Statistical package - STATA
Randomization Unit
Unit of randomization - village
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
10 clusters of intervention and non-intervention villages
Sample size: planned number of observations
729 farmers /households
Sample size (or number of clusters) by treatment arms
5 treatment clusters, and 5 control clusters
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials