Key indicator groups (modules) generated by all RHoMIS applications
(Fig. 1 from: The Rural Household Multiple Indicator Survey,
data from 13,310 farm households in 21 countries).
Out this week is the first public release of a huge dataset
generated by recent surveys of more than
13,000 households in 21 countries
a novel tool that makes household surveys
efficient, robust and comparable.
The tool was co-developed by Mark van Wijk and Jim Hammond,
scientists at the International Livestock Research Institute (ILRI),
and the data generated by the tool’s use by (literally) dozens of partners.
Mark van Wijk, a senior scientist at the International Livestock Research Institute (ILRI), and his colleagues are excited to announce the first public release of a dataset generated by surveys of more than 13,000 households in 21 countries using a novel survey tool that van Wijk co-developed, as reported this week in Scientific Data, a peer-reviewed, open-access journal for descriptions of scientifically valuable datasets, and research that advances the sharing and reuse of scientific data, which is part of the Nature Group.
What’s the tool? A rapid, standardized and cost-effective tool employing a minimalist data approach for surveying rural smallholder households and tracking agricultural performance. Rolled out in May 2015, this flexible digital platform is built on open-source software that can be easily modified to meet a range of needs while collecting a core set of data.
The household survey tool includes data storage and analysis functions designed to rapidly characterize the state and change in farming households by a series of standardized indicators. It was designed in response to an expressed need from development practitioners to improve current approaches in targeting and prioritizing intervention options and monitoring farm households.
The big picture Since it was designed in 2015, RHoMIS, which stands for ‘Rural Household Multi-Indicator Survey’, has now been used in Central America; West, East and Central Africa; and South and Southeast Asia to characterize more than 7,000 farm households, evaluate management options, identify locally best-performing farmers, track changes in farm households over time and relate observed changes in farm household performance to changes in farm management and land use.
Use of this streamlined, modular survey tool has greatly reduced the costs of conducting household surveys in the field (traditional household surveys typically take 2–3 hours per household) and of the subsequent data analysis and reporting.
How does it work? This farm household survey is conducted on a digital platform using smart phones or tablets using the Open Data Kit suite of software installed on Android-based mobile phones or tablets. Data can be directly uploaded to a web server and an associated set of analysis tools programmed in R extract the data and calculate indicators.
What does it cost? Because RHoMIS is digital and implemented on open-source software, it is accessible to all institutions that have access to a computer and internet, for free.
What does it collect? RHoMIS collects information on 758 variables covering household demographics, farm area, crops grown and their production, livestock holdings and their production, agricultural product use and variables underlying standard socio-economic and food security indicators such as the Probability of Poverty Index, the Household Food Insecurity Access Scale, and household dietary diversity. Additional modules of questions and indicators can be incorporated and analysed depending on the local study needs.
What’s the purpose? These variables are used to quantify more than 40 different indicators on farm and household characteristics, welfare, productivity, and economic performance. The rapid characterization of farm systems, including household and farm welfare and livelihood strategies, that is made possible by RHoMIS supports planning, management and monitoring of specific agricultural and other development interventions and projects.
What’s new? Data from applications of the survey instrument in 21 countries in Central America, sub-Saharan Africa and Asia presented in this Nature article this week include the raw survey response data, the indicator calculation code, and the resulting indicator values.
Why it matters The standardization of indicators is a line of research that has been largely ignored in the current literature. The lack of standardized indicators means that only a small amount of the information collected during lengthy household surveys can actually be used for cross-site comparisons. This hampers our knowledge of trade-offs and/or synergies between indicators at farm household level and of how these relationships and trade-offs are shaped by farm management and by social and biophysical environments.
Furthermore, choosing suitable farm-level intervention options is challenging because different contexts require different recommendations and trade-offs can exist between different objectives, causing dilemmas between multiple household goals. These standardized data can be used to quantify on- and off-farm pathways to food security, diverse diets, and changes in poverty for rural smallholder farm households.
What’s so special about the new dataset? The ‘harmonized’ (aka, ‘useful to many’) dataset was developed from all the applications of RHoMIS that took place during the years 2015, 2016, 2017 and the first three months of 2018, resulting in a dataset collected from 13,310 farm households across 21 low- and middle-income countries.
The power of ‘harmonized’ datasets may be obscure to most of us but is critical to scientists. Lack of standardization of agricultural household surveys, especially in international ‘agriculture for development’ research, has resulted in a proliferation of survey tools and indicators leading to datasets that are often badly documented, incoherent and with limited interoperability. The current state of affairs limits our ability to compare outcomes across studies and to draw general conclusions on the effectiveness of interventions and trade-offs among outcomes.
The overall database (available at the Harvard Dataverse RHoMIS data repository) consists of the raw data (758 variables) and 41 indicators calculated based on the information provided by these variables.
What they’re saying Co-developers Mark van Wijk and Jim Hammond say:
- The RHoMIS tool has been designed to collect reliable information while putting a minimum burden on respondents, providing a rapid characterization of farm systems and performance indicators.
- The survey tool takes 40–60 minutes to administer per household using a digital implementation platform. This is linked to a set of automated analysis procedures that enable immediate cross-site bench-marking and intra-site characterization.
- With RHoMIS, we gain a consistent level of detail over a wide range of topics. This allows us to look for system interactions that are observable in a widespread diversity of locations, projects, cultures, or climate zones. The data can shed light on previously unknown interactions that should be acknowledged in the high-level design of development programs.
- RHoMIS reduces the costs, time requirements and reporting burdens for those who carry out household surveys. The development team have built and used a bank of survey questions based on internationally recognized indicators, covering all aspects of farming systems, including livestock.
- The database contains a wealth of information that may unlock important solutions to livestock challenges.
Between the lines The objective of RHoMIS is to gather information on the common variables of interest in all agricultural development research but not to go too deep into any one topic. The overall strategy of RHoMIS is to collect data that permits an overview of the farming system and the main livelihood activities. Based on this information, agricultural scientists and development experts can identify farm-level constraints, deficiencies or successes, and sift meaning from the high degree of variation observed among smallholder households. This is in contrast to the design of many impact assessment studies, which collect data on a narrow topic but at a higher resolution, thus permitting evaluation of that specific topic but limiting the ability to assess the over-arching farming system and rural livelihoods.
How are the raw data used? The raw data and indicators generated by RHoMIS have already been used for a wide range of studies at site level, for regional analyses and for continental analysis. Different aspects of smallholder households have been analysed, including gender equity, dietary diversity, nutritional gaps, poverty and greenhouse gas emissions in relation to production intensification, subsistence- versus market-oriented strategies and on-farm vs. off-farm activities.
State of play The survey itself is conducted on android smartphone or tablet. Data are uploaded to an internet server, either via a laptop or direct from the android device, for storage in a confidential database. The back-end analytical engine runs automated analysis routines that support almost real-time information delivery to front-line workers and program managers.
This near immediate feedback means that the time lag between data collection in the field and actionable information becomes very small. Shortening the duration is critical to improve adaptive management by helping development professionals to quickly identify successes and scale up what is working well while also quickly moving past what is not working without wasting time and money.
Go deeper The indicators captured by the RHoMIS tool were chosen to represent important factors across the agricultural production, nutrition and poverty relationships, while also capturing key indicators of interest related to climate smart agriculture, such as greenhouse gas emissions and gender equity. The survey tool was constructed in a modular way, with each module collecting the information needed to be able to calculate the performance an indicator of interest. New indicators can be added or removed as necessary for a given survey campaign.
RHoMIS aims to be a generic indicator framework. With additional modules bolted on, the potential applications of this tool are large and include integrated natural resource management, integrated nutrient management, conservation agriculture, organic agriculture, integrated pest management, agroforestry, integrated soil fertility management. RHoMIS can also be used to construct farm types to better target interventions across farming systems or to generate the appropriate inputs in modelling exercises for forecasting the impacts of interventions.
What is this survey tool designed to accomplish?
- To be quick to administer to avoid participant fatigue or annoyance while keeping costs low to allow for larger sample sizes on limited budgets.
- To be utilitarian, with all questions asked in the survey used in pre-defined analyses to minimize superfluous data collection.
- To be user friendly so that all participants in the process of collecting and analysing data are able to perform their tasks with minimum hassle and resistance, thereby increasing both the speed of their data collection and the quality of the data.
- To be flexible so that the survey tool can be easily modified to suit given the local context of the farming systems and farm households where it is deployed while maintaining its systematic and harmonized core indicator set.
- To generate reliable data, with the questions easy for respondents to understand and the answers based on observable criteria or respondents’ direct experience rather than abstract scales or abstract concepts.
The bottom line
- RHoMIS is a rapid, cheap, digital farm household-level survey and analytical engine for characterizing, targeting and monitoring agricultural performance.
- RHoMIS captures information describing farm productivity and practices, nutrition, food security, gender equity, climate and poverty.
- RHoMIS is action-ready, tested and adapted for diverse systems in more than 7,000 households across the global tropics.
What’s the future? RHoMIS is an on-going initiative, and its developers welcome interested parties to the community of practice (see http://www.rhomis.org for up-to-date information and downloadable survey questionnaires). Records continue to be submitted to the central data repository: in the latter part of 2018 more than 10,000 households were additionally interviewed, and their information added to the database. Further releases will be made public in the near future.
- The simplicity and flexibility of RHoMIS have catalyzed its spontaneous adoption.
- Uptake of RHoMIS by 12 organizations, including CGIAR centres, international non-governmental organizations and national agricultural research organizations, has happened only by word-of-mouth and without significant promotion of the tool.
- Users are not viewed as clients but as collaborators in the iterative development of the RHoMIS approach, which contributes to continuous improvements in the tool and the subsequent data analyses.
- RHoMIS is a starting point for a grassroots community of researchers and development practitioners working to solve the (large) challenges of targeting and monitoring projects.
- As this is an emergent community, the RHoMIS developers are always seeking new ideas and partners to extend and improve this approach.
- The RHoMIS developers welcome interested parties to the community of practice (see www.rhomis.org for up-to-date information and downloadable survey questionnaires). Records continue to be submitted to the central data repository: in the latter part of 2018 more than 10,000 households were additionally interviewed and their information added to the database. Further releases will be made public in the near future.
Get in touch For further information, please contact ILRI scientists Mark van Wijk (m.vanwijk [at] cgiar.org) or Jim Hammond (j.hammond [at] cgiar.org).
Read the whole article by ILRI’s Mark Van Wijk, Jim Hammond and others: The Rural Household Multiple Indicator Survey, data from 13,310 farm households in 21 countries, Scientific Data, 11 Feb 2020. DOI: https://doi.org/10.1038/s41597-020-0388-8
The RHoMIS data can be found at Harvard Dataverse.
Visit www.rhomis.org for more information, publications, data and analysis.
This work was supported by the United States Agency for International Development (USAID) and was implemented as part of the collaborative, ILRI-led CGIAR Research Program on Livestock, which is carried out with support from the CGIAR Trust Fund and through bilateral funding agreements.